PERF: performance regressions in 1.2.0rc #38591

jorisvandenbossche · 2020-12-20T16:39:36Z

I did a full benchmark run on a dedicated machine comparing v1.2.0rc0 with v1.1.5.

The top results:

     [b5958ee1]       [7688d3cf]
     <v1.1.5^0>       <v1.2.0rc0^0>
+     4.25±0.05μs          366±9ms 86187.42  index_object.IndexEquals.time_non_object_equals_multiindex
+        50.2±4μs          104±5ms  2075.23  indexing.NumericSeriesIndexing.time_getitem_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+         120±8μs          102±5ms   848.38  indexing.NumericSeriesIndexing.time_loc_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+        268±20μs          103±5ms   383.87  indexing.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+        289±40μs          111±8ms   382.15  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 1000000)
+        273±20μs          102±5ms   374.69  indexing.NumericSeriesIndexing.time_getitem_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      1.75±0.1ms          103±5ms    58.77  indexing.NumericSeriesIndexing.time_loc_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      2.11±0.1ms          103±5ms    48.69  indexing.NumericSeriesIndexing.time_getitem_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      7.38±0.7ms          109±7ms    14.80  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 1000000)
+      3.66±0.2μs       13.4±0.4μs     3.66  index_cached_properties.IndexCache.time_is_all_dates('Float64Index')
+         336±2ms          1.22±0s     3.65  groupby.TransformEngine.time_series_numba(True)
+         287±2ms          1.02±0s     3.55  groupby.AggEngine.time_series_numba(True)
+         289±2ms          1.02±0s     3.52  groupby.AggEngine.time_dataframe_numba(True)
+      3.64±0.2μs       12.8±0.5μs     3.52  index_cached_properties.IndexCache.time_is_all_dates('IntervalIndex')
+     1.89±0.09μs       6.49±0.1μs     3.44  index_cached_properties.IndexCache.time_is_all_dates('PeriodIndex')
+      3.80±0.2μs       12.7±0.4μs     3.35  index_cached_properties.IndexCache.time_is_all_dates('UInt64Index')
+      3.27±0.1μs       10.9±0.2μs     3.33  index_cached_properties.IndexCache.time_is_all_dates('MultiIndex')
+        728±30ns      2.39±0.06μs     3.28  index_cached_properties.IndexCache.time_is_all_dates('RangeIndex')
+     2.02±0.09μs       6.49±0.1μs     3.21  index_cached_properties.IndexCache.time_is_all_dates('DatetimeIndex')
+        719±30ns      2.27±0.06μs     3.15  index_cached_properties.IndexCache.time_is_all_dates('Int64Index')
+      4.11±0.2μs       12.4±0.3μs     3.02  index_cached_properties.IndexCache.time_is_all_dates('TimedeltaIndex')
+         430±3ms          1.23±0s     2.85  groupby.TransformEngine.time_dataframe_numba(True)
+     7.44±0.08ms         17.1±2ms     2.29  hash_functions.UniqueAndFactorizeArange.time_unique(6)
+        75.3±1ms          172±1ms     2.28  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 1000000)
+     7.64±0.04ms         17.1±2ms     2.24  hash_functions.UniqueAndFactorizeArange.time_unique(5)
+     1.62±0.04ms      3.51±0.02ms     2.16  arithmetic.Timeseries.time_series_timestamp_compare(None)
+     1.60±0.03ms      3.45±0.04ms     2.16  arithmetic.Timeseries.time_timestamp_series_compare(None)
+     1.62±0.03ms      3.49±0.09ms     2.16  arithmetic.Timeseries.time_timestamp_series_compare('US/Eastern')
+      10.9±0.7ms         23.2±1ms     2.14  hash_functions.UniqueAndFactorizeArange.time_factorize(6)
+      11.1±0.5ms         23.2±1ms     2.10  hash_functions.UniqueAndFactorizeArange.time_factorize(5)
+     1.63±0.04ms      3.41±0.03ms     2.09  arithmetic.Timeseries.time_series_timestamp_compare('US/Eastern')
+      4.11±0.1ms      8.52±0.09ms     2.08  index_object.SetDisjoint.time_datetime_difference_disjoint
+      74.5±0.3ms          153±1ms     2.05  replace.ReplaceList.time_replace_list_one_match(True)
+       124±0.9ms        248±0.4ms     2.00  gil.ParallelDatetimeFields.time_datetime_to_period

The first, biggest regression in the IndexEquals benchmark is probably already fixed in #38560

Full results:

       before           after         ratio
     [b5958ee1]       [7688d3cf]
     <v1.1.5^0>       <v1.2.0rc0^0>
+     4.25±0.05μs          366±9ms 86187.42  index_object.IndexEquals.time_non_object_equals_multiindex
+        50.2±4μs          104±5ms  2075.23  indexing.NumericSeriesIndexing.time_getitem_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+         120±8μs          102±5ms   848.38  indexing.NumericSeriesIndexing.time_loc_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+        268±20μs          103±5ms   383.87  indexing.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+        289±40μs          111±8ms   382.15  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 1000000)
+        273±20μs          102±5ms   374.69  indexing.NumericSeriesIndexing.time_getitem_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      1.75±0.1ms          103±5ms    58.77  indexing.NumericSeriesIndexing.time_loc_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      2.11±0.1ms          103±5ms    48.69  indexing.NumericSeriesIndexing.time_getitem_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      7.38±0.7ms          109±7ms    14.80  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 1000000)
+      3.66±0.2μs       13.4±0.4μs     3.66  index_cached_properties.IndexCache.time_is_all_dates('Float64Index')
+         336±2ms          1.22±0s     3.65  groupby.TransformEngine.time_series_numba(True)
+         287±2ms          1.02±0s     3.55  groupby.AggEngine.time_series_numba(True)
+         289±2ms          1.02±0s     3.52  groupby.AggEngine.time_dataframe_numba(True)
+      3.64±0.2μs       12.8±0.5μs     3.52  index_cached_properties.IndexCache.time_is_all_dates('IntervalIndex')
+     1.89±0.09μs       6.49±0.1μs     3.44  index_cached_properties.IndexCache.time_is_all_dates('PeriodIndex')
+      3.80±0.2μs       12.7±0.4μs     3.35  index_cached_properties.IndexCache.time_is_all_dates('UInt64Index')
+      3.27±0.1μs       10.9±0.2μs     3.33  index_cached_properties.IndexCache.time_is_all_dates('MultiIndex')
+        728±30ns      2.39±0.06μs     3.28  index_cached_properties.IndexCache.time_is_all_dates('RangeIndex')
+     2.02±0.09μs       6.49±0.1μs     3.21  index_cached_properties.IndexCache.time_is_all_dates('DatetimeIndex')
+        719±30ns      2.27±0.06μs     3.15  index_cached_properties.IndexCache.time_is_all_dates('Int64Index')
+      4.11±0.2μs       12.4±0.3μs     3.02  index_cached_properties.IndexCache.time_is_all_dates('TimedeltaIndex')
+         430±3ms          1.23±0s     2.85  groupby.TransformEngine.time_dataframe_numba(True)
+     7.44±0.08ms         17.1±2ms     2.29  hash_functions.UniqueAndFactorizeArange.time_unique(6)
+        75.3±1ms          172±1ms     2.28  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 1000000)
+     7.64±0.04ms         17.1±2ms     2.24  hash_functions.UniqueAndFactorizeArange.time_unique(5)
+     1.62±0.04ms      3.51±0.02ms     2.16  arithmetic.Timeseries.time_series_timestamp_compare(None)
+     1.60±0.03ms      3.45±0.04ms     2.16  arithmetic.Timeseries.time_timestamp_series_compare(None)
+     1.62±0.03ms      3.49±0.09ms     2.16  arithmetic.Timeseries.time_timestamp_series_compare('US/Eastern')
+      10.9±0.7ms         23.2±1ms     2.14  hash_functions.UniqueAndFactorizeArange.time_factorize(6)
+      11.1±0.5ms         23.2±1ms     2.10  hash_functions.UniqueAndFactorizeArange.time_factorize(5)
+     1.63±0.04ms      3.41±0.03ms     2.09  arithmetic.Timeseries.time_series_timestamp_compare('US/Eastern')
+      4.11±0.1ms      8.52±0.09ms     2.08  index_object.SetDisjoint.time_datetime_difference_disjoint
+      74.5±0.3ms          153±1ms     2.05  replace.ReplaceList.time_replace_list_one_match(True)
+       124±0.9ms        248±0.4ms     2.00  gil.ParallelDatetimeFields.time_datetime_to_period
+        950±30μs      1.83±0.05ms     1.92  reindex.LevelAlign.time_reindex_level
+        2.84±0ms      5.46±0.02ms     1.92  tslibs.normalize.Normalize.time_is_date_array_normalized(1000000, datetime.timezone(datetime.timedelta(seconds=3600)))
+         283±2ms          535±1ms     1.89  groupby.AggEngine.time_series_numba(False)
+         284±2ms          534±2ms     1.88  groupby.AggEngine.time_dataframe_numba(False)
+      11.8±0.3ms       21.4±0.4ms     1.81  io.csv.ToCSVDatetime.time_frame_date_formatting
+         331±3ms          599±2ms     1.81  groupby.TransformEngine.time_series_numba(False)
+      9.45±0.5ms         17.1±2ms     1.81  hash_functions.UniqueAndFactorizeArange.time_unique(4)
+      16.8±0.1ms       30.3±0.8ms     1.80  gil.ParallelDatetimeFields.time_datetime_field_normalize
+        65.8±2ms          113±2ms     1.71  series_methods.IsInFloat64.time_isin_many_different
+        192±20μs         326±20μs     1.70  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 500000)
+      37.2±0.1μs      62.3±0.09μs     1.68  tslibs.normalize.Normalize.time_is_date_array_normalized(10000, datetime.timezone(datetime.timedelta(seconds=3600)))
+      13.9±0.4ms         23.2±1ms     1.67  hash_functions.UniqueAndFactorizeArange.time_factorize(4)
+         508±2ms          837±5ms     1.65  stat_ops.FrameMultiIndexOps.time_op([0, 1], 'skew')
+      11.7±0.1ms       18.5±0.2ms     1.58  period.PeriodIndexConstructor.time_from_ints('D', False)
+      11.8±0.1ms       18.5±0.5ms     1.57  period.PeriodIndexConstructor.time_from_ints('D', True)
+     4.24±0.03ms      6.58±0.03ms     1.55  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 100000)
+         438±1ms          680±2ms     1.55  series_methods.IsInLongSeriesValuesDominate.time_isin('float64', 'monotone')
+      94.0±0.2ms        143±0.8ms     1.52  groupby.GroupByMethods.time_dtype_as_field('datetime', 'unique', 'transformation')
+      94.2±0.1ms        143±0.5ms     1.52  groupby.GroupByMethods.time_dtype_as_field('datetime', 'unique', 'direct')
+         616±3μs          937±8μs     1.52  stat_ops.FrameOps.time_op('sum', 'int', 0)
+      21.3±0.8ms       32.4±0.7ms     1.52  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 19)
+      41.1±0.3ms        62.2±20ms     1.51  rolling.Apply.time_rolling('Series', 300, 'int', <built-in function sum>, False)
+     1.87±0.02ms         2.80±0ms     1.49  series_methods.IsIn.time_isin('uint64')
+      27.7±0.2μs       41.1±0.1μs     1.48  tslibs.normalize.Normalize.time_is_date_array_normalized(10000, None)
+         1.32±0s       1.95±0.01s     1.48  stat_ops.FrameMultiIndexOps.time_op([0, 1], 'mad')
+     2.97±0.01ms       4.40±0.7ms     1.48  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'count')
+     2.93±0.02ms       4.32±0.6ms     1.47  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'count')
+        1.46±0ms      2.14±0.04ms     1.47  groupby.FillNA.time_srs_ffill
+     1.47±0.01ms      2.14±0.01ms     1.46  groupby.FillNA.time_srs_bfill
+      96.7±0.1ms          141±2ms     1.46  rolling.Groupby.time_rolling_int('min')
+      28.0±0.2μs       40.8±0.2μs     1.46  tslibs.normalize.Normalize.time_is_date_array_normalized(10000, datetime.timezone.utc)
+     2.82±0.01ms      4.11±0.01ms     1.46  tslibs.normalize.Normalize.time_is_date_array_normalized(1000000, None)
+      97.4±0.6ms          142±1ms     1.46  rolling.Groupby.time_rolling_int('kurt')
+      18.5±0.1ms       27.0±0.2ms     1.46  frame_methods.Iteration.time_items
+      2.81±0.1ms      4.09±0.01ms     1.46  tslibs.normalize.Normalize.time_is_date_array_normalized(1000000, datetime.timezone.utc)
+      97.5±0.5ms        142±0.9ms     1.45  rolling.Groupby.time_rolling_int('mean')
+      97.4±0.9ms          141±1ms     1.45  rolling.Groupby.time_rolling_int('max')
+         751±4μs      1.09±0.01ms     1.45  stat_ops.FrameOps.time_op('mean', 'int', 0)
+      99.9±0.7ms          144±1ms     1.44  rolling.Groupby.time_rolling_int('median')
+        97.0±1ms        140±0.9ms     1.44  rolling.Groupby.time_rolling_int('sum')
+        11.9±2ms         17.1±2ms     1.44  hash_functions.UniqueAndFactorizeArange.time_unique(15)
+     2.98±0.02ms       4.27±0.7ms     1.43  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'count')
+      9.98±0.4ms       14.3±0.5ms     1.43  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 18)
+     8.77±0.02ms      12.5±0.03ms     1.43  frame_methods.ToString.time_to_string_floats
+      4.25±0.03s       6.03±0.01s     1.42  replace.ReplaceDict.time_replace_series(False)
+        786±10μs         1.11±0ms     1.41  stat_ops.FrameOps.time_op('prod', 'int', 0)
+       122±0.5ms        173±0.8ms     1.41  frame_methods.Iteration.time_iteritems_indexing
+       157±0.8ms          221±2ms     1.41  stat_ops.FrameMultiIndexOps.time_op(1, 'mad')
+       134±0.8ms         188±20ms     1.41  gil.ParallelGroupbyMethods.time_loop(8, 'count')
+         427±1ms          599±1ms     1.40  groupby.TransformEngine.time_dataframe_numba(False)
+         284±3ms          398±8ms     1.40  frame_methods.GetDtypeCounts.time_info
+      33.5±0.2ms         46.6±5ms     1.39  gil.ParallelGroupbyMethods.time_loop(2, 'count')
+      74.5±0.4ms        103±0.9ms     1.38  stat_ops.FrameMultiIndexOps.time_op(1, 'skew')
+        60.9±3ms         84.0±3ms     1.38  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 20)
+     6.97±0.05ms      9.60±0.06ms     1.38  groupby.Apply.time_scalar_function_single_col
+         223±5ms          304±4ms     1.36  groupby.MultiColumn.time_lambda_sum
+     11.7±0.01ms      15.8±0.03ms     1.35  stat_ops.Correlation.time_corr_wide('pearson')
+      25.4±0.4μs       34.2±0.2μs     1.35  indexing.NumericSeriesIndexing.time_iloc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+     12.2±0.06μs       16.4±0.1μs     1.34  categoricals.CategoricalSlicing.time_getitem_slice('monotonic_decr')
+      3.98±0.2ms      5.33±0.04ms     1.34  rolling.Methods.time_rolling('DataFrame', 1000, 'float', 'kurt')
+        229±10ms         306±10ms     1.34  replace.ReplaceList.time_replace_list_one_match(False)
+      26.1±0.4μs       34.9±0.7μs     1.34  indexing.NumericSeriesIndexing.time_iloc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 'unique_monotonic_inc')
+     14.4±0.06ms       19.2±0.7ms     1.33  strings.Cat.time_cat(0, None, None, 0.0)
+         122±3ms          162±2ms     1.33  groupby.MultiColumn.time_col_select_lambda_sum
+     12.3±0.05μs       16.4±0.3μs     1.33  categoricals.CategoricalSlicing.time_getitem_slice('non_monotonic')
+     12.3±0.08μs       16.3±0.3μs     1.32  categoricals.CategoricalSlicing.time_getitem_slice('monotonic_incr')
+      25.9±0.4μs       34.3±0.4μs     1.32  indexing.NumericSeriesIndexing.time_iloc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'nonunique_monotonic_inc')
+      58.0±0.2μs         76.3±1μs     1.32  indexing.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 'unique_monotonic_inc')
+      67.3±0.3ms         88.6±9ms     1.32  gil.ParallelGroupbyMethods.time_loop(4, 'count')
+     2.53±0.03ms      3.33±0.02ms     1.32  categoricals.Indexing.time_sort_values
+      26.3±0.4μs       34.6±0.2μs     1.31  indexing.NumericSeriesIndexing.time_iloc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'unique_monotonic_inc')
+     20.7±0.03ms      27.1±0.03ms     1.31  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 8000, -2)
+      25.9±0.1μs       33.9±0.2μs     1.31  indexing.NumericSeriesIndexing.time_iloc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     2.38±0.02ms      3.12±0.03ms     1.31  groupby.FillNA.time_df_ffill
+     4.33±0.01ms      5.66±0.02ms     1.31  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 1000, -2)
+      60.3±0.4μs       78.8±0.8μs     1.31  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 100000)
+     3.92±0.02ms       5.12±0.2ms     1.31  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'float', 'kurt')
+        90.4±2ms        118±0.7ms     1.30  timeseries.Iteration.time_iter_preexit(<function timedelta_range at 0x7f2782aecaf0>)
+     3.94±0.01ms      5.12±0.02ms     1.30  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'kurt')
+      31.0±0.2μs       40.2±0.2μs     1.30  indexing.NumericSeriesIndexing.time_getitem_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 'nonunique_monotonic_inc')
+      3.09±0.3ms      4.00±0.06ms     1.30  rolling.ExpandingMethods.time_expanding('DataFrame', 'float', 'kurt')
+     2.42±0.01ms      3.13±0.02ms     1.29  groupby.FillNA.time_df_bfill
+      8.73±0.3ms      11.3±0.05ms     1.29  timeseries.ResampleSeries.time_resample('period', '5min', 'ohlc')
+         105±1ms        135±0.6ms     1.29  groupby.Groups.time_series_groups('int64_large')
+      29.4±0.2ms       37.8±0.2ms     1.29  rolling.Apply.time_rolling('Series', 3, 'int', <built-in function sum>, False)
+         455±1ms        584±0.7ms     1.28  series_methods.IsInLongSeriesValuesDominate.time_isin('float32', 'monotone')
+     8.95±0.03ms       11.5±0.1ms     1.28  index_object.IntervalIndexMethod.time_intersection(100000)
+     2.60±0.02ms      3.33±0.09ms     1.28  stat_ops.SeriesMultiIndexOps.time_op(1, 'prod')
+     3.20±0.02ms      4.10±0.03ms     1.28  rolling.ExpandingMethods.time_expanding('DataFrame', 'int', 'kurt')
+      29.9±0.2ms       38.2±0.2ms     1.28  rolling.Apply.time_rolling('Series', 3, 'float', <built-in function sum>, False)
+      29.9±0.3μs       38.2±0.1μs     1.28  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('string', 'non_monotonic')
+        18.3±1ms         23.2±1ms     1.27  hash_functions.UniqueAndFactorizeArange.time_factorize(14)
+      30.2±0.6μs       38.4±0.4μs     1.27  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('string', 'unique_monotonic_inc')
+      31.3±0.4μs       39.8±0.3μs     1.27  indexing.NumericSeriesIndexing.time_getitem_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'unique_monotonic_inc')
+     16.1±0.09ms       20.5±0.2ms     1.27  groupby.AggFunctions.time_different_str_functions
+      30.3±0.5μs       38.3±0.4μs     1.27  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('string', 'nonunique_monotonic_inc')
+      30.2±0.2ms       38.2±0.4ms     1.27  rolling.Apply.time_rolling('DataFrame', 3, 'float', <built-in function sum>, False)
+     2.62±0.02ms       3.32±0.1ms     1.27  stat_ops.SeriesMultiIndexOps.time_op(1, 'sum')
+     3.07±0.02ms      3.89±0.02ms     1.26  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'int', 'kurt')
+     3.90±0.01ms      4.91±0.02ms     1.26  series_methods.IsInDatetime64.time_isin_cat_values
+       115±0.5μs          145±1μs     1.26  tslibs.normalize.Normalize.time_is_date_array_normalized(10000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+     2.61±0.03ms       3.30±0.1ms     1.26  stat_ops.SeriesMultiIndexOps.time_op(0, 'sum')
+      14.6±0.6ms         18.3±1ms     1.26  tslibs.normalize.Normalize.time_is_date_array_normalized(1000000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+     21.1±0.04ms      26.4±0.07ms     1.25  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 8000, 2)
+     4.50±0.03μs      5.64±0.05μs     1.25  categoricals.CategoricalSlicing.time_getitem_scalar('non_monotonic')
+     16.1±0.08ms      20.2±0.09ms     1.25  groupby.AggFunctions.time_different_numpy_functions
+      30.2±0.3ms       37.9±0.1ms     1.25  rolling.Apply.time_rolling('DataFrame', 3, 'int', <built-in function sum>, False)
+       359±0.5ms          450±4ms     1.25  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 5, 'monotone_misses')
+      37.2±0.5μs       46.4±0.4μs     1.25  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('string', 'non_monotonic')
+       142±0.4ms        177±0.3ms     1.24  hash_functions.IsinWithArange.time_isin(<class 'object'>, 2000, 2)
+      21.7±0.1ms       26.9±0.1ms     1.24  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 8000, -2)
+         271±2μs          335±3μs     1.24  reindex.Fillna.time_float_32('backfill')
+     6.95±0.04ms       8.58±0.1ms     1.23  stat_ops.Correlation.time_corrwith_cols('pearson')
+      48.1±0.4μs       59.3±0.6μs     1.23  frame_methods.GetNumericData.time_frame_get_numeric_data
+      45.7±0.2μs       56.4±0.8μs     1.23  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 10000)
+      6.06±0.4ms       7.47±0.4ms     1.23  stat_ops.FrameMultiIndexOps.time_op([0, 1], 'sum')
+     6.22±0.07μs      7.67±0.04μs     1.23  dtypes.Dtypes.time_pandas_dtype('Int8')
+       108±0.2ms        133±0.2ms     1.23  hash_functions.IsinWithArange.time_isin(<class 'object'>, 2000, -2)
+     9.49±0.08μs       11.7±0.1μs     1.23  indexing.CategoricalIndexIndexing.time_getitem_scalar('monotonic_incr')
+     4.53±0.04μs      5.57±0.04μs     1.23  categoricals.CategoricalSlicing.time_getitem_scalar('monotonic_decr')
+      4.13±0.1ms       5.07±0.2ms     1.23  tslibs.normalize.Normalize.time_normalize_i8_timestamps(1000000, None)
+     4.54±0.04μs      5.57±0.03μs     1.23  categoricals.CategoricalSlicing.time_getitem_scalar('monotonic_incr')
+         137±1μs          168±2μs     1.23  hash_functions.IsinWithArangeSorted.time_isin(<class 'object'>, 1000)
+      56.4±0.3μs         68.9±1μs     1.22  indexing.NumericSeriesIndexing.time_iloc_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     1.54±0.02ms      1.88±0.03ms     1.22  reindex.ReindexMethod.time_reindex_method('pad', <function period_range at 0x7f2782af4700>)
+      3.76±0.2ms      4.59±0.03ms     1.22  rolling.Methods.time_rolling('DataFrame', 1000, 'float', 'skew')
+      25.7±0.3ms       31.4±0.4ms     1.22  groupby.ApplyDictReturn.time_groupby_apply_dict_return
+     1.19±0.02μs      1.45±0.01μs     1.21  tslibs.normalize.Normalize.time_is_date_array_normalized(100, None)
+      56.6±0.6μs         68.8±1μs     1.21  indexing.NumericSeriesIndexing.time_iloc_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+      62.8±0.1μs         76.0±1μs     1.21  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 100000)
+     4.14±0.01ms      5.00±0.02ms     1.21  tslibs.normalize.Normalize.time_normalize_i8_timestamps(1000000, datetime.timezone.utc)
+        325±20ms         392±10ms     1.21  indexing.NumericSeriesIndexing.time_getitem_lists(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+      28.4±0.2ms       34.3±0.2ms     1.21  groupby.AggFunctions.time_different_python_functions_multicol
+         134±2μs        161±0.9μs     1.21  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'object'>, 10)
+        58.9±2ms         71.2±3ms     1.21  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 20)
+     5.19±0.04ms      6.27±0.04ms     1.21  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 17)
+      9.55±0.1μs      11.5±0.09μs     1.20  indexing.CategoricalIndexIndexing.time_getitem_scalar('monotonic_decr')
+       513±0.7μs          617±3μs     1.20  stat_ops.Correlation.time_corr('pearson')
+        891±10μs      1.07±0.01ms     1.20  reindex.ReindexMethod.time_reindex_method('pad', <function date_range at 0x7f2782b204c0>)
+        321±20ms         385±10ms     1.20  indexing.NumericSeriesIndexing.time_loc_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     2.62±0.02ms      3.15±0.07ms     1.20  stat_ops.SeriesMultiIndexOps.time_op(0, 'prod')
+        323±20ms         388±10ms     1.20  indexing.NumericSeriesIndexing.time_loc_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+      86.7±0.6ms        104±0.5ms     1.20  frame_methods.ToHTML.time_to_html_mixed
+     8.31±0.07ms      9.97±0.05ms     1.20  frame_methods.Apply.time_apply_pass_thru
+      44.7±0.3μs      53.6±0.08μs     1.20  tslibs.normalize.Normalize.time_normalize_i8_timestamps(10000, None)
+       182±0.4μs          218±2μs     1.20  hash_functions.IsinWithArangeSorted.time_isin(<class 'object'>, 2000)
+     4.64±0.02ms      5.56±0.05ms     1.20  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'kurt')
+     5.25±0.03ms       6.28±0.1ms     1.20  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 2000, 0)
+         653±2μs          782±4μs     1.20  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 8000)
+     7.09±0.06μs      8.48±0.03μs     1.20  dtypes.Dtypes.time_pandas_dtype('Int16')
+         609±4ms          728±2ms     1.20  frame_methods.Nunique.time_frame_nunique
+        324±20ms         387±10ms     1.20  indexing.NumericSeriesIndexing.time_getitem_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     22.0±0.07ms      26.3±0.04ms     1.20  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 8000, 2)
+      50.4±0.2μs       60.2±0.3μs     1.20  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('string', 'nonunique_monotonic_inc')
+         130±2μs        155±0.7μs     1.19  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('datetime', 'unique_monotonic_inc')
+         595±2ms          710±1ms     1.19  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 5000000)
+     44.9±0.08μs       53.5±0.1μs     1.19  tslibs.normalize.Normalize.time_normalize_i8_timestamps(10000, datetime.timezone.utc)
+       122±0.9μs          146±3μs     1.19  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('datetime', 'non_monotonic')
+         603±5μs          719±4μs     1.19  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 7000)
+      82.4±0.8μs       98.2±0.3μs     1.19  indexing.DataFrameNumericIndexing.time_iloc
+      16.0±0.1ms      19.1±0.05ms     1.19  categoricals.Isin.time_isin_categorical('object')
+      90.3±0.6μs          107±2μs     1.19  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.Int64Index'>, 'unique_monotonic_inc')
+       196±0.6ms          233±2ms     1.19  groupby.Groups.time_series_groups('object_large')
+      47.3±0.2μs       56.2±0.7μs     1.19  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 10000)
+     9.68±0.07μs      11.5±0.09μs     1.19  indexing.CategoricalIndexIndexing.time_getitem_scalar('non_monotonic')
+        1.06±0ms         1.25±0ms     1.19  timeseries.DatetimeIndex.time_unique('repeated')
+     4.66±0.02ms      5.52±0.02ms     1.18  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'kurt')
+     4.54±0.03ms      5.38±0.05ms     1.18  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'kurt')
+     2.59±0.01ms      3.07±0.01ms     1.18  series_methods.IsInForObjects.time_isin_long_series_short_values
+     5.40±0.02ms      6.39±0.06ms     1.18  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 8000, 0)
+         136±1ms        161±0.5ms     1.18  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 900000)
+        326±20ms         385±20ms     1.18  indexing.NumericSeriesIndexing.time_getitem_list_like(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+        66.1±1μs       78.1±0.4μs     1.18  indexing.DataFrameNumericIndexing.time_loc
+      3.86±0.2ms      4.55±0.01ms     1.18  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 0, 'midpoint')
+        667±10μs          785±5μs     1.18  groupby.GroupManyLabels.time_sum(1)
+     3.38±0.04ms      3.98±0.08ms     1.18  stat_ops.SeriesMultiIndexOps.time_op(1, 'var')
+     7.51±0.03μs       8.84±0.2μs     1.18  dtypes.Dtypes.time_pandas_dtype('interval')
+     2.88±0.02ms      3.39±0.06ms     1.18  stat_ops.FrameMultiIndexOps.time_op(0, 'sum')
+     2.84±0.04ms      3.34±0.05ms     1.18  stat_ops.FrameMultiIndexOps.time_op(0, 'mean')
+      26.8±0.1ms      31.5±0.06ms     1.18  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 2000, -2)
+      3.86±0.2ms      4.54±0.02ms     1.18  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 0, 'lower')
+         145±1μs        170±0.8μs     1.18  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'object'>, 10)
+     9.33±0.05μs      11.0±0.04μs     1.18  dtypes.Dtypes.time_pandas_dtype('UInt8')
+     2.89±0.02ms      3.39±0.03ms     1.17  stat_ops.FrameMultiIndexOps.time_op(1, 'prod')
+     5.21±0.02ms      6.12±0.01ms     1.17  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 1000, 0)
+     2.24±0.03ms      2.63±0.04ms     1.17  stat_ops.FrameOps.time_op('var', 'int', 0)
+     7.86±0.03μs      9.22±0.03μs     1.17  dtypes.Dtypes.time_pandas_dtype('Int32')
+     5.03±0.03ms      5.90±0.03ms     1.17  tslibs.normalize.Normalize.time_normalize_i8_timestamps(1000000, datetime.timezone(datetime.timedelta(seconds=3600)))
+     26.4±0.03ms      30.9±0.04ms     1.17  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 2000, 2)
+     4.72±0.06ms      5.53±0.06ms     1.17  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 80000)
+         909±4μs      1.06±0.05ms     1.17  dtypes.SelectDtypes.time_select_dtype_int_include(<class 'float'>)
+     2.86±0.04ms      3.35±0.02ms     1.17  stat_ops.FrameMultiIndexOps.time_op(1, 'mean')
+     8.63±0.05μs       10.1±0.5μs     1.17  dtypes.Dtypes.time_pandas_dtype('Int64')
+      90.9±0.8μs          106±1μs     1.17  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'nonunique_monotonic_inc')
+       199±0.3ms          232±1ms     1.17  frame_methods.Apply.time_apply_axis_1
+      3.89±0.2ms      4.54±0.01ms     1.17  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 0, 'nearest')
+        76.3±1ms       89.0±0.6ms     1.17  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 750000)
+         739±5μs          862±2μs     1.17  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 8000)
+     4.64±0.03ms      5.41±0.03ms     1.17  index_object.Indexing.time_get_loc_non_unique('Float')
+     3.39±0.03ms      3.96±0.05ms     1.17  stat_ops.SeriesMultiIndexOps.time_op(0, 'var')
+        997±40ns      1.16±0.04μs     1.17  index_cached_properties.IndexCache.time_is_monotonic_increasing('Int64Index')
+     3.04±0.02ms      3.55±0.04ms     1.17  stat_ops.FrameMultiIndexOps.time_op(0, 'prod')
+     2.94±0.02ms      3.42±0.05ms     1.17  groupby.TransformBools.time_transform_mean
+      3.88±0.1ms      4.52±0.02ms     1.16  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 0, 'higher')
+     2.22±0.06ms      2.58±0.02ms     1.16  frame_methods.Lookup.time_frame_fancy_lookup
+     11.3±0.04ms       13.1±0.3ms     1.16  index_object.IntervalIndexMethod.time_intersection_one_duplicate(100000)
+        1.41±0ms         1.63±0ms     1.16  timeseries.DatetimeIndex.time_normalize('tz_naive')
+      81.8±0.2ms       95.0±0.2ms     1.16  series_methods.IsInFloat64.time_isin_nan_values
+      3.90±0.2ms      4.53±0.03ms     1.16  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 1, 'lower')
+      3.90±0.2ms      4.53±0.02ms     1.16  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 1, 'nearest')
+        90.6±1μs          105±1μs     1.16  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.Int64Index'>, 'nonunique_monotonic_inc')
+      41.0±0.3ms       47.6±0.1ms     1.16  rolling.Apply.time_rolling('Series', 300, 'float', <built-in function sum>, False)
+      3.91±0.2ms      4.53±0.02ms     1.16  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 1, 'linear')
+      95.6±0.2ms        111±0.9ms     1.16  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 1000, 'random_misses')
+         662±3μs          768±2μs     1.16  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 7000)
+        90.6±1μs          105±2μs     1.16  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+     2.31±0.02ms      2.67±0.03ms     1.16  stat_ops.FrameOps.time_op('std', 'int', 0)
+      78.7±0.4ms         91.1±1ms     1.16  period.PeriodIndexConstructor.time_from_ints_daily('D', True)
+      3.91±0.1ms      4.53±0.01ms     1.16  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 1, 'midpoint')
+         366±2μs          424±4μs     1.16  groupby.GroupByMethods.time_dtype_as_field('datetime', 'shift', 'direct')
+         129±1ms          149±1ms     1.16  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 900000)
+         894±7ns      1.03±0.02μs     1.16  tslibs.normalize.Normalize.time_is_date_array_normalized(0, None)
+         440±5μs          509±3μs     1.16  hash_functions.IsinWithArangeSorted.time_isin(<class 'object'>, 8000)
+     3.75±0.04ms      4.33±0.04ms     1.16  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 70000)
+      91.0±0.4μs          105±2μs     1.15  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+      10.7±0.2ms       12.4±0.4ms     1.15  stat_ops.SeriesMultiIndexOps.time_op([0, 1], 'median')
+      40.5±0.8ms         46.7±1ms     1.15  stat_ops.FrameMultiIndexOps.time_op(0, 'mad')
+     25.4±0.08ms      29.3±0.09ms     1.15  hash_functions.IsinWithArange.time_isin(<class 'numpy.int64'>, 1000, 2)
+        91.4±1μs        105±0.9μs     1.15  indexing.NumericSeriesIndexing.time_iloc_array(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'unique_monotonic_inc')
+     1.42±0.01ms      1.63±0.01ms     1.15  timeseries.DatetimeIndex.time_normalize('repeated')
+         177±1μs        204±0.9μs     1.15  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'object'>, 11)
+     15.8±0.03ms       18.2±0.3ms     1.15  eval.Query.time_query_datetime_index
+     10.1±0.04μs      11.6±0.03μs     1.15  dtypes.Dtypes.time_pandas_dtype('UInt16')
+      41.6±0.2ms       47.9±0.3ms     1.15  rolling.Apply.time_rolling('DataFrame', 300, 'int', <built-in function sum>, False)
+      6.71±0.3μs       7.72±0.3μs     1.15  index_cached_properties.IndexCache.time_shape('IntervalIndex')
+      79.0±0.2ms       90.9±0.9ms     1.15  period.PeriodIndexConstructor.time_from_ints_daily('D', False)
+         532±9ns         612±10ns     1.15  multiindex_object.Integer.time_is_monotonic
+         141±3μs        162±0.4μs     1.15  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('datetime', 'nonunique_monotonic_inc')
+       481±0.9ms          554±4ms     1.15  frame_methods.Iteration.time_iterrows
+      27.4±0.3μs       31.5±0.8μs     1.15  indexing.CategoricalIndexIndexing.time_getitem_slice('monotonic_decr')
+     3.72±0.01ms      4.28±0.01ms     1.15  series_methods.IsInDatetime64.time_isin
+        1.74±0ms      2.00±0.01ms     1.15  timeseries.DatetimeAccessor.time_dt_accessor_normalize(tzutc())
+         453±3μs          520±2μs     1.15  reindex.DropDuplicates.time_series_drop_dups_int(False)
+      31.8±0.6ms       36.5±0.8ms     1.15  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 19)
+     2.94±0.04ms      3.38±0.02ms     1.15  stat_ops.FrameMultiIndexOps.time_op(1, 'sum')
+         409±2μs          469±3μs     1.15  reindex.DropDuplicates.time_series_drop_dups_string(True)
+     5.21±0.03ms       5.97±0.3ms     1.15  stat_ops.SeriesMultiIndexOps.time_op(1, 'median')
+      3.94±0.2ms      4.52±0.02ms     1.15  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 0, 'linear')
+        895±10ns      1.03±0.01μs     1.15  tslibs.normalize.Normalize.time_is_date_array_normalized(1, None)
+      27.6±0.3μs       31.6±0.5μs     1.15  indexing.CategoricalIndexIndexing.time_getitem_slice('monotonic_incr')
+      47.7±0.4ms       54.7±0.7ms     1.15  gil.ParallelGroupbyMethods.time_parallel(4, 'max')
+      73.3±0.4μs       84.0±0.9μs     1.15  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 10000)
+         154±1μs          177±2μs     1.15  groupby.GroupByMethods.time_dtype_as_group('int', 'size', 'direct')
+       258±0.4μs          295±1μs     1.14  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 2000)
+         369±2μs          422±4μs     1.14  groupby.GroupByMethods.time_dtype_as_field('datetime', 'shift', 'transformation')
+        1.72±0ms      1.96±0.02ms     1.14  timeseries.DatetimeAccessor.time_dt_accessor_normalize(None)
+     1.78±0.02ms      2.04±0.01ms     1.14  series_methods.IsIn.time_isin('int64')
+     3.42±0.03ms       3.91±0.1ms     1.14  stat_ops.SeriesMultiIndexOps.time_op(0, 'std')
+      74.6±0.7μs       85.2±0.4μs     1.14  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 10000)
+      41.7±0.2ms       47.6±0.2ms     1.14  rolling.Apply.time_rolling('DataFrame', 300, 'float', <built-in function sum>, False)
+      54.2±0.4μs       61.8±0.4μs     1.14  inference.ToNumeric.time_from_float('coerce')
+      15.2±0.2ms       17.3±0.2ms     1.14  eval.Query.time_query_datetime_column
+         279±2μs        318±0.8μs     1.14  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 2000)
+      54.2±0.9μs       61.8±0.5μs     1.14  inference.ToNumeric.time_from_float('ignore')
+     3.23±0.02ms      3.69±0.02ms     1.14  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 70000)
+     2.16±0.01μs      2.46±0.06μs     1.14  attrs_caching.SeriesArrayAttribute.time_array('object')
+     1.74±0.01ms      1.98±0.01ms     1.14  timeseries.DatetimeAccessor.time_dt_accessor_normalize('UTC')
+     2.16±0.02μs      2.45±0.05μs     1.14  attrs_caching.SeriesArrayAttribute.time_array('numeric')
+         441±3μs          501±4μs     1.14  strings.Encode.time_encode_decode
+     2.70±0.01ms      3.07±0.01ms     1.14  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 16)
+      26.7±0.2ms       30.3±0.3ms     1.14  groupby.Groups.time_series_groups('int64_small')
+      63.0±0.2μs       71.5±0.3μs     1.13  tslibs.normalize.Normalize.time_normalize_i8_timestamps(10000, datetime.timezone(datetime.timedelta(seconds=3600)))
+         252±2μs          286±3μs     1.13  arithmetic.NumericInferOps.time_subtract(<class 'numpy.int8'>)
+         254±2μs          288±2μs     1.13  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'object'>, 12)
+       214±0.7ms        242±0.4ms     1.13  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 1000, 'monotone_hits')
+        74.6±1ms       84.6±0.6ms     1.13  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 750000)
+      27.6±0.3μs       31.3±0.5μs     1.13  indexing.CategoricalIndexIndexing.time_getitem_slice('non_monotonic')
+         323±2μs          366±2μs     1.13  index_object.IntervalIndexMethod.time_intersection(1000)
+         152±1μs          173±1μs     1.13  groupby.GroupByMethods.time_dtype_as_field('datetime', 'size', 'transformation')
+     27.4±0.03ms      31.0±0.06ms     1.13  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 8000, 2)
+      14.3±0.2μs       16.2±0.1μs     1.13  indexing.NonNumericSeriesIndexing.time_getitem_scalar('datetime', 'non_monotonic')
+      65.5±0.7μs       74.1±0.3μs     1.13  indexing.NumericSeriesIndexing.time_getitem_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     79.4±0.09ms       89.9±0.2ms     1.13  hash_functions.IsinWithArange.time_isin(<class 'object'>, 1000, -2)
+      4.01±0.2ms      4.53±0.02ms     1.13  rolling.Quantile.time_quantile('DataFrame', 1000, 'float', 1, 'higher')
+       127±0.1ms        144±0.3ms     1.13  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 5, 'monotone_hits')
+         915±5μs      1.03±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_string_include('Int8')
+      21.6±0.4μs      24.4±0.08μs     1.13  indexing.NonNumericSeriesIndexing.time_getitem_scalar('datetime', 'nonunique_monotonic_inc')
+      64.5±0.4μs       72.9±0.4μs     1.13  indexing.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
+     4.15±0.04ms      4.69±0.06ms     1.13  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 80000)
+         907±4μs      1.02±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_string_include(<class 'int'>)
+         534±1μs          603±7μs     1.13  frame_methods.Iteration.time_itertuples_raw_start
+       156±0.9ms          176±1ms     1.13  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 5, 'random_hits')
+        81.4±1ms       92.0±0.7ms     1.13  arithmetic.BinaryOpsMultiIndex.time_binary_op_multiindex('add')
+         535±5μs          604±6μs     1.13  frame_methods.Iteration.time_itertuples_raw_read_first
+         943±7μs      1.06±0.02ms     1.13  dtypes.SelectDtypes.time_select_dtype_string_include('m8[ns]')
+         154±1μs        174±0.8μs     1.13  groupby.GroupByMethods.time_dtype_as_group('int', 'size', 'transformation')
+       152±0.8μs          172±1μs     1.13  groupby.GroupByMethods.time_dtype_as_group('object', 'size', 'transformation')
+      27.8±0.2ms      31.3±0.05ms     1.13  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 2000, -2)
+         121±2μs          136±2μs     1.13  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 10)
+         939±4μs      1.06±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_int_include('int8')
+         153±1μs          172±2μs     1.13  groupby.GroupByMethods.time_dtype_as_field('datetime', 'size', 'direct')
+     11.0±0.06μs      12.4±0.07μs     1.13  dtypes.Dtypes.time_pandas_dtype('UInt32')
+     1.00±0.01ms      1.13±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_int_exclude('Int64')
+         920±4μs      1.04±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_string_include('Int64')
+         924±3μs      1.04±0.01ms     1.13  dtypes.SelectDtypes.time_select_dtype_int_include('Int32')
+        946±10μs      1.06±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include('float32')
+         152±1μs          171±2μs     1.12  groupby.GroupByMethods.time_dtype_as_field('int', 'size', 'transformation')
+         987±3μs         1.11±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_exclude(<class 'int'>)
+       205±0.6μs        230±0.9μs     1.12  tslibs.normalize.Normalize.time_is_date_array_normalized(10000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+         938±5μs      1.05±0.02ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('bool')
+      79.2±0.4μs       89.0±0.6μs     1.12  hash_functions.NumericSeriesIndexing.time_loc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 10000)
+     1.00±0.01ms      1.13±0.01ms     1.12  frame_methods.SelectDtypes.time_select_dtypes(1000)
+         938±8μs         1.05±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('int32')
+         903±3μs         1.01±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include(<class 'bool'>)
+         944±4μs      1.06±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include('uint64')
+         936±5μs         1.05±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include('complex64')
+         938±5μs      1.05±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('uint16')
+      44.9±0.5μs       50.4±0.1μs     1.12  arithmetic.Ops2.time_series_dot
+         935±1μs      1.05±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('M8[ns]')
+         917±2μs      1.03±0.02ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include('Int16')
+         937±5μs      1.05±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('float64')
+       158±0.9μs        177±0.9μs     1.12  groupby.GroupByMethods.time_dtype_as_group('float', 'size', 'transformation')
+     14.8±0.08μs       16.6±0.1μs     1.12  dtypes.Dtypes.time_pandas_dtype('float32')
+     3.71±0.02ms      4.16±0.04ms     1.12  hash_functions.IsinWithArangeSorted.time_isin(<class 'object'>, 100000)
+         152±2μs        171±0.9μs     1.12  groupby.GroupByMethods.time_dtype_as_group('object', 'size', 'direct')
+         226±7ms          254±5ms     1.12  indexing.NumericSeriesIndexing.time_loc_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+         941±1μs         1.05±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include('uint8')
+     11.7±0.06μs      13.1±0.03μs     1.12  dtypes.Dtypes.time_pandas_dtype('UInt64')
+         930±6μs      1.04±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include('UInt16')
+         922±4μs         1.03±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('UInt8')
+         152±1μs        171±0.9μs     1.12  groupby.GroupByMethods.time_dtype_as_field('object', 'size', 'transformation')
+         913±6μs      1.02±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include(<class 'bool'>)
+     3.76±0.05ms      4.21±0.05ms     1.12  stat_ops.FrameMultiIndexOps.time_op(1, 'var')
+     5.85±0.02ms      6.55±0.01ms     1.12  frame_methods.Repr.time_html_repr_trunc_si
+         923±6μs      1.03±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_include('UInt16')
+         944±9μs      1.06±0.02ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('timedelta64[ns]')
+     1.61±0.02ms      1.81±0.02ms     1.12  reindex.ReindexMethod.time_reindex_method('backfill', <function period_range at 0x7f2782af4700>)
+         927±2μs      1.04±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include('UInt32')
+         880±8μs          984±9μs     1.12  groupby.GroupByMethods.time_dtype_as_group('float', 'sem', 'direct')
+      24.3±0.4ms       27.2±0.8ms     1.12  timeseries.ToDatetimeFormat.time_no_exact
+         939±4μs      1.05±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('uint16')
+      27.4±0.1ms      30.7±0.07ms     1.12  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 2000, 2)
+       937±0.6μs         1.05±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include('m8[ns]')
+         112±3ms        125±0.9ms     1.12  io.json.ToJSON.time_to_json('split', 'df_date_idx')
+         937±2μs      1.05±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('float64')
+         109±1ms          122±2ms     1.12  io.json.ToJSON.time_to_json('records', 'df')
+     3.26±0.01ms      3.65±0.02ms     1.12  rolling.ExpandingMethods.time_expanding('DataFrame', 'int', 'skew')
+     25.6±0.06ms       28.6±0.2ms     1.12  groupby.Apply.time_scalar_function_multi_col
+        90.7±2ms          101±3ms     1.12  gil.ParallelGroupbyMethods.time_parallel(8, 'min')
+         905±3μs      1.01±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include(<class 'float'>)
+        1.02±0ms         1.14±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_int_exclude('int64')
+         943±4μs         1.05±0ms     1.12  dtypes.SelectDtypes.time_select_dtype_float_include('uint32')
+     4.68±0.04ms      5.22±0.07ms     1.12  stat_ops.SeriesMultiIndexOps.time_op(1, 'sem')
+         952±6μs      1.06±0.02ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('timedelta64[ns]')
+         932±4μs      1.04±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('UInt32')
+         937±3μs      1.04±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_string_include('uint8')
+         907±3μs      1.01±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include(<class 'int'>)
+         922±1μs      1.03±0.01ms     1.12  dtypes.SelectDtypes.time_select_dtype_bool_include('Int8')
+      87.0±0.7μs       97.0±0.4μs     1.11  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('datetime', 'unique_monotonic_inc')
+         938±2μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('int8')
+     3.74±0.05ms      4.17±0.02ms     1.11  stat_ops.FrameMultiIndexOps.time_op(0, 'var')
+         948±3μs         1.06±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_int_include('complex128')
+     2.47±0.01ms      2.76±0.09ms     1.11  io.csv.ReadCSVFloatPrecision.time_read_csv(',', '_', None)
+     14.8±0.05μs       16.5±0.2μs     1.11  dtypes.Dtypes.time_pandas_dtype('int8')
+         991±4μs      1.10±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_exclude(<class 'bool'>)
+         945±8μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('complex64')
+         945±7μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('datetime64[ns]')
+         911±5μs         1.01±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include(<class 'bool'>)
+         944±9μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('float32')
+         846±6μs          941±3μs     1.11  groupby.GroupByMethods.time_dtype_as_field('float', 'sem', 'direct')
+      46.8±0.6ms       52.0±0.2ms     1.11  gil.ParallelGroupbyMethods.time_parallel(4, 'sum')
+         945±7μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('timedelta64[ns]')
+        82.5±3ms         91.7±1ms     1.11  arithmetic.BinaryOpsMultiIndex.time_binary_op_multiindex('div')
+         943±5μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('int16')
+        992±10μs         1.10±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_exclude(<class 'float'>)
+     1.95±0.02ms      2.17±0.01ms     1.11  groupby.Datelike.time_sum('date_range')
+         929±3μs         1.03±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('UInt64')
+     2.92±0.02ms      3.24±0.02ms     1.11  reindex.DropDuplicates.time_frame_drop_dups_bool(False)
+         940±7μs      1.04±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('datetime64[ns]')
+        943±10μs         1.05±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_int_include('int16')
+         942±3μs         1.05±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('uint8')
+         931±5μs      1.03±0.02ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('UInt32')
+         909±5μs      1.01±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include(<class 'float'>)
+         525±3μs          583±5μs     1.11  groupby.GroupByMethods.time_dtype_as_field('datetime', 'ffill', 'direct')
+     3.44±0.02ms      3.82±0.09ms     1.11  stat_ops.SeriesMultiIndexOps.time_op(1, 'std')
+      10.1±0.4ms       11.3±0.2ms     1.11  stat_ops.SeriesMultiIndexOps.time_op([0, 1], 'sem')
+      68.4±0.3ms       76.0±0.5ms     1.11  groupby.GroupByMethods.time_dtype_as_field('int', 'unique', 'direct')
+         909±9μs         1.01±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include(<class 'int'>)
+         938±6μs      1.04±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_bool_include('int64')
+         926±6μs      1.03±0.02ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('Int8')
+     1.09±0.04μs      1.21±0.08μs     1.11  index_cached_properties.IndexCache.time_is_monotonic_increasing('RangeIndex')
+         945±6μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('uint64')
+         943±2μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('uint64')
+         945±8μs      1.05±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('int16')
+      76.7±0.2μs       85.1±0.3μs     1.11  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 10000)
+         252±4μs          280±2μs     1.11  arithmetic.NumericInferOps.time_subtract(<class 'numpy.uint8'>)
+     3.58±0.04ms      3.97±0.05ms     1.11  io.csv.ReadCSVFloatPrecision.time_read_csv(',', '.', 'round_trip')
+         909±4μs         1.01±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include(<class 'complex'>)
+     4.35±0.02ms      4.83±0.03ms     1.11  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'skew')
+        948±10μs         1.05±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('complex128')
+        1.02±0ms      1.14±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_exclude('float64')
+      69.2±0.7μs         76.7±1μs     1.11  ctors.SeriesConstructors.time_series_constructor(<function no_change at 0x7f277fe8ac10>, False, 'float')
+       677±0.6μs          751±3μs     1.11  groupby.GroupByMethods.time_dtype_as_group('object', 'unique', 'transformation')
+       932±0.9μs         1.03±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('UInt16')
+         868±3μs          963±9μs     1.11  groupby.GroupByMethods.time_dtype_as_group('int', 'sem', 'transformation')
+         850±5μs          943±9μs     1.11  groupby.GroupByMethods.time_dtype_as_field('int', 'sem', 'direct')
+      86.0±0.3μs         95.3±1μs     1.11  frame_methods.XS.time_frame_xs(0)
+     3.73±0.02ms      4.13±0.04ms     1.11  series_methods.IsInForObjects.time_isin_long_series_long_values
+     14.9±0.06μs       16.5±0.1μs     1.11  dtypes.Dtypes.time_pandas_dtype('int64')
+         926±6μs         1.03±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('Int64')
+         880±7μs          975±5μs     1.11  groupby.GroupByMethods.time_dtype_as_group('float', 'sem', 'transformation')
+         526±5μs          583±5μs     1.11  groupby.GroupByMethods.time_dtype_as_field('datetime', 'bfill', 'transformation')
+      6.01±0.1μs       6.66±0.1μs     1.11  index_object.Indexing.time_get_loc('Int')
+         926±7μs         1.02±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('Int16')
+         926±3μs      1.02±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_int_include('Int8')
+     14.9±0.07μs       16.5±0.2μs     1.11  dtypes.Dtypes.time_pandas_dtype('int16')
+         952±6μs      1.05±0.03ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('uint16')
+         909±3μs         1.01±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_int_include(<class 'complex'>)
+        46.2±1ms       51.1±0.4ms     1.11  gil.ParallelGroupbyMethods.time_parallel(4, 'last')
+        87.1±1μs         96.4±2μs     1.11  ctors.SeriesConstructors.time_series_constructor(<function no_change at 0x7f277fe8ac10>, True, 'float')
+     14.9±0.05μs      16.5±0.04μs     1.11  dtypes.Dtypes.time_pandas_dtype('float64')
+         939±6μs         1.04±0ms     1.11  dtypes.SelectDtypes.time_select_dtype_int_include('uint32')
+         290±2μs          321±2μs     1.11  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'object'>, 12)
+         109±2ms          121±1ms     1.11  io.json.ToJSON.time_to_json('records', 'df_date_idx')
+      25.8±0.1ms      28.5±0.09ms     1.11  hash_functions.IsinWithArange.time_isin(<class 'object'>, 1000, 0)
+     14.9±0.07μs      16.5±0.09μs     1.11  dtypes.Dtypes.time_pandas_dtype('uint32')
+         738±2ms          817±2ms     1.11  groupby.GroupByMethods.time_dtype_as_group('int', 'mad', 'direct')
+      87.9±0.5μs       97.2±0.6μs     1.11  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('datetime', 'nonunique_monotonic_inc')
+     14.9±0.06μs      16.5±0.06μs     1.11  dtypes.Dtypes.time_pandas_dtype('uint8')
+         932±2μs      1.03±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_string_include('Int32')
+         167±1μs          185±2μs     1.11  groupby.GroupByMethods.time_dtype_as_field('datetime', 'count', 'transformation')
+     9.24±0.07ms      10.2±0.04ms     1.11  frame_methods.Repr.time_html_repr_trunc_mi
+       191±0.5ms          211±2ms     1.11  io.style.RenderApply.time_render(24, 120)
+         159±1μs          176±1μs     1.11  groupby.GroupByMethods.time_dtype_as_group('float', 'size', 'direct')
+        945±10μs      1.04±0.01ms     1.11  dtypes.SelectDtypes.time_select_dtype_float_include('bool')
+     15.0±0.03μs      16.6±0.04μs     1.11  dtypes.Dtypes.time_pandas_dtype('datetime64')
+        936±10μs      1.03±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_float_include('UInt8')
+         940±6μs      1.04±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_string_include('M8[ns]')
+         867±3μs          958±4μs     1.10  groupby.GroupByMethods.time_dtype_as_group('int', 'sem', 'direct')
+         932±7μs      1.03±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_bool_include('UInt64')
+         944±5μs      1.04±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('M8[ns]')
+         228±5ms          252±7ms     1.10  indexing.NumericSeriesIndexing.time_getitem_array(<class 'pandas.core.indexes.numeric.Float64Index'>, 'nonunique_monotonic_inc')
+       155±0.8μs        171±0.7μs     1.10  groupby.GroupByMethods.time_dtype_as_field('object', 'size', 'direct')
+         1.15±0s          1.27±0s     1.10  groupby.GroupByMethods.time_dtype_as_group('float', 'mad', 'transformation')
+       150±0.3μs          166±1μs     1.10  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.int64'>, 2000)
+        852±10μs          940±4μs     1.10  groupby.GroupByMethods.time_dtype_as_field('float', 'sem', 'transformation')
+         933±9μs         1.03±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_bool_include('UInt16')
+         954±3μs         1.05±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_bool_include('m8[ns]')
+     1.03±0.01ms         1.13±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_bool_exclude('bool')
+         128±2μs          141±2μs     1.10  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 10)
+     2.92±0.03ms       3.22±0.1ms     1.10  stat_ops.SeriesMultiIndexOps.time_op(1, 'mean')
+         928±2μs         1.02±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_float_include('Int16')
+         412±3μs          454±2μs     1.10  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 14)
+     1.20±0.02μs      1.32±0.01μs     1.10  tslibs.normalize.Normalize.time_is_date_array_normalized(100, datetime.timezone.utc)
+     3.16±0.08μs       3.48±0.1μs     1.10  attrs_caching.SeriesArrayAttribute.time_extract_array('object')
+         265±2μs          292±2μs     1.10  arithmetic.NumericInferOps.time_multiply(<class 'numpy.int8'>)
+         947±4μs         1.04±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_string_include('uint32')
+      26.4±0.2ms      29.1±0.05ms     1.10  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 1000, 2)
+      88.3±0.6μs       97.4±0.5μs     1.10  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('datetime', 'non_monotonic')
+         720±4μs          794±4μs     1.10  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 15)
+         259±3μs          285±2μs     1.10  arithmetic.NumericInferOps.time_add(<class 'numpy.int8'>)
+      1.87±0.1μs      2.06±0.09μs     1.10  index_cached_properties.IndexCache.time_inferred_type('PeriodIndex')
+         191±1μs        211±0.7μs     1.10  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'numpy.float64'>, 1300)
+         100±2ms          110±1ms     1.10  io.json.ToJSON.time_to_json('values', 'df_date_idx')
+        952±10μs      1.05±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('datetime64[ns]')
+     2.46±0.01ms      2.71±0.01ms     1.10  reindex.DropDuplicates.time_frame_drop_dups_bool(True)
+        952±10μs      1.05±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('bool')
+         948±8μs         1.04±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('m8[ns]')
+         948±9μs      1.04±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_float_include('int8')
+         528±8μs          581±2μs     1.10  groupby.GroupByMethods.time_dtype_as_field('datetime', 'ffill', 'transformation')
+      14.9±0.1μs      16.4±0.04μs     1.10  dtypes.Dtypes.time_pandas_dtype('int32')
+     4.33±0.01ms      4.76±0.01ms     1.10  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'skew')
+         942±8μs      1.04±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('uint16')
+       503±0.4ms          554±1ms     1.10  groupby.GroupByMethods.time_dtype_as_field('int', 'mad', 'direct')
+       503±0.9ms        553±0.9ms     1.10  groupby.GroupByMethods.time_dtype_as_field('int', 'mad', 'transformation')
+         918±3μs      1.01±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_string_include(<class 'complex'>)
+         927±4μs         1.02±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_bool_include('Int64')
+      20.5±0.2μs       22.5±0.2μs     1.10  indexing.NonNumericSeriesIndexing.time_getitem_scalar('datetime', 'unique_monotonic_inc')
+        832±10μs         915±20μs     1.10  series_methods.Clip.time_clip(50)
+         279±1ms        307±0.8ms     1.10  io.style.RenderApply.time_render(36, 120)
+       101±0.2ms          111±1ms     1.10  groupby.GroupByMethods.time_dtype_as_group('int', 'unique', 'direct')
+       104±0.7ms        115±0.8ms     1.10  hash_functions.IsinWithArange.time_isin(<class 'object'>, 8000, -2)
+       171±0.9μs          188±2μs     1.10  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 100000)
+      1.16±0.01s          1.27±0s     1.10  groupby.GroupByMethods.time_dtype_as_group('float', 'mad', 'direct')
+         929±4μs         1.02±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_string_include('UInt8')
+       264±0.9μs          290±1μs     1.10  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 13)
+         936±9μs      1.03±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('UInt8')
+         257±2ms        282±0.9ms     1.10  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 1000, 'random_hits')
+     3.59±0.03ms      3.94±0.06ms     1.10  io.csv.ReadCSVFloatPrecision.time_read_csv(';', '.', 'round_trip')
+         943±6μs         1.03±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_float_include('uint8')
+         296±2μs          324±2μs     1.10  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.int64'>, 8000)
+       741±0.7ms          812±1ms     1.10  groupby.GroupByMethods.time_dtype_as_group('int', 'mad', 'transformation')
+     2.46±0.02ms      2.69±0.03ms     1.10  io.csv.ReadCSVFloatPrecision.time_read_csv(',', '_', 'high')
+        954±10μs         1.05±0ms     1.10  dtypes.SelectDtypes.time_select_dtype_string_include('complex64')
+     14.8±0.04μs      16.3±0.04μs     1.10  dtypes.Dtypes.time_pandas_dtype('object')
+         936±3μs      1.03±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_int_include('UInt64')
+     14.9±0.03μs      16.3±0.07μs     1.10  dtypes.DtypesInvalid.time_pandas_dtype_invalid('scalar-string')
+         130±1ms          142±2ms     1.10  io.json.ToJSON.time_to_json('index', 'df')
+      68.9±0.4ms       75.4±0.5ms     1.10  groupby.GroupByMethods.time_dtype_as_field('int', 'unique', 'transformation')
+         138±2ms          151±2ms     1.10  io.json.ToJSON.time_to_json('index', 'df_date_idx')
+         947±3μs      1.04±0.01ms     1.10  dtypes.SelectDtypes.time_select_dtype_float_include('M8[ns]')
+         448±3μs          491±3μs     1.09  period.Algorithms.time_drop_duplicates('series')
+         949±7μs         1.04±0ms     1.09  dtypes.SelectDtypes.time_select_dtype_bool_include('uint32')
+     15.0±0.05μs      16.4±0.08μs     1.09  dtypes.Dtypes.time_pandas_dtype('uint64')
+         857±2μs          937±6μs     1.09  groupby.GroupByMethods.time_dtype_as_field('int', 'sem', 'transformation')
+         414±4μs          453±3μs     1.09  index_object.SetOperations.time_operation('datetime', 'intersection')
+         508±3μs          555±5μs     1.09  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 14)
+         201±5ms        220±0.6ms     1.09  io.stata.StataMissing.time_read_stata('td')
+        934±10μs      1.02±0.01ms     1.09  dtypes.SelectDtypes.time_select_dtype_bool_include('Int32')
+         108±2ms          118±2ms     1.09  io.stata.Stata.time_read_stata('tq')
+        959±20μs         1.05±0ms     1.09  dtypes.SelectDtypes.time_select_dtype_int_include('timedelta64[ns]')
+     1.44±0.01ms      1.58±0.02ms     1.09  dtypes.SelectDtypes.time_select_dtype_int_include(<class 'int'>)
+         205±5ms          224±6ms     1.09  io.stata.StataMissing.time_read_stata('tq')
+     15.4±0.07μs      16.8±0.02μs     1.09  dtypes.Dtypes.time_pandas_dtype('timedelta64')
+       101±0.4ms        111±0.7ms     1.09  groupby.GroupByMethods.time_dtype_as_group('int', 'unique', 'transformation')
+         514±2ms        561±0.7ms     1.09  groupby.GroupByMethods.time_dtype_as_field('float', 'mad', 'transformation')
+         148±1μs        162±0.7μs     1.09  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 11)
+     1.55±0.01ms      1.69±0.02ms     1.09  dtypes.SelectDtypes.time_select_dtype_float_exclude('M8[ns]')
+       161±0.5ms        176±0.3ms     1.09  groupby.GroupByMethods.time_dtype_as_group('datetime', 'unique', 'transformation')
+      52.9±0.8ms       57.7±0.9ms     1.09  gil.ParallelGroupbyMethods.time_parallel(4, 'var')
+     1.46±0.01ms      1.60±0.01ms     1.09  dtypes.SelectDtypes.time_select_dtype_int_include('Int64')
+      94.7±0.7ms        103±0.7ms     1.09  frame_methods.Repr.time_frame_repr_wide
+         168±5ms          183±2ms     1.09  io.json.ToJSONLines.time_floats_with_int_idex_lines
+      87.6±0.7μs       95.4±0.8μs     1.09  timeseries.SortIndex.time_get_slice(True)
+     3.17±0.08μs      3.45±0.08μs     1.09  attrs_caching.SeriesArrayAttribute.time_extract_array('numeric')
+         233±2ms          254±1ms     1.09  io.json.ToJSONLines.time_float_longint_str_lines
+         855±5μs          932±2μs     1.09  series_methods.Clip.time_clip(1000)
+         185±2ms        201±0.7ms     1.09  io.json.ToJSONLines.time_delta_int_tstamp_lines
+        920±10μs         1.00±0ms     1.09  dtypes.SelectDtypes.time_select_dtype_bool_include(<class 'complex'>)
+        1.49±0ms      1.62±0.02ms     1.09  frame_methods.Iteration.time_items_cached
+       158±0.7μs        172±0.5μs     1.09  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 11)
+         951±8μs         1.03±0ms     1.09  dtypes.SelectDtypes.time_select_dtype_bool_include('datetime64[ns]')
+     1.54±0.01ms      1.68±0.01ms     1.09  dtypes.SelectDtypes.time_select_dtype_float_exclude('int32')
+     1.55±0.02ms      1.68±0.02ms     1.09  dtypes.SelectDtypes.time_select_dtype_float_exclude('int8')
+     4.63±0.01ms      5.03±0.02ms     1.09  stat_ops.FrameOps.time_op('median', 'int', 0)
+     22.0±0.05ms      23.9±0.07ms     1.09  groupby.MultiColumn.time_cython_sum
+         417±6μs          453±1μs     1.09  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'object'>, 13)
+      99.3±0.5ms        108±0.4ms     1.09  join_merge.MergeOrdered.time_merge_ordered
+         896±5μs          973±8μs     1.09  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 15)
+       168±0.7μs          183±2μs     1.09  groupby.GroupByMethods.time_dtype_as_group('datetime', 'size', 'direct')
+        1.55±0ms      1.68±0.01ms     1.09  dtypes.SelectDtypes.time_select_dtype_float_exclude('timedelta64[ns]')
+         126±2μs          137±2μs     1.09  timeseries.AsOf.time_asof_single_early('DataFrame')
+     4.29±0.02ms      4.66±0.01ms     1.09  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'skew')
+       161±0.4ms        175±0.2ms     1.09  groupby.GroupByMethods.time_dtype_as_group('datetime', 'unique', 'direct')
+      24.9±0.3ms       27.0±0.3ms     1.09  gil.ParallelGroupbyMethods.time_parallel(2, 'prod')
+     11.9±0.07ms      12.9±0.09ms     1.08  stat_ops.FrameMultiIndexOps.time_op(1, 'sem')
+         210±2ms        228±0.5ms     1.08  io.json.ToJSONLines.time_float_int_str_lines
+     15.1±0.08μs      16.3±0.04μs     1.08  dtypes.Dtypes.time_pandas_dtype('uint16')
+      24.6±0.8ms       26.7±0.3ms     1.08  gil.ParallelGroupbyMethods.time_parallel(2, 'sum')
+       155±0.4ms          168±2ms     1.08  frame_methods.Duplicated.time_frame_duplicated_wide
+     1.57±0.01ms      1.70±0.02ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude('bool')
+         222±2μs        241±0.4μs     1.08  arithmetic.OffsetArrayArithmetic.time_add_series_offset(<Day>)
+     12.1±0.09ms      13.2±0.09ms     1.08  multiindex_object.Integer.time_get_indexer
+         212±2μs          230±1μs     1.08  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 12)
+     1.54±0.01ms      1.67±0.03ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('Int64')
+     3.96±0.02ms      4.29±0.08ms     1.08  index_cached_properties.IndexCache.time_is_unique('Float64Index')
+        699±40ns         757±30ns     1.08  index_cached_properties.IndexCache.time_inferred_type('RangeIndex')
+     2.38±0.02ms      2.58±0.02ms     1.08  groupby.TransformNaN.time_first
+     12.0±0.07ms      13.0±0.07ms     1.08  stat_ops.FrameMultiIndexOps.time_op(0, 'sem')
+      98.0±0.8μs          106±3μs     1.08  ctors.SeriesConstructors.time_series_constructor(<function no_change at 0x7f277fe8ac10>, True, 'int')
+     1.38±0.01ms      1.49±0.01ms     1.08  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.int64'>, 16)
+        1.56±0ms      1.69±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude('uint16')
+         215±1ms          233±3ms     1.08  io.json.ToJSONLines.time_float_int_lines
+         630±5μs          682±5μs     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'nunique', 'transformation')
+         131±2μs          142±2μs     1.08  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.int64'>, 1000)
+         360±2μs          390±7μs     1.08  arithmetic.NumericInferOps.time_subtract(<class 'numpy.int16'>)
+      27.2±0.1ms       29.5±0.3ms     1.08  gil.ParallelGroupbyMethods.time_parallel(2, 'var')
+         361±6μs          390±6μs     1.08  arithmetic.NumericInferOps.time_subtract(<class 'numpy.uint16'>)
+      11.6±0.2ms      12.5±0.06ms     1.08  reshape.PivotTable.time_pivot_table_categorical_observed
+      22.3±0.2μs      24.1±0.09μs     1.08  boolean.TimeLogicalOps.time_and_array
+         366±2μs          396±4μs     1.08  arithmetic.NumericInferOps.time_add(<class 'numpy.int16'>)
+      24.4±0.4ms       26.4±0.5ms     1.08  gil.ParallelGroupbyMethods.time_parallel(2, 'mean')
+         364±3μs          394±4μs     1.08  arithmetic.NumericInferOps.time_multiply(<class 'numpy.uint16'>)
+     7.25±0.01ms      7.84±0.03ms     1.08  timeseries.ResampleSeries.time_resample('period', '5min', 'mean')
+         880±3ms          952±2ms     1.08  join_merge.I8Merge.time_i8merge('right')
+     1.56±0.01ms      1.68±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('int16')
+       119±0.3ms        129±0.2ms     1.08  strings.Repeat.time_repeat('array')
+        1.55±0ms      1.68±0.02ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude('Int8')
+      73.0±0.2ms       78.9±0.6ms     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'unique', 'direct')
+         634±5μs          685±3μs     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'nunique', 'direct')
+         518±3ms          560±2ms     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'mad', 'direct')
+     2.47±0.02ms      2.67±0.05ms     1.08  io.csv.ReadCSVFloatPrecision.time_read_csv(';', '_', 'high')
+     1.80±0.03ms      1.95±0.03ms     1.08  groupby.Datelike.time_sum('period_range')
+        953±10μs         1.03±0ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_include('float64')
+     1.53±0.02ms      1.65±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('Int8')
+       142±0.5μs        154±0.7μs     1.08  tslibs.normalize.Normalize.time_normalize_i8_timestamps(10000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+      15.9±0.2ms       17.2±0.1ms     1.08  io.csv.ReadCSVSkipRows.time_skipprows(10000, 'c')
+         169±2μs          183±3μs     1.08  hash_functions.IsinWithRandomFloat.time_isin(<class 'numpy.float64'>, 1300)
+      73.0±0.3ms       78.8±0.2ms     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'unique', 'transformation')
+         364±3μs          393±2μs     1.08  arithmetic.NumericInferOps.time_add(<class 'numpy.uint16'>)
+       267±0.8μs          288±2μs     1.08  groupby.GroupByMethods.time_dtype_as_group('object', 'tail', 'direct')
+     1.58±0.02ms      1.71±0.02ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude('complex128')
+     3.46±0.02μs       3.73±0.1μs     1.08  tslibs.normalize.Normalize.time_normalize_i8_timestamps(1, None)
+     1.55±0.01ms      1.67±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('bool')
+         310±5μs        335±0.5μs     1.08  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.int64'>, 13)
+     8.51±0.04ms       9.18±0.1ms     1.08  stat_ops.SeriesMultiIndexOps.time_op(0, 'skew')
+      74.5±0.5ms       80.3±0.8ms     1.08  groupby.GroupByMethods.time_dtype_as_field('object', 'unique', 'transformation')
+     2.49±0.01ms      2.69±0.03ms     1.08  io.csv.ReadCSVFloatPrecision.time_read_csv(',', '_', 'round_trip')
+      28.2±0.4μs      30.4±0.09μs     1.08  indexing.NumericSeriesIndexing.time_loc_scalar(<class 'pandas.core.indexes.numeric.Int64Index'>, 'unique_monotonic_inc')
+         361±2μs          389±2μs     1.08  groupby.GroupByMethods.time_dtype_as_field('float', 'sum', 'transformation')
+        85.6±1ms         92.2±1ms     1.08  stat_ops.SeriesMultiIndexOps.time_op(1, 'mad')
+         756±5μs         814±10μs     1.08  series_methods.IsInForObjects.time_isin_nans
+     1.47±0.01ms      1.58±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_include('float64')
+     1.55±0.01ms      1.67±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('complex64')
+     1.58±0.01ms      1.70±0.02ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude('uint32')
+     1.54±0.01ms      1.66±0.02ms     1.08  dtypes.SelectDtypes.time_select_dtype_int_exclude(<class 'bool'>)
+     1.54±0.01ms      1.66±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_exclude('Int16')
+      81.3±0.8μs       87.4±0.4μs     1.08  ctors.SeriesConstructors.time_series_constructor(<function no_change at 0x7f277fe8ac10>, False, 'int')
+         251±1μs          271±3μs     1.08  groupby.GroupByMethods.time_dtype_as_group('object', 'head', 'transformation')
+     1.43±0.01ms      1.54±0.01ms     1.08  dtypes.SelectDtypes.time_select_dtype_float_include(<class 'float'>)
+      3.64±0.1μs       3.91±0.2μs     1.08  index_cached_properties.IndexCache.time_inferred_type('IntervalIndex')
+     1.59±0.01ms      1.71±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('uint64')
+         515±3μs         553±20μs     1.07  multiindex_object.Duplicates.time_remove_unused_levels
+        41.2±1ms       44.2±0.1ms     1.07  io.csv.ReadCSVCategorical.time_convert_direct('c')
+         886±8ms          952±7ms     1.07  gil.ParallelGroups.time_get_groups(8)
+      18.3±0.2μs       19.6±0.2μs     1.07  series_methods.NanOps.time_func('prod', 1000, 'boolean')
+         438±4μs          471±3μs     1.07  series_methods.Map.time_map('Series', 'int')
+     1.53±0.01ms      1.65±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('UInt64')
+     1.57±0.01ms      1.69±0.02ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('int8')
+       160±0.4ms          172±2ms     1.07  groupby.GroupByMethods.time_dtype_as_group('float', 'unique', 'transformation')
+       287±0.5ms        308±0.4ms     1.07  join_merge.MergeCategoricals.time_merge_cat
+     1.56±0.02ms      1.67±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('uint32')
+      5.87±0.2ms      6.30±0.02ms     1.07  io.csv.ReadCSVParseSpecialDate.time_read_special_date('mdY', 'c')
+     1.58±0.02ms      1.70±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('datetime64[ns]')
+     16.2±0.05ms       17.4±0.7ms     1.07  stat_ops.SeriesMultiIndexOps.time_op(0, 'mad')
+     1.57±0.02ms      1.69±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('UInt16')
+     2.46±0.01ms      2.64±0.04ms     1.07  io.csv.ReadCSVFloatPrecision.time_read_csv(';', '_', None)
+      22.7±0.1μs       24.4±0.1μs     1.07  boolean.TimeLogicalOps.time_or_array
+     1.44±0.01ms         1.54±0ms     1.07  groupby.SumMultiLevel.time_groupby_sum_multiindex
+      39.0±0.5ms       41.8±0.3ms     1.07  indexing.ChainIndexing.time_chained_indexing('warn')
+         364±5μs          390±4μs     1.07  arithmetic.NumericInferOps.time_multiply(<class 'numpy.int16'>)
+       169±0.9μs        181±0.6μs     1.07  groupby.GroupByMethods.time_dtype_as_group('datetime', 'size', 'transformation')
+      29.1±0.3ms      31.2±0.08ms     1.07  hash_functions.IsinWithArange.time_isin(<class 'object'>, 8000, 0)
+       160±0.6ms          171±1ms     1.07  groupby.GroupByMethods.time_dtype_as_group('float', 'unique', 'direct')
+      21.9±0.1μs       23.5±0.2μs     1.07  boolean.TimeLogicalOps.time_xor_array
+      39.0±0.6ms      41.7±0.06ms     1.07  indexing.ChainIndexing.time_chained_indexing(None)
+      45.2±0.1μs       48.4±0.5μs     1.07  boolean.TimeLogicalOps.time_xor_scalar
+     1.56±0.02ms      1.67±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('uint16')
+     1.58±0.01ms      1.69±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('float64')
+         275±2μs          294±7μs     1.07  groupby.GroupByMethods.time_dtype_as_group('datetime', 'head', 'direct')
+         430±5μs          460±3μs     1.07  groupby.GroupByMethods.time_dtype_as_group('float', 'min', 'direct')
+      74.4±0.2ms       79.6±0.7ms     1.07  groupby.GroupByMethods.time_dtype_as_field('object', 'unique', 'direct')
+     1.56±0.02ms      1.66±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('float32')
+     30.0±0.03ms       32.0±0.1ms     1.07  reshape.Crosstab.time_crosstab_values
+     1.59±0.01ms      1.70±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('m8[ns]')
+     1.42±0.08μs      1.52±0.05μs     1.07  index_cached_properties.IndexCache.time_is_monotonic_decreasing('Int64Index')
+      16.3±0.2ms      17.5±0.05ms     1.07  frame_methods.Repr.time_repr_tall
+     1.51±0.01ms      1.62±0.02ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_include('int64')
+      31.8±0.1ms       34.0±0.1ms     1.07  reshape.Crosstab.time_crosstab
+     4.99±0.01ms      5.33±0.02ms     1.07  timeseries.AsOf.time_asof_nan('Series')
+     1.59±0.01ms         1.70±0ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('int16')
+      28.5±0.4ms       30.5±0.4ms     1.07  reshape.PivotTable.time_pivot_table
+      12.6±0.2ms      13.5±0.07ms     1.07  frame_methods.Apply.time_apply_ref_by_name
+      56.4±0.3μs       60.2±0.5μs     1.07  boolean.TimeLogicalOps.time_and_scalar
+     1.56±0.03ms      1.67±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('int64')
+     1.57±0.02ms      1.68±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('Int16')
+     1.56±0.01ms      1.67±0.02ms     1.07  dtypes.SelectDtypes.time_select_dtype_float_exclude('datetime64[ns]')
+         366±2μs          390±3μs     1.07  groupby.GroupByMethods.time_dtype_as_field('float', 'max', 'transformation')
+      47.8±0.5ms       51.0±0.3ms     1.07  gil.ParallelGroupbyMethods.time_parallel(4, 'prod')
+         398±4μs          425±2μs     1.07  groupby.GroupByMethods.time_dtype_as_group('object', 'first', 'transformation')
+      48.4±0.6ms       51.6±0.4ms     1.07  gil.ParallelGroupbyMethods.time_parallel(4, 'mean')
+     1.59±0.02ms         1.69±0ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('int32')
+         365±2μs          389±3μs     1.07  groupby.GroupByMethods.time_dtype_as_field('float', 'max', 'direct')
+     1.57±0.01ms      1.68±0.01ms     1.07  dtypes.SelectDtypes.time_select_dtype_int_exclude('Int32')
+         174±2μs        186±0.9μs     1.07  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 100000)
+         430±4μs          459±2μs     1.07  groupby.GroupByMethods.time_dtype_as_group('float', 'min', 'transformation')
+      50.5±0.3ms       53.8±0.8ms     1.07  groupby.Nth.time_series_nth_all('object')
+      98.1±0.1ms          105±1ms     1.07  join_merge.MergeAsof.time_by_object('backward', None)
+         274±2μs          292±2μs     1.07  groupby.GroupByMethods.time_dtype_as_group('datetime', 'head', 'transformation')
+      51.1±0.5ms       54.4±0.4ms     1.07  groupby.Nth.time_series_nth_any('object')
+       180±0.5ms        191±0.9ms     1.07  groupby.GroupByMethods.time_dtype_as_group('int', 'skew', 'direct')
+       252±0.7μs          269±2μs     1.07  groupby.GroupByMethods.time_dtype_as_group('object', 'head', 'direct')
+         160±1μs          171±1μs     1.06  groupby.GroupByMethods.time_dtype_as_group('object', 'count', 'transformation')
+         502±9μs          534±3μs     1.06  indexing_engines.NumericEngineIndexing.time_get_loc((<class 'pandas._libs.index.UInt8Engine'>, <class 'numpy.uint8'>), 'monotonic_incr')
+         825±1ms          878±2ms     1.06  join_merge.I8Merge.time_i8merge('outer')
+     7.94±0.04ms      8.45±0.06ms     1.06  stat_ops.SeriesMultiIndexOps.time_op(0, 'kurt')
+      5.02±0.1μs       5.34±0.2μs     1.06  index_cached_properties.IndexCache.time_shape('MultiIndex')
+         297±2μs          316±2μs     1.06  groupby.GroupByMethods.time_dtype_as_group('float', 'tail', 'direct')
+         268±2μs          285±1μs     1.06  groupby.GroupByMethods.time_dtype_as_group('object', 'tail', 'transformation')
+     1.52±0.02ms      1.62±0.02ms     1.06  dtypes.SelectDtypes.time_select_dtype_float_exclude(<class 'int'>)
+         755±4ms        803±0.7ms     1.06  stat_ops.SeriesMultiIndexOps.time_op([0, 1], 'mad')
+         395±3μs          420±4μs     1.06  groupby.GroupByMethods.time_dtype_as_group('object', 'last', 'direct')
+         430±2μs          457±2μs     1.06  groupby.GroupByMethods.time_dtype_as_group('float', 'first', 'direct')
+         408±3μs         433±10μs     1.06  groupby.GroupByMethods.time_dtype_as_field('int', 'last', 'transformation')
+     3.46±0.04μs      3.68±0.02μs     1.06  tslibs.normalize.Normalize.time_normalize_i8_timestamps(0, None)
+     1.92±0.09μs       2.04±0.2μs     1.06  index_cached_properties.IndexCache.time_inferred_type('DatetimeIndex')
+     1.56±0.02ms      1.66±0.01ms     1.06  dtypes.SelectDtypes.time_select_dtype_float_exclude('uint8')
+         365±2μs          387±2μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'min', 'direct')
+     1.72±0.02ms      1.82±0.01ms     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'pct_change', 'transformation')
+         367±4μs          390±4μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'sum', 'direct')
+         351±2μs          372±1μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'last', 'transformation')
+     1.56±0.02ms      1.65±0.01ms     1.06  dtypes.SelectDtypes.time_select_dtype_int_exclude(<class 'complex'>)
+         289±1μs          306±3μs     1.06  groupby.GroupByMethods.time_dtype_as_field('object', 'count', 'transformation')
+       160±0.7μs          170±2μs     1.06  groupby.GroupByMethods.time_dtype_as_group('object', 'count', 'direct')
+         369±3μs          391±3μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'min', 'transformation')
+         827±2ms          877±3ms     1.06  join_merge.I8Merge.time_i8merge('inner')
+     1.54±0.01ms      1.63±0.01ms     1.06  dtypes.SelectDtypes.time_select_dtype_float_exclude('Int32')
+         361±2μs          382±1μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'prod', 'transformation')
+     1.60±0.01ms      1.69±0.01ms     1.06  dtypes.SelectDtypes.time_select_dtype_int_exclude('float32')
+         317±2μs          336±5μs     1.06  groupby.GroupByMethods.time_dtype_as_field('datetime', 'head', 'direct')
+       319±0.6μs          338±4μs     1.06  groupby.GroupByMethods.time_dtype_as_group('int', 'head', 'transformation')
+         315±1μs          334±2μs     1.06  groupby.GroupByMethods.time_dtype_as_group('float', 'std', 'direct')
+      24.8±0.4ms       26.3±0.5ms     1.06  gil.ParallelGroupbyMethods.time_parallel(2, 'last')
+       122±0.3ms        129±0.3ms     1.06  groupby.GroupByMethods.time_dtype_as_field('int', 'skew', 'direct')
+      83.9±0.6μs       88.8±0.5μs     1.06  frame_ctor.FromSeries.time_mi_series
+      4.41±0.1μs      4.67±0.07μs     1.06  attrs_caching.SeriesArrayAttribute.time_extract_array_numpy('numeric')
+         338±3μs          358±2μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'shift', 'direct')
+     5.16±0.02ms      5.46±0.02ms     1.06  timeseries.AsOf.time_asof('Series')
+         343±3μs          363±3μs     1.06  reshape.Explode.time_explode(100, 5)
+         370±2μs          391±4μs     1.06  groupby.GroupByMethods.time_dtype_as_field('datetime', 'min', 'direct')
+         351±3μs          371±1μs     1.06  groupby.GroupByMethods.time_dtype_as_field('datetime', 'last', 'direct')
+         432±2μs          457±3μs     1.06  groupby.GroupByMethods.time_dtype_as_group('float', 'first', 'transformation')
+         310±1μs          328±5μs     1.06  groupby.GroupByMethods.time_dtype_as_field('datetime', 'any', 'direct')
+     1.17±0.01ms         1.24±0ms     1.06  index_object.IntervalIndexMethod.time_intersection_both_duplicate(1000)
+     1.22±0.05μs      1.29±0.08μs     1.06  index_cached_properties.IndexCache.time_is_monotonic('Int64Index')
+         501±2μs          530±2μs     1.06  groupby.GroupByMethods.time_dtype_as_field('object', 'last', 'transformation')
+      43.2±0.2ms       45.7±0.3ms     1.06  frame_methods.SortIndexByColumns.time_frame_sort_values_by_columns
+     5.56±0.01ms       5.88±0.1ms     1.06  timeseries.ToDatetimeFormatQuarters.time_infer_quarter
+         163±1ms          172±1ms     1.06  io.json.ReadJSON.time_read_json('split', 'int')
+      17.9±0.1μs       18.9±0.2μs     1.06  series_methods.NanOps.time_func('sum', 1000, 'boolean')
+         328±2μs          347±2μs     1.06  groupby.GroupByMethods.time_dtype_as_field('int', 'tail', 'transformation')
+     2.88±0.01ms      3.05±0.02ms     1.06  dtypes.SelectDtypes.time_select_dtype_string_exclude(<class 'float'>)
+         320±2ms          337±2ms     1.06  io.excel.ReadExcel.time_read_excel('openpyxl')
+       161±0.8μs          169±1μs     1.06  groupby.GroupByMethods.time_dtype_as_field('int', 'count', 'transformation')
+         338±1μs          356±3μs     1.06  groupby.GroupByMethods.time_dtype_as_field('float', 'shift', 'transformation')
+         309±2μs          327±2μs     1.06  groupby.GroupByMethods.time_dtype_as_field('datetime', 'all', 'transformation')
+       169±0.9ms        178±0.5ms     1.05  io.json.ReadJSON.time_read_json('split', 'datetime')
+         317±1μs          335±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'head', 'direct')
+     26.8±0.09ms       28.2±0.1ms     1.05  categoricals.Constructor.time_regular
+       180±0.8ms          190±1ms     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'skew', 'transformation')
+     14.2±0.05ms       14.9±0.1ms     1.05  stat_ops.Correlation.time_corrwith_rows('pearson')
+     4.01±0.02μs         4.23±0μs     1.05  tslibs.normalize.Normalize.time_normalize_i8_timestamps(100, None)
+         856±1ms          902±3ms     1.05  join_merge.I8Merge.time_i8merge('left')
+         432±3μs          456±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'max', 'transformation')
+         426±3μs          450±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'min', 'transformation')
+         400±2μs          422±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('object', 'first', 'direct')
+       396±0.6μs          418±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('object', 'last', 'transformation')
+       161±0.8μs          170±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'count', 'direct')
+      11.0±0.1μs       11.6±0.2μs     1.05  timeseries.DatetimeIndex.time_get('tz_naive')
+     1.57±0.01ms      1.66±0.01ms     1.05  dtypes.SelectDtypes.time_select_dtype_float_exclude('m8[ns]')
+         281±1ms          296±2ms     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'skew', 'direct')
+       128±0.7μs          135±1μs     1.05  arithmetic.CategoricalComparisons.time_categorical_op('__le__')
+         360±2μs          379±4μs     1.05  groupby.GroupByMethods.time_dtype_as_field('datetime', 'max', 'transformation')
+      63.6±0.7ms       67.0±0.7ms     1.05  frame_methods.Count.time_count_level_multi(0)
+       163±0.7μs          171±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'count', 'direct')
+       122±0.5ms        129±0.4ms     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'skew', 'transformation')
+         427±1μs          449±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'first', 'direct')
+        2.92±0ms      3.07±0.01ms     1.05  dtypes.SelectDtypes.time_select_dtype_string_exclude('complex64')
+         284±2μs          299±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'head', 'direct')
+         223±2ms          235±2ms     1.05  frame_methods.Apply.time_apply_user_func
+         315±1μs          331±2μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'head', 'direct')
+      61.2±0.5μs       64.4±0.6μs     1.05  boolean.TimeLogicalOps.time_or_scalar
+      50.4±0.2μs       53.0±0.3μs     1.05  arithmetic.CategoricalComparisons.time_categorical_op('__ge__')
+       277±0.6ms          291±2ms     1.05  join_merge.MergeAsof.time_by_object('nearest', 5)
+      34.2±0.1μs       36.0±0.2μs     1.05  timeseries.AsOf.time_asof_single('Series')
+         328±2μs        345±0.8μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'tail', 'direct')
+       277±0.4ms          291±2ms     1.05  join_merge.MergeAsof.time_by_object('nearest', None)
+         340±2μs          357±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('object', 'tail', 'transformation')
+         417±4μs          438±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'last', 'direct')
+       162±0.6μs          171±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'count', 'transformation')
+         227±1μs        239±0.7μs     1.05  tslibs.normalize.Normalize.time_normalize_i8_timestamps(10000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+        959±20μs      1.01±0.02ms     1.05  ctors.SeriesConstructors.time_series_constructor(<function list_of_str at 0x7f277fe8aaf0>, True, 'int')
+         333±2μs          350±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'tail', 'direct')
+         429±2μs          451±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'min', 'direct')
+         296±2μs          311±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('datetime', 'cumcount', 'transformation')
+         354±2μs          372±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('datetime', 'last', 'transformation')
+         297±2μs          312±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('datetime', 'cumcount', 'direct')
+         428±2μs        450±0.7μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'last', 'direct')
+         454±3μs          477±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('object', 'nunique', 'transformation')
+      85.8±0.7ms         90.1±1ms     1.05  groupby.Groups.time_series_groups('object_small')
+       502±0.9μs          527±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('object', 'last', 'direct')
+         303±2μs          318±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'cumcount', 'direct')
+         294±2μs          308±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('datetime', 'tail', 'direct')
+         411±1μs          431±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'last', 'direct')
+         238±1ms          250±2ms     1.05  gil.ParallelGroups.time_get_groups(2)
+         304±1μs          319±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'cumcount', 'transformation')
+     2.39±0.01ms      2.50±0.02ms     1.05  timeseries.AsOf.time_asof_single('DataFrame')
+         329±1μs          345±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'tail', 'transformation')
+       310±0.9μs          325±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('datetime', 'any', 'transformation')
+         434±4μs          455±2μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'first', 'direct')
+         427±3μs          448±4μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'max', 'direct')
+      97.4±0.2ms          102±6ms     1.05  tslibs.resolution.TimeResolution.time_get_resolution('D', 1000000, None)
+         319±1μs          334±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'std', 'direct')
+         303±1μs          317±2μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'std', 'direct')
+      50.6±0.3μs       53.0±0.4μs     1.05  arithmetic.CategoricalComparisons.time_categorical_op('__gt__')
+         303±2μs          317±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'std', 'transformation')
+        926±20μs         969±20μs     1.05  ctors.SeriesConstructors.time_series_constructor(<function list_of_str at 0x7f277fe8aaf0>, False, 'int')
+       308±0.4μs          322±2μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'std', 'direct')
+       347±0.9μs          363±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'shift', 'direct')
+       198±0.4ms        207±0.7ms     1.05  join_merge.MergeAsof.time_by_object('forward', None)
+         307±1μs          322±2μs     1.05  groupby.GroupByMethods.time_dtype_as_field('int', 'std', 'transformation')
+         282±2ms          295±3ms     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'skew', 'transformation')
+         315±3μs          330±3μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'head', 'transformation')
+     2.93±0.01ms      3.06±0.01ms     1.05  dtypes.SelectDtypes.time_select_dtype_string_exclude('int16')
+         431±3μs          451±3μs     1.05  groupby.GroupByMethods.time_dtype_as_group('int', 'max', 'transformation')
+         280±2μs          293±1μs     1.05  groupby.GroupByMethods.time_dtype_as_group('object', 'cumcount', 'direct')
+     2.94±0.01ms      3.07±0.03ms     1.05  dtypes.SelectDtypes.time_select_dtype_string_exclude('int8')
+         331±2μs          346±2μs     1.05  groupby.GroupByMethods.time_dtype_as_field('object', 'cumcount', 'transformation')
+     2.08±0.01ms      2.17±0.03ms     1.05  series_methods.Clip.time_clip(100000)
+         311±3μs          325±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('datetime', 'all', 'direct')
+      37.1±0.4ms       38.8±0.1ms     1.05  arithmetic.IrregularOps.time_add
+      15.2±0.1μs       15.9±0.2μs     1.05  series_methods.NanOps.time_func('min', 1000, 'boolean')
+         569±6ns          595±4ns     1.05  dtypes.Dtypes.time_pandas_dtype(dtype('int32'))
+       331±0.7μs          345±1μs     1.05  groupby.GroupByMethods.time_dtype_as_field('float', 'cumcount', 'transformation')
+     2.01±0.02ms         2.10±0ms     1.05  groupby.GroupByMethods.time_dtype_as_group('float', 'pct_change', 'direct')
+      23.1±0.1ms       24.2±0.2ms     1.04  reshape.Explode.time_explode(10000, 10)
+         333±1μs          348±3μs     1.04  groupby.GroupByMethods.time_dtype_as_field('datetime', 'tail', 'transformation')
+         369±7μs          385±2μs     1.04  groupby.GroupByMethods.time_dtype_as_field('float', 'first', 'direct')
+         374±3μs          391±5μs     1.04  groupby.GroupByMethods.time_dtype_as_field('datetime', 'first', 'direct')
+         506±8ns          529±3ns     1.04  dtypes.DtypesInvalid.time_pandas_dtype_invalid('array-string')
+         316±2μs          330±3μs     1.04  groupby.GroupByMethods.time_dtype_as_group('int', 'std', 'transformation')
+        736±30ns         769±40ns     1.04  index_cached_properties.IndexCache.time_is_unique('RangeIndex')
+     2.52±0.02ms      2.63±0.03ms     1.04  frame_methods.Interpolate.time_interpolate_some_good(None)
+     1.24±0.06μs      1.30±0.04μs     1.04  index_cached_properties.IndexCache.time_is_monotonic('RangeIndex')
+         625±8ns         652±10ns     1.04  dtypes.Dtypes.time_pandas_dtype(period[D])
+     2.94±0.03ms      3.07±0.01ms     1.04  dtypes.SelectDtypes.time_select_dtype_string_exclude('complex128')
+     2.92±0.01ms      3.05±0.02ms     1.04  dtypes.SelectDtypes.time_select_dtype_string_exclude('uint16')
+         428±2μs          446±2μs     1.04  groupby.GroupByMethods.time_dtype_as_field('int', 'min', 'direct')
+      3.79±0.2μs       3.95±0.2μs     1.04  index_cached_properties.IndexCache.time_inferred_type('TimedeltaIndex')
+     2.92±0.02ms      3.04±0.03ms     1.04  dtypes.SelectDtypes.time_select_dtype_string_exclude('Int64')
+         569±5ns          593±2ns     1.04  dtypes.Dtypes.time_pandas_dtype(dtype('int16'))
+     2.93±0.03ms      3.05±0.02ms     1.04  dtypes.SelectDtypes.time_select_dtype_string_exclude('UInt64')
+     1.88±0.01ms      1.96±0.01ms     1.04  groupby.GroupByMethods.time_dtype_as_group('int', 'pct_change', 'direct')
+         401±2μs         417±20μs     1.04  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(10000, 6000)
+         330±2μs          344±1μs     1.04  groupby.GroupByMethods.time_dtype_as_field('object', 'cumcount', 'direct')
+       198±0.3ms        206±0.8ms     1.04  join_merge.MergeAsof.time_by_object('forward', 5)
+     4.15±0.03ms      4.32±0.04ms     1.04  reshape.Melt.time_melt_dataframe
+         318±1μs          332±2μs     1.04  groupby.GroupByMethods.time_dtype_as_field('datetime', 'head', 'transformation')
+         362±2μs          377±3μs     1.04  groupby.GroupByMethods.time_dtype_as_group('object', 'bfill', 'direct')
+     3.82±0.02ms      3.97±0.02ms     1.04  timeseries.ToDatetimeCache.time_dup_string_dates(True)
+      30.1±0.1ms       31.4±0.2ms     1.04  timeseries.ToDatetimeFormat.time_same_offset_to_utc
+         761±4μs          791±6μs     1.04  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.uint64'>, 15)
+       323±0.9μs          335±2μs     1.04  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.uint64'>, 13)
+     15.9±0.09μs       16.5±0.2μs     1.04  series_methods.NanOps.time_func('prod', 1000, 'Int64')
+       164±0.6ms        171±0.6ms     1.04  series_methods.IsInLongSeriesValuesDominate.time_isin('object', 'monotone')
+        952±10μs         990±10μs     1.04  index_cached_properties.IndexCache.time_is_monotonic_decreasing('MultiIndex')
+         434±3μs          451±2μs     1.04  groupby.GroupByMethods.time_dtype_as_group('int', 'max', 'direct')
+         372±2μs          387±3μs     1.04  groupby.GroupByMethods.time_dtype_as_field('datetime', 'first', 'transformation')
+       226±0.5μs          235±2μs     1.04  indexing.CategoricalIndexIndexing.time_getitem_bool_array('monotonic_decr')
+        1.06±0ms      1.10±0.06ms     1.04  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 1011, datetime.timezone.utc)
+     1.17±0.01ms      1.21±0.02ms     1.04  frame_methods.Quantile.time_frame_quantile(1)
+         570±5ns          592±2ns     1.04  dtypes.Dtypes.time_pandas_dtype(dtype('float32'))
+         335±3μs          348±1μs     1.04  groupby.GroupByMethods.time_dtype_as_group('int', 'tail', 'transformation')
+     1.79±0.02ms      1.86±0.01ms     1.04  groupby.GroupByMethods.time_dtype_as_field('int', 'pct_change', 'direct')
+       327±0.5μs        339±0.6μs     1.04  groupby.GroupByMethods.time_dtype_as_field('object', 'head', 'direct')
+         331±2μs          343±2μs     1.04  groupby.GroupByMethods.time_dtype_as_field('datetime', 'cumcount', 'transformation')
+     1.34±0.01ms         1.39±0ms     1.04  timeseries.InferFreq.time_infer_freq('D')
+       157±0.3ms        163±0.7ms     1.04  frame_ctor.FromDicts.time_nested_dict_int64
+         348±1μs          361±2μs     1.04  groupby.GroupByMethods.time_dtype_as_group('float', 'shift', 'transformation')
+      15.1±0.1μs       15.7±0.3μs     1.04  series_methods.NanOps.time_func('max', 1000, 'boolean')
+         623±2μs          646±3μs     1.04  groupby.GroupByMethods.time_dtype_as_field('int', 'nunique', 'direct')
+     8.62±0.04ms       8.95±0.1ms     1.04  reindex.DropDuplicates.time_frame_drop_dups(True)
+         341±3μs          354±2μs     1.04  groupby.GroupByMethods.time_dtype_as_field('object', 'tail', 'direct')
+         569±6ns          590±2ns     1.04  dtypes.Dtypes.time_pandas_dtype(dtype('float64'))
+     2.39±0.01ms      2.48±0.01ms     1.04  timeseries.ResampleDataFrame.time_method('min')
+         361±2μs          374±6μs     1.04  groupby.GroupByMethods.time_dtype_as_group('object', 'ffill', 'transformation')
+         361±2μs          374±2μs     1.04  groupby.GroupByMethods.time_dtype_as_group('object', 'ffill', 'direct')
+         682±3μs         707±40μs     1.04  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(10000, 8000)
+      45.7±0.4ms       47.4±0.2ms     1.04  groupby.Nth.time_frame_nth_any('datetime')
+     5.76±0.03ms      5.97±0.02ms     1.04  dtypes.InferDtypes.time_infer('py-object')
+     2.89±0.02ms      2.99±0.02ms     1.04  dtypes.SelectDtypes.time_select_dtype_string_exclude(<class 'int'>)
+     16.6±0.09ms      17.2±0.05ms     1.04  reindex.DropDuplicates.time_frame_drop_dups(False)
+         593±1ms          614±1ms     1.04  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 1000, 'monotone_misses')
+         573±4ns          593±2ns     1.04  dtypes.Dtypes.time_pandas_dtype(dtype('int64'))
+     2.96±0.03ms      3.07±0.01ms     1.03  dtypes.SelectDtypes.time_select_dtype_string_exclude('timedelta64[ns]')
+         640±4μs          663±2μs     1.03  groupby.GroupByMethods.time_dtype_as_group('float', 'nunique', 'transformation')
+     1.74±0.01ms         1.80±0ms     1.03  series_methods.Map.time_map('Series', 'object')
+      65.5±0.5μs       67.8±0.7μs     1.03  frame_ctor.FromNDArray.time_frame_from_ndarray
+         588±3μs          609±3μs     1.03  groupby.GroupByMethods.time_dtype_as_field('object', 'nunique', 'direct')
+      65.2±0.6ms       67.4±0.6ms     1.03  io.style.RenderApply.time_render(36, 12)
+       162±0.5μs        168±0.9μs     1.03  timedelta.ToTimedelta.time_convert_int
+         571±5ns          590±1ns     1.03  dtypes.Dtypes.time_pandas_dtype(dtype('uint64'))
+     6.08±0.03ms      6.29±0.04ms     1.03  io.csv.ParseDateComparison.time_to_datetime_dayfirst(True)
+      28.6±0.2ms       29.6±0.2ms     1.03  stat_ops.SeriesMultiIndexOps.time_op(1, 'skew')
+     2.94±0.02ms      3.04±0.02ms     1.03  dtypes.SelectDtypes.time_select_dtype_string_exclude('int32')
+         529±2μs          546±4μs     1.03  series_methods.Map.time_map('Series', 'category')
+     36.5±0.09ms      37.7±0.08ms     1.03  groupby.AggEngine.time_dataframe_cython(False)
+        993±10μs      1.03±0.01ms     1.03  ctors.SeriesConstructors.time_series_constructor(<function list_of_str at 0x7f277fe8aaf0>, True, 'float')
+     1.28±0.01ms      1.32±0.03ms     1.03  stat_ops.Covariance.time_cov_series
+     2.95±0.03ms      3.05±0.01ms     1.03  dtypes.SelectDtypes.time_select_dtype_string_exclude('float32')
+     2.21±0.01ms      2.28±0.01ms     1.03  io.csv.ReadCSVDInferDatetimeFormat.time_read_csv(False, 'ymd')
+         572±6ns        591±0.7ns     1.03  dtypes.Dtypes.time_pandas_dtype(dtype('uint8'))
+     2.98±0.04ms      3.07±0.01ms     1.03  dtypes.SelectDtypes.time_select_dtype_string_exclude('m8[ns]')
+       267±0.7μs        275±0.8μs     1.03  groupby.GroupByMethods.time_dtype_as_group('int', 'all', 'transformation')
+         646±3μs        667±0.9μs     1.03  groupby.GroupByMethods.time_dtype_as_group('int', 'nunique', 'transformation')
+      47.3±0.4ms       48.8±0.4ms     1.03  io.sql.WriteSQLDtypes.time_to_sql_dataframe_column('sqlite', 'datetime')
+         572±6ns          590±2ns     1.03  dtypes.Dtypes.time_pandas_dtype(dtype('uint32'))
+         714±2μs          736±2μs     1.03  groupby.GroupByMethods.time_dtype_as_field('datetime', 'nunique', 'direct')
+     8.94±0.04ms      9.22±0.09ms     1.03  frame_ctor.FromArrays.time_frame_from_arrays_sparse
+      34.7±0.2ms      35.8±0.07ms     1.03  sparse.ToCoo.time_sparse_series_to_coo
+      33.5±0.1ms       34.5±0.1ms     1.03  groupby.AggEngine.time_series_cython(True)
+         646±4μs          665±4μs     1.03  groupby.GroupByMethods.time_dtype_as_group('float', 'nunique', 'direct')
+     7.18±0.08ms       7.40±0.1ms     1.03  ctors.SeriesConstructors.time_series_constructor(<function gen_of_tuples at 0x7f277fe8a280>, False, 'int')
+         333±1μs          342±1μs     1.03  groupby.GroupByMethods.time_dtype_as_field('datetime', 'cumcount', 'direct')
+         574±7ns          592±2ns     1.03  dtypes.Dtypes.time_pandas_dtype(dtype('int8'))
+     2.55±0.01ms      2.63±0.01ms     1.03  timeseries.ResampleDataFrame.time_method('max')
+     1.00±0.01μs      1.03±0.01μs     1.03  dtypes.Dtypes.time_pandas_dtype(<class 'pandas.core.arrays.integer.UInt8Dtype'>)
+        1.29±0ms      1.32±0.06ms     1.03  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 1011, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+     7.19±0.08ms      7.40±0.09ms     1.03  ctors.SeriesConstructors.time_series_constructor(<function gen_of_tuples at 0x7f277fe8a280>, False, 'float')
+      67.1±0.5ms       69.1±0.3ms     1.03  gil.ParallelFactorize.time_loop(4)
+      1.18±0.01s          1.21±0s     1.03  join_merge.MergeAsof.time_multiby('nearest', 5)
+         762±1ms          784±1ms     1.03  join_merge.MergeCategoricals.time_merge_object
+         625±3μs          643±2μs     1.03  groupby.GroupByMethods.time_dtype_as_field('int', 'nunique', 'transformation')
+         268±1μs          275±2μs     1.03  groupby.GroupByMethods.time_dtype_as_group('int', 'all', 'direct')
+         623±5ns          640±8ns     1.03  dtypes.Dtypes.time_pandas_dtype(datetime64[ns, UTC])
+      91.1±0.3ms       93.6±0.4ms     1.03  plotting.SeriesPlotting.time_series_plot('bar')
+     5.65±0.03ms      5.81±0.03ms     1.03  groupby.CountMultiDtype.time_multi_count
+     1.06±0.01ms         1.09±0ms     1.03  groupby.GroupByMethods.time_dtype_as_group('object', 'value_counts', 'direct')
+         524±4μs          538±2μs     1.03  groupby.GroupByMethods.time_dtype_as_group('float', 'bfill', 'direct')
+         656±2μs          674±2μs     1.03  categoricals.Repr.time_rendering
+     14.7±0.03μs      15.1±0.05μs     1.03  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(100, 8000)
+         526±2μs          539±4μs     1.03  groupby.GroupByMethods.time_dtype_as_field('int', 'ffill', 'transformation')
+         451±3ms        462±0.7ms     1.02  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.UInt64Index'>, 5000000)
+         683±2μs          700±4μs     1.02  groupby.GroupByMethods.time_dtype_as_field('object', 'bfill', 'transformation')
+         1.07±0s       1.09±0.01s     1.02  join_merge.MergeAsof.time_multiby('forward', 5)
+         714±4μs          731±2μs     1.02  groupby.GroupByMethods.time_dtype_as_field('datetime', 'nunique', 'transformation')
+      92.2±0.4ms       94.5±0.8ms     1.02  plotting.TimeseriesPlotting.time_plot_regular
+     5.25±0.02ms      5.38±0.03ms     1.02  dtypes.InferDtypes.time_infer('bytes')
+      33.5±0.1ms       34.3±0.1ms     1.02  groupby.AggEngine.time_series_cython(False)
+         1.06±0s          1.09±0s     1.02  join_merge.MergeAsof.time_multiby('forward', None)
+         830±2μs          850±2μs     1.02  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(10000, 9000)
+       256±0.3ms          262±6ms     1.02  io.stata.StataMissing.time_write_stata('th')
+       272±0.9μs          278±1μs     1.02  groupby.GroupByMethods.time_dtype_as_group('float', 'any', 'transformation')
+      11.9±0.1ms      12.1±0.03ms     1.02  timedelta.ToTimedelta.time_convert_string_seconds
+     1.21±0.01ms         1.24±0ms     1.02  index_object.SetOperations.time_operation('int', 'union')
+        1.68±0ms      1.72±0.01ms     1.02  arithmetic.OffsetArrayArithmetic.time_add_series_offset(<MonthEnd>)
+         529±2μs          540±2μs     1.02  groupby.GroupByMethods.time_dtype_as_group('int', 'ffill', 'direct')
+     15.8±0.04ms      16.1±0.02ms     1.02  timeseries.SortIndex.time_sort_index(False)
+     5.65±0.07ms      5.77±0.07ms     1.02  index_cached_properties.IndexCache.time_is_all_dates('CategoricalIndex')
+      109±0.09ms        111±0.6ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 2000, datetime.timezone(datetime.timedelta(seconds=3600)))
+       106±0.1ms        108±0.4ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 1000, datetime.timezone(datetime.timedelta(seconds=3600)))
+     1.38±0.01ms      1.41±0.01ms     1.02  frame_ctor.FromRecords.time_frame_from_records_generator(1000)
+       105±0.3ms        108±0.3ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 1011, datetime.timezone(datetime.timedelta(seconds=3600)))
+     4.30±0.03ms      4.39±0.04ms     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_tuples_with_none at 0x7f277fe8a3a0>, True, 'float')
+       122±0.4ms        125±0.2ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 2011, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+        1.13±0ms      1.16±0.02ms     1.02  tslibs.resolution.TimeResolution.time_get_resolution('D', 10000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+       119±0.3ms        122±0.6ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 1000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+     18.0±0.06ms      18.4±0.07ms     1.02  join_merge.MergeAsof.time_on_int('forward', 5)
+         577±2μs          588±2μs     1.02  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(10000, 7000)
+        975±20μs         993±10μs     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_str at 0x7f277fe8aaf0>, False, 'float')
+        697±40μs         709±30μs     1.02  index_cached_properties.IndexCache.time_is_monotonic('MultiIndex')
+     4.30±0.03ms      4.38±0.04ms     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_tuples at 0x7f277fe8a040>, True, 'float')
+         458±3ms          466±1ms     1.02  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Int64Index'>, 5000000)
+         530±2μs          539±3μs     1.02  groupby.GroupByMethods.time_dtype_as_group('int', 'ffill', 'transformation')
+       133±0.4ms        135±0.1ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 4006, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+     4.28±0.04ms      4.35±0.04ms     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_tuples at 0x7f277fe8a040>, False, 'float')
+     12.4±0.07ms      12.6±0.03ms     1.02  index_object.IndexAppend.time_append_range_list
+     4.28±0.03ms      4.35±0.04ms     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_tuples_with_none at 0x7f277fe8a3a0>, False, 'float')
+       150±0.3ms        152±0.1ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 4000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+      98.6±0.1ms          100±4ms     1.02  tslibs.resolution.TimeResolution.time_get_resolution('D', 1000000, datetime.timezone(datetime.timedelta(seconds=3600)))
+        1.09±0ms         1.11±0ms     1.02  tslibs.resolution.TimeResolution.time_get_resolution('us', 10000, datetime.timezone(datetime.timedelta(seconds=3600)))
+        1.20±0ms         1.22±0ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 4000, datetime.timezone(datetime.timedelta(seconds=3600)))
+       298±0.6ms        303±0.4ms     1.02  index_object.Indexing.time_get_loc_non_unique_sorted('String')
+     7.59±0.02ms      7.71±0.01ms     1.02  dtypes.InferDtypes.time_infer_skipna('np-object')
+       131±0.2ms        133±0.1ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 1011, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+       119±0.4ms        121±0.3ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 1011, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+        1.30±0ms      1.32±0.01ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 4000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+        1.10±0ms         1.12±0ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 3000, datetime.timezone(datetime.timedelta(seconds=3600)))
+     100.0±0.1ms        102±0.1ms     1.02  tslibs.resolution.TimeResolution.time_get_resolution('m', 1000000, datetime.timezone(datetime.timedelta(seconds=3600)))
+       148±0.3ms        150±0.4ms     1.02  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 8000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+     4.30±0.04ms      4.37±0.05ms     1.02  ctors.SeriesConstructors.time_series_constructor(<function list_of_lists at 0x7f277fe8a550>, True, 'float')
+     4.27±0.04ms      4.34±0.04ms     1.01  ctors.SeriesConstructors.time_series_constructor(<function list_of_lists at 0x7f277fe8a550>, False, 'float')
+        1.47±0ms         1.49±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 4000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+        1.58±0ms      1.60±0.02ms     1.01  arithmetic.OffsetArrayArithmetic.time_add_series_offset(<SemiMonthBegin: day_of_month=15>)
+        1.20±0ms         1.21±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 3000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+       100±0.1ms       102±0.03ms     1.01  tslibs.resolution.TimeResolution.time_get_resolution('ns', 1000000, datetime.timezone(datetime.timedelta(seconds=3600)))
+        1.45±0ms         1.47±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 8000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+        1.11±0ms         1.12±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 2011, datetime.timezone(datetime.timedelta(seconds=3600)))
+     4.27±0.03ms      4.33±0.04ms     1.01  ctors.SeriesConstructors.time_series_constructor(<function list_of_lists_with_none at 0x7f277fe8a1f0>, False, 'float')
+        1.17±0ms         1.19±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 1000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+        1.02±0ms      1.03±0.01ms     1.01  tslibs.resolution.TimeResolution.time_get_resolution('ns', 10000, datetime.timezone(datetime.timedelta(seconds=3600)))
+        1.11±0ms         1.12±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 2000, datetime.timezone(datetime.timedelta(seconds=3600)))
+        1.32±0ms      1.34±0.01ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 2000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+       149±0.3ms        151±0.2ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 11000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+        1.32±0ms         1.34±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 2011, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+       150±0.6ms        152±0.1ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 4006, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+     4.30±0.04ms      4.35±0.04ms     1.01  ctors.SeriesConstructors.time_series_constructor(<function list_of_lists_with_none at 0x7f277fe8a1f0>, True, 'float')
+        1.30±0ms         1.32±0ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(10000, 4006, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
+       147±0.4ms        149±0.2ms     1.01  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1000000, 6000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
+     99.6±0.06ms        101±0.2ms     1.01  tslibs.resolution.TimeResolution.time_get_resolution('m', 1000000, None)
+     3.39±0.02ms      3.43±0.03ms     1.01  ctors.SeriesConstructors.time_series_constructor(<class 'list'>, True, 'int')
+            116M             117M     1.01  rolling.VariableWindowMethods.peakmem_rolling('DataFrame', '50s', 'float', 'count')
+      75.1±0.4ms       75.9±0.2ms     1.01  io.parsers.DoesStringLookLikeDatetime.time_check_datetimes('2Q2005')
+         997±1μs         1.01±0ms     1.01  tslibs.resolution.TimeResolution.time_get_resolution('ns', 10000, datetime.timezone.utc)
+            109M             110M     1.01  stat_ops.Correlation.peakmem_corr_wide('pearson')
-            112M             111M     0.99  rolling.ForwardWindowMethods.peakmem_rolling('DataFrame', 1000, 'float', 'mean')
-            112M             111M     0.99  rolling.ForwardWindowMethods.peakmem_rolling('DataFrame', 1000, 'float', 'sum')
-            114M             113M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'float', 'mean')
-            114M             113M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'float', 'kurt')
-            116M             115M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'std')
-            115M             114M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'sum')
-            115M             114M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'skew')
-            115M             114M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'mean')
-            118M             117M     0.99  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'count')
-            115M             114M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'kurt')
-     1.57±0.08ms      1.55±0.02ms     0.99  ctors.SeriesConstructors.time_series_constructor(<class 'list'>, True, 'float')
-     3.30±0.01ms         3.26±0ms     0.99  arithmetic.OffsetArrayArithmetic.time_add_series_offset(<BusinessDay>)
-            118M             117M     0.99  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'count')
-            118M             117M     0.99  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'count')
-            115M             113M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'min')
-            115M             113M     0.99  rolling.Methods.peakmem_rolling('DataFrame', 10, 'int', 'max')
-      67.2±0.3ms      66.2±0.08ms     0.99  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'int', 'median')
-      22.9±0.1μs      22.5±0.08μs     0.98  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(100, 2011, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-         551±2ns          542±1ns     0.98  tslibs.period.PeriodProperties.time_property('M', 'daysinmonth')
-     6.73±0.01ms      6.62±0.01ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('datetime', 10000, datetime.timezone(datetime.timedelta(seconds=3600)))
-         863±4ns          849±2ns     0.98  tslibs.timestamp.TimestampOps.time_to_pydatetime(tzutc())
-         2.35±0s          2.31±0s     0.98  stat_ops.Correlation.time_corr_wide_nans('kendall')
-      81.8±0.4μs       80.5±0.3μs     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('timestamp', 100, datetime.timezone.utc)
-            114M             112M     0.98  rolling.Methods.peakmem_rolling('DataFrame', 10, 'float', 'min')
-            114M             112M     0.98  rolling.Methods.peakmem_rolling('DataFrame', 10, 'float', 'max')
-     5.49±0.03ms      5.40±0.01ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('time', 10000, datetime.timezone(datetime.timedelta(seconds=3600)))
-     8.18±0.02ms      8.04±0.02ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('timestamp', 10000, None)
-     5.45±0.02ms      5.36±0.01ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('time', 10000, datetime.timezone.utc)
-     6.54±0.03ms      6.43±0.02ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('datetime', 10000, datetime.timezone.utc)
-     2.99±0.01ms         2.94±0ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('date', 10000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-        3.01±0ms         2.95±0ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('date', 10000, None)
-            120M             118M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'median')
-     3.00±0.02ms      2.94±0.01ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('date', 10000, datetime.timezone(datetime.timedelta(seconds=3600)))
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'median')
-         502±4μs          493±2μs     0.98  groupby.GroupByMethods.time_dtype_as_group('int', 'var', 'direct')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'skew')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'skew')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'kurt')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'kurt')
-     21.6±0.09μs       21.2±0.5μs     0.98  tslibs.resolution.TimeResolution.time_get_resolution('ns', 100, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'kurt')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'min')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'max')
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'median')
-            119M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'median')
-       175±0.6ms          171±1ms     0.98  index_cached_properties.IndexCache.time_is_monotonic_decreasing('IntervalIndex')
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'median')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'kurt')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'skew')
-       168±0.8ms        165±0.8ms     0.98  index_cached_properties.IndexCache.time_engine('IntervalIndex')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'median')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'skew')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'skew')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'kurt')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'kurt')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'std')
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'std')
-     15.1±0.07ms      14.8±0.06ms     0.98  stat_ops.Correlation.time_corrwith_cols('spearman')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'std')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'skew')
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'std')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'sum')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'std')
-            119M             117M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'std')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'mean')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'mean')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'sum')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'mean')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'sum')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'sum')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'mean')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'sum')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'min')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'mean')
-     8.24±0.03ms      8.07±0.02ms     0.98  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('timestamp', 10000, datetime.timezone.utc)
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'min')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'sum')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'min')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'int', 'max')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'float', 'mean')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'min')
-            118M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'float', 'max')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1d', 'int', 'max')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'max')
-            117M             115M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '50s', 'float', 'min')
-            118M             116M     0.98  rolling.VariableWindowMethods.peakmem_rolling('Series', '1h', 'int', 'max')
-            203M             199M     0.98  io.json.ToJSON.peakmem_to_json_wide('columns', 'df_int_float_str')
-      16.4±0.1ms      16.1±0.08ms     0.98  join_merge.Concat.time_concat_small_frames(1)
-     6.43±0.03ms      6.28±0.05ms     0.98  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 8000, 0)
-      18.2±0.1ms      17.8±0.03ms     0.98  join_merge.MergeAsof.time_on_uint64('forward', 5)
-            197M             193M     0.98  io.json.ToJSON.peakmem_to_json_wide('columns', 'df_td_int_ts')
-            318M             311M     0.98  frame_methods.Iteration.peakmem_itertuples_start
-            318M             311M     0.98  frame_methods.Iteration.peakmem_itertuples
-         870±3ns          850±3ns     0.98  tslibs.timestamp.TimestampOps.time_to_pydatetime(<UTC>)
-       169±0.5ms        165±0.9ms     0.98  index_cached_properties.IndexCache.time_is_monotonic('IntervalIndex')
-     9.48±0.05ms      9.26±0.02ms     0.98  stat_ops.Correlation.time_corrwith_cols('kendall')
-      55.9±0.4ms       54.5±0.1ms     0.98  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0.5, 'nearest')
-            193M             189M     0.98  io.json.ToJSON.peakmem_to_json_wide('index', 'df_int_float_str')
-            193M             189M     0.98  io.json.ToJSON.peakmem_to_json_wide('records', 'df_int_float_str')
-       169±0.7ms        165±0.7ms     0.98  index_cached_properties.IndexCache.time_is_monotonic_increasing('IntervalIndex')
-     16.8±0.06ms      16.4±0.09ms     0.98  join_merge.MergeAsof.time_on_uint64('backward', 5)
-            187M             183M     0.98  io.json.ToJSON.peakmem_to_json_wide('records', 'df_td_int_ts')
-            187M             183M     0.98  io.json.ToJSON.peakmem_to_json_wide('index', 'df_td_int_ts')
-         140±1μs          137±1μs     0.97  tslibs.offsets.OffestDatetimeArithmetic.time_subtract_10(<CustomBusinessMonthEnd>)
-       193±0.5ms        188±0.4ms     0.97  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(1000000, 2000)
-         434±4ns        423±0.7ns     0.97  tslibs.timestamp.TimestampProperties.time_dayofyear(tzfile('/usr/share/zoneinfo/US/Central'), 'B')
-      55.7±0.3ms       54.2±0.2ms     0.97  strings.Methods.time_match
-            181M             176M     0.97  io.json.ToJSON.peakmem_to_json_wide('split', 'df_int_float_str')
-     6.28±0.08ms      6.12±0.02ms     0.97  hash_functions.IsinWithArange.time_isin(<class 'numpy.uint64'>, 2000, 0)
-            196M             191M     0.97  io.json.ToJSON.peakmem_to_json_wide('columns', 'df_date_idx')
-            116M             113M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'median')
-            181M             176M     0.97  io.json.ToJSON.peakmem_to_json_wide('values', 'df_int_float_str')
-            188M             183M     0.97  io.json.ToJSON.peakmem_to_json_wide('index', 'df')
-            188M             183M     0.97  io.json.ToJSON.peakmem_to_json_wide('records', 'df')
-            188M             183M     0.97  io.json.ToJSON.peakmem_to_json_wide('index', 'df_date_idx')
-            188M             183M     0.97  io.json.ToJSON.peakmem_to_json_wide('records', 'df_date_idx')
-            187M             181M     0.97  io.json.ToJSON.peakmem_to_json_wide('columns', 'df')
-            175M             170M     0.97  io.json.ToJSON.peakmem_to_json_wide('split', 'df_td_int_ts')
-            175M             170M     0.97  io.json.ToJSON.peakmem_to_json_wide('values', 'df_td_int_ts')
-     22.3±0.09ms       21.7±0.1ms     0.97  join_merge.MergeAsof.time_on_uint64('nearest', None)
-            115M             112M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'median')
-     4.60±0.05μs      4.47±0.02μs     0.97  categoricals.Contains.time_categorical_index_contains
-         380±3ns          370±2ns     0.97  tslibs.timestamp.TimestampOps.time_tz_localize(tzutc())
-       194±0.9ms        188±0.1ms     0.97  tslibs.period.TimePeriodArrToDT64Arr.time_periodarray_to_dt64arr(1000000, 2011)
-     6.08±0.02ms      5.89±0.03ms     0.97  tslibs.offsets.OnOffset.time_on_offset(<CustomBusinessMonthBegin>)
-      66.3±0.5ms       64.3±0.1ms     0.97  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0.5, 'midpoint')
-      15.2±0.2μs       14.8±0.2μs     0.97  tslibs.offsets.OffestDatetimeArithmetic.time_apply_np_dt64(<BYearBegin: month=1>)
-       178±0.8μs        172±0.6μs     0.97  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<CustomBusinessMonthBegin>)
-            177M             172M     0.97  io.json.ToJSON.peakmem_to_json_wide('split', 'df')
-            177M             172M     0.97  io.json.ToJSON.peakmem_to_json_wide('values', 'df')
-            177M             172M     0.97  io.json.ToJSON.peakmem_to_json_wide('split', 'df_date_idx')
-      1.79±0.01s       1.73±0.01s     0.97  groupby.GroupByMethods.time_dtype_as_field('float', 'describe', 'direct')
-            177M             172M     0.97  io.json.ToJSON.peakmem_to_json_wide('values', 'df_date_idx')
-     4.75±0.02ms      4.61±0.02ms     0.97  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(';', '_', 'high')
-         310±4ms          301±1ms     0.97  arithmetic.OffsetArrayArithmetic.time_add_dti_offset(<CustomBusinessDay>)
-         317±4ms          307±2ms     0.97  arithmetic.OffsetArrayArithmetic.time_add_series_offset(<CustomBusinessDay>)
-         196±2μs        189±0.9μs     0.97  tslibs.offsets.OffestDatetimeArithmetic.time_add_10(<CustomBusinessMonthBegin>)
-      5.05±0.7ms      4.89±0.03ms     0.97  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'int', 'skew')
-     20.7±0.08μs      20.0±0.06μs     0.97  tslibs.resolution.TimeResolution.time_get_resolution('h', 100, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'median')
-     3.31±0.01ms      3.20±0.02ms     0.97  arithmetic.ApplyIndex.time_apply_index(<BusinessDay>)
-      98.8±0.5ms       95.6±0.6ms     0.97  rolling.Quantile.time_quantile('Series', 1000, 'float', 0.5, 'linear')
-      56.2±0.5ms       54.4±0.1ms     0.97  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0.5, 'lower')
-     4.78±0.02ms      4.62±0.02ms     0.97  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(';', '_', None)
-       138±0.4ms        133±0.4ms     0.97  hash_functions.IsinWithArange.time_isin(<class 'object'>, 8000, 2)
-      10.2±0.2μs       9.82±0.1μs     0.97  tslibs.timestamp.TimestampProperties.time_weekday_name(datetime.timezone(datetime.timedelta(seconds=3600)), None)
-      91.6±0.4ms       88.5±0.4ms     0.97  rolling.Quantile.time_quantile('Series', 1000, 'float', 0.5, 'nearest')
-            117M             113M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'kurt')
-            116M             112M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'kurt')
-            204M             197M     0.97  io.json.ToJSON.peakmem_to_json_wide('columns', 'df_int_floats')
-            117M             113M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'median')
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'max')
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'min')
-            116M             112M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'mean')
-            116M             112M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'int', 'sum')
-            116M             112M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'skew')
-            116M             112M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'sum')
-            116M             112M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'median')
-      1.79±0.04s          1.73±0s     0.97  groupby.GroupByMethods.time_dtype_as_field('float', 'describe', 'transformation')
-            117M             113M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'skew')
-       372±0.7ms          359±2ms     0.97  series_methods.IsInLongSeriesValuesDominate.time_isin('int64', 'monotone')
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'kurt')
-     2.01±0.01μs      1.94±0.02μs     0.97  tslibs.resolution.TimeResolution.time_get_resolution('ns', 1, None)
-      11.1±0.5μs       10.7±0.3μs     0.97  index_cached_properties.IndexCache.time_engine('Float64Index')
-            117M             113M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'std')
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'mean')
-            114M             110M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'max')
-      94.6±0.3ms       91.3±0.3ms     0.97  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'median')
-      29.2±0.1ms       28.2±0.2ms     0.97  io.csv.ToCSV.time_frame('mixed')
-            118M             114M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'median')
-            114M             110M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'min')
-            116M             112M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'kurt')
-            116M             112M     0.97  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'mean')
-            115M             111M     0.97  rolling.ForwardWindowMethods.peakmem_rolling('Series', 10, 'float', 'sum')
-            117M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'mean')
-            194M             187M     0.96  io.json.ToJSON.peakmem_to_json_wide('records', 'df_int_floats')
-            194M             187M     0.96  io.json.ToJSON.peakmem_to_json_wide('index', 'df_int_floats')
-            118M             114M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'std')
-     5.78±0.03μs      5.58±0.03μs     0.96  dtypes.InferDtypes.time_infer('np-floating')
-      11.4±0.1μs      11.0±0.08μs     0.96  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('timestamp', 0, datetime.timezone(datetime.timedelta(seconds=3600)))
-     4.80±0.02ms      4.63±0.01ms     0.96  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '_', 'high')
-            117M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'sum')
-      72.3±0.4ms       69.7±0.5ms     0.96  io.csv.ToCSVDatetimeBig.time_frame(10000)
-      2.06±0.01s       1.99±0.01s     0.96  frame_methods.Iteration.time_itertuples_to_list
-      15.4±0.1μs      14.8±0.05μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_apply_np_dt64(<YearBegin: month=1>)
-            117M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'min')
-      43.1±0.3ms       41.5±0.3ms     0.96  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'median')
-      98.7±0.3ms       95.0±0.5ms     0.96  rolling.Quantile.time_quantile('Series', 1000, 'float', 0.5, 'midpoint')
-         161±1μs          155±1μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_subtract_10(<CustomBusinessMonthBegin>)
-            117M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'max')
-      3.80±0.2μs       3.66±0.2μs     0.96  index_cached_properties.IndexCache.time_inferred_type('UInt64Index')
-         407±3ms        392±0.3ms     0.96  index_object.Indexing.time_get_loc('String')
-         129±2ms        124±0.6ms     0.96  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'median')
-      91.9±0.1ms       88.4±0.4ms     0.96  rolling.Quantile.time_quantile('Series', 1000, 'float', 0.5, 'lower')
-      2.96±0.8ms      2.85±0.02ms     0.96  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0, 'linear')
-      44.6±0.2ms       42.9±0.5ms     0.96  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'median')
-            138M             132M     0.96  io.pickle.Pickle.peakmem_write_pickle
-            182M             175M     0.96  io.json.ToJSON.peakmem_to_json_wide('split', 'df_int_floats')
-            116M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'max')
-            116M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'min')
-            182M             175M     0.96  io.json.ToJSON.peakmem_to_json_wide('values', 'df_int_floats')
-            116M             111M     0.96  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'kurt')
-      56.6±0.4ms       54.4±0.5ms     0.96  reshape.Cut.time_cut_int(1000)
-      92.2±0.3ms       88.5±0.2ms     0.96  rolling.Quantile.time_quantile('Series', 1000, 'float', 0.5, 'higher')
-     15.2±0.06μs       14.6±0.3μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_add(<BusinessDay>)
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'median')
-     7.48±0.04μs       7.18±0.1μs     0.96  tslibs.normalize.Normalize.time_is_date_array_normalized(0, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-         179±1μs          171±1μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_apply_np_dt64(<CustomBusinessMonthBegin>)
-     9.99±0.04μs      9.58±0.09μs     0.96  tslibs.timestamp.TimestampProperties.time_month_name(<UTC>, None)
-       116±0.3μs        111±0.9μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<CustomBusinessMonthEnd>)
-      10.2±0.1μs      9.81±0.08μs     0.96  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 6000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      2.96±0.8ms      2.84±0.01ms     0.96  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0, 'nearest')
-      1.61±0.01s       1.55±0.02s     0.96  frame_methods.Iteration.time_itertuples
-         186±1μs          179±1μs     0.96  stat_ops.SeriesOps.time_op('mean', 'int')
-     9.93±0.08μs      9.53±0.03μs     0.96  tslibs.timestamp.TimestampProperties.time_month_name(None, 'B')
-      10.3±0.1μs      9.86±0.08μs     0.96  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 1000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            145M             139M     0.96  io.pickle.Pickle.peakmem_read_pickle
-            117M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'skew')
-            117M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'sum')
-      3.70±0.2μs       3.55±0.2μs     0.96  index_cached_properties.IndexCache.time_values('UInt64Index')
-       166±0.4μs        159±0.4μs     0.96  series_methods.NanOps.time_func('median', 1000, 'Int64')
-     4.21±0.01ms      4.04±0.02ms     0.96  io.csv.ReadCSVCachedParseDates.time_read_csv_cached(True, 'python')
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'std')
-            117M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'mean')
-            117M             112M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'kurt')
-       164±0.8μs        157±0.7μs     0.96  period.DataFramePeriodColumn.time_setitem_period_column
-            119M             114M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'std')
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'kurt')
-            114M             109M     0.96  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'max')
-         185±3μs          177±1μs     0.96  tslibs.offsets.OffestDatetimeArithmetic.time_subtract(<CustomBusinessMonthEnd>)
-         250±2μs          239±1μs     0.96  series_methods.NanOps.time_func('sem', 1000, 'int32')
-            115M             110M     0.96  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'min')
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'skew')
-            115M             110M     0.96  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'max')
-            114M             109M     0.96  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'min')
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'mean')
-         252±1μs          241±3μs     0.96  series_methods.NanOps.time_func('sem', 1000, 'int8')
-            119M             114M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'median')
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'sum')
-     10.3±0.08μs       9.81±0.1μs     0.96  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 4006, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'min')
-     10.2±0.06μs       9.77±0.1μs     0.96  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 3000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            118M             113M     0.96  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'max')
-     10.2±0.08μs       9.76±0.2μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 2000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      73.9±0.2ms       70.6±0.3ms     0.95  index_object.SetOperations.time_operation('strings', 'intersection')
-      10.2±0.2μs      9.73±0.04μs     0.95  tslibs.timestamp.TimestampProperties.time_weekday_name(tzlocal(), None)
-     3.78±0.06μs      3.61±0.03μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 2000, None)
-     4.95±0.04μs      4.72±0.03μs     0.95  tslibs.timedelta.TimedeltaConstructor.time_from_np_timedelta
-     10.5±0.06μs      10.0±0.09μs     0.95  tslibs.timestamp.TimestampProperties.time_month_name(tzfile('/usr/share/zoneinfo/US/Central'), 'B')
-      10.2±0.1μs       9.75±0.1μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 2011, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            115M             110M     0.95  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'kurt')
-     3.49±0.02μs      3.32±0.09μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 10000, datetime.timezone.utc)
-      91.0±0.6ms       86.8±0.4ms     0.95  replace.Convert.time_replace('Series', 'Timedelta')
-            116M             110M     0.95  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'mean')
-        443±10ns          422±2ns     0.95  tslibs.timestamp.TimestampProperties.time_dayofyear(tzfile('/usr/share/zoneinfo/US/Central'), None)
-            116M             110M     0.95  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'int', 'sum')
-            115M             109M     0.95  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'mean')
-       101±0.6ms       95.9±0.3ms     0.95  rolling.Methods.time_rolling('Series', 1000, 'float', 'median')
-      10.2±0.1μs      9.76±0.05μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 5000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     5.71±0.05ms      5.44±0.03ms     0.95  rolling.Engine.time_rolling_apply('Series', 'float', <function sum at 0x7f2790e97430>, 'cython')
-            115M             109M     0.95  rolling.ForwardWindowMethods.peakmem_rolling('Series', 1000, 'float', 'sum')
-      3.00±0.8ms      2.86±0.02ms     0.95  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 0, 'midpoint')
-      95.6±0.3ms       91.0±0.2ms     0.95  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'median')
-      3.02±0.8ms      2.87±0.03ms     0.95  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 1, 'nearest')
-     2.03±0.03μs      1.93±0.02μs     0.95  tslibs.resolution.TimeResolution.time_get_resolution('h', 1, None)
-      10.9±0.1μs       10.3±0.1μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 9000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     10.2±0.07μs       9.75±0.1μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 3000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-            120M             114M     0.95  rolling.Methods.peakmem_rolling('Series', 1000, 'float', 'count')
-            117M             111M     0.95  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'max')
-     8.15±0.02ms      7.75±0.03ms     0.95  io.csv.ToCSVDatetimeBig.time_frame(1000)
-            117M             111M     0.95  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'min')
-     10.1±0.08μs      9.60±0.04μs     0.95  tslibs.timestamp.TimestampProperties.time_weekday_name(tzfile('/usr/share/zoneinfo/US/Central'), 'B')
-      16.7±0.4ms       15.9±0.8ms     0.95  stat_ops.FrameOps.time_op('std', 'float', 1)
-      15.3±0.1μs       14.5±0.3μs     0.95  tslibs.offsets.OffestDatetimeArithmetic.time_apply_np_dt64(<BusinessMonthBegin>)
-      11.6±0.2μs       11.0±0.1μs     0.95  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('time', 1, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     5.69±0.04ms      5.41±0.03ms     0.95  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '.', None)
-     1.59±0.01ms      1.51±0.01ms     0.95  io.parsers.ConcatDateCols.time_check_concat(1234567890, 1)
-      38.4±0.3ms       36.5±0.5ms     0.95  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'median')
-     10.7±0.04μs       10.2±0.3μs     0.95  tslibs.resolution.TimeResolution.time_get_resolution('us', 1, datetime.timezone(datetime.timedelta(seconds=3600)))
-     10.5±0.04μs      9.96±0.09μs     0.95  tslibs.timestamp.TimestampProperties.time_month_name(datetime.timezone(datetime.timedelta(seconds=3600)), 'B')
-         287±3μs          273±4μs     0.95  series_methods.NanOps.time_func('sem', 1000, 'boolean')
-       125±0.9μs          119±2μs     0.95  series_methods.NanOps.time_func('skew', 1000, 'float64')
-     8.96±0.05μs       8.50±0.2μs     0.95  tslibs.resolution.TimeResolution.time_get_resolution('ns', 1, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     1.69±0.02μs      1.61±0.01μs     0.95  attrs_caching.SeriesArrayAttribute.time_array('datetime64')
-     4.32±0.01ms      4.10±0.06ms     0.95  rolling.Apply.time_rolling('Series', 300, 'int', <function sum at 0x7f2790e97430>, True)
-     2.43±0.01ms      2.31±0.02ms     0.95  rolling.ExpandingMethods.time_expanding('DataFrame', 'int', 'mean')
-      3.03±0.8ms      2.88±0.01ms     0.95  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 1, 'lower')
-     11.0±0.05μs       10.4±0.1μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 2011, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      69.6±0.1ms       66.0±0.3ms     0.95  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'median')
-     5.73±0.07ms      5.44±0.04ms     0.95  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(';', '.', None)
-     5.73±0.04ms      5.44±0.03ms     0.95  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '.', 'high')
-      10.3±0.1μs       9.77±0.1μs     0.95  tslibs.normalize.Normalize.time_normalize_i8_timestamps(1, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      33.2±0.2μs       31.4±0.2μs     0.95  indexing.NumericSeriesIndexing.time_loc_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
-      10.8±0.1μs      10.2±0.06μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 7000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     5.73±0.04ms      5.43±0.02ms     0.95  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(';', '.', 'high')
-     8.19±0.04μs      7.76±0.07μs     0.95  series_methods.SearchSorted.time_searchsorted('float64')
-     2.60±0.02ms       2.46±0.1ms     0.95  rolling.Apply.time_rolling('Series', 3, 'float', <built-in function sum>, True)
-      95.1±0.7μs       90.0±0.6μs     0.95  indexing.NumericSeriesIndexing.time_getitem_scalar(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'unique_monotonic_inc')
-      44.0±0.3ms       41.7±0.2ms     0.95  rolling.Quantile.time_quantile('DataFrame', 10, 'float', 0.5, 'linear')
-         736±9ms          697±4ms     0.95  io.csv.ToCSVDatetimeBig.time_frame(100000)
-      15.2±0.3μs       14.4±0.3μs     0.95  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<BusinessMonthEnd>)
-            120M             114M     0.95  rolling.Methods.peakmem_rolling('Series', 1000, 'int', 'count')
-      99.3±0.3ms       94.0±0.6ms     0.95  rolling.Methods.time_rolling('Series', 1000, 'int', 'median')
-         119±2μs        112±0.5μs     0.95  tslibs.offsets.OffestDatetimeArithmetic.time_apply_np_dt64(<CustomBusinessMonthEnd>)
-     10.3±0.09μs       9.72±0.1μs     0.95  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 7000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     2.03±0.02μs      1.92±0.02μs     0.95  tslibs.resolution.TimeResolution.time_get_resolution('m', 1, None)
-     9.54±0.09μs       9.02±0.1μs     0.95  tslibs.resolution.TimeResolution.time_get_resolution('s', 1, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-       138±0.8μs        131±0.6μs     0.95  stat_ops.SeriesOps.time_op('sum', 'int')
-      9.73±0.2μs       9.20±0.2μs     0.95  tslibs.timestamp.TimestampProperties.time_weekday_name(<UTC>, None)
-       187±0.5μs          177±1μs     0.95  series_methods.NanOps.time_func('median', 1000, 'float64')
-      3.03±0.8ms      2.86±0.02ms     0.94  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 1, 'midpoint')
-     10.2±0.07μs      9.63±0.07μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 7000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     2.04±0.02μs      1.93±0.02μs     0.94  tslibs.resolution.TimeResolution.time_get_resolution('s', 1, None)
-      9.69±0.1μs      9.15±0.05μs     0.94  tslibs.timestamp.TimestampProperties.time_weekday_name(<UTC>, 'B')
-      3.02±0.8ms      2.86±0.02ms     0.94  rolling.Quantile.time_quantile('DataFrame', 1000, 'int', 1, 'higher')
-      43.4±0.6ms       41.0±0.3ms     0.94  rolling.Methods.time_rolling('DataFrame', 10, 'float', 'median')
-            121M             114M     0.94  rolling.Methods.peakmem_rolling('Series', 10, 'float', 'count')
-     10.9±0.06μs      10.3±0.09μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 1000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      5.05±0.7ms      4.77±0.02ms     0.94  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'int', 'skew')
-      37.9±0.3ms       35.7±0.3ms     0.94  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0.5, 'linear')
-      15.5±0.2μs       14.6±0.3μs     0.94  tslibs.offsets.OffestDatetimeArithmetic.time_add(<MonthEnd>)
-      15.1±0.2μs       14.2±0.1μs     0.94  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<MonthBegin>)
-      40.8±0.4ms       38.5±0.3ms     0.94  rolling.Quantile.time_quantile('DataFrame', 10, 'float', 0.5, 'nearest')
-     10.3±0.04μs       9.68±0.2μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 2011, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     10.9±0.08μs      10.3±0.05μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 12000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-            121M             114M     0.94  rolling.Methods.peakmem_rolling('Series', 10, 'int', 'count')
-     10.2±0.06μs      9.62±0.07μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 11000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     6.10±0.03ms      5.74±0.03ms     0.94  rolling.Engine.time_expanding_apply('Series', 'int', <function sum at 0x7f2790e97430>, 'cython')
-     9.64±0.06μs       9.08±0.2μs     0.94  tslibs.timestamp.TimestampProperties.time_weekday_name(None, None)
-     6.74±0.06ms      6.35±0.03ms     0.94  rolling.Engine.time_rolling_apply('Series', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'cython')
-      44.3±0.4ms       41.7±0.2ms     0.94  rolling.Quantile.time_quantile('DataFrame', 10, 'float', 0.5, 'midpoint')
-      5.76±0.1ms      5.42±0.02ms     0.94  io.csv.ReadCSVFloatPrecision.time_read_csv_python_engine(',', '.', 'round_trip')
-     3.26±0.02ms      3.07±0.02ms     0.94  io.csv.ReadCSVParseDates.time_multiple_date('python')
-     6.10±0.04ms      5.74±0.06ms     0.94  rolling.Engine.time_expanding_apply('Series', 'float', <function sum at 0x7f2790e97430>, 'cython')
-      38.0±0.4ms         35.8±1ms     0.94  stat_ops.FrameOps.time_op('median', 'float', 1)
-      21.1±0.3μs       19.9±0.3μs     0.94  indexing.NumericSeriesIndexing.time_getitem_scalar(<class 'pandas.core.indexes.numeric.Float64Index'>, 'unique_monotonic_inc')
-     5.10±0.06ms      4.80±0.09ms     0.94  ctors.SeriesConstructors.time_series_constructor(<function gen_of_str at 0x7f277fe8a310>, False, 'int')
-       117±0.9μs          110±1μs     0.94  series_methods.NanOps.time_func('var', 1000, 'float64')
-     1.84±0.05ms      1.73±0.01ms     0.94  reshape.Explode.time_explode(1000, 3)
-     3.86±0.02ms      3.63±0.03ms     0.94  index_cached_properties.IndexCache.time_is_unique('DatetimeIndex')
-     3.15±0.04μs      2.96±0.03μs     0.94  tslibs.timestamp.TimestampConstruction.time_parse_now
-        71.3±1ms         67.0±1ms     0.94  io.csv.ReadCSVCategorical.time_convert_post('c')
-      41.0±0.4ms       38.4±0.2ms     0.94  rolling.Quantile.time_quantile('DataFrame', 10, 'float', 0.5, 'lower')
-     1.36±0.02ms      1.28±0.01ms     0.94  arithmetic.NumericInferOps.time_divide(<class 'numpy.uint8'>)
-       110±0.6μs        103±0.9μs     0.94  indexing.NumericSeriesIndexing.time_loc_scalar(<class 'pandas.core.indexes.numeric.UInt64Index'>, 'unique_monotonic_inc')
-     10.9±0.03μs      10.2±0.09μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 3000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      9.91±0.1μs       9.29±0.1μs     0.94  tslibs.timestamp.TimestampProperties.time_weekday_name(tzutc(), 'B')
-      9.88±0.1μs       9.26±0.2μs     0.94  tslibs.timestamp.TimestampProperties.time_weekday_name(tzutc(), None)
-     3.85±0.06μs      3.61±0.01μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 1011, None)
-     2.57±0.04ms       2.41±0.1ms     0.94  rolling.Apply.time_rolling('Series', 3, 'int', <built-in function sum>, True)
-     11.0±0.05μs       10.3±0.2μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 2000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      41.6±0.5ms       38.9±0.3ms     0.94  rolling.Quantile.time_quantile('DataFrame', 10, 'float', 0.5, 'higher')
-       143±0.7ms        134±0.2ms     0.94  rolling.Apply.time_rolling('Series', 3, 'float', <function sum at 0x7f2790e97430>, False)
-      10.4±0.2μs      9.69±0.03μs     0.94  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 4000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      10.3±0.1μs       9.61±0.2μs     0.93  tslibs.timestamp.TimestampProperties.time_weekday_name(datetime.timezone(datetime.timedelta(seconds=3600)), 'B')
-      10.8±0.2μs       10.1±0.2μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 4000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      16.5±0.4ms       15.4±0.1ms     0.93  reshape.PivotTable.time_pivot_table_categorical
-       143±0.8ms        134±0.7ms     0.93  rolling.Apply.time_rolling('Series', 3, 'int', <function sum at 0x7f2790e97430>, False)
-     1.37±0.01ms      1.28±0.02ms     0.93  arithmetic.NumericInferOps.time_divide(<class 'numpy.int8'>)
-       146±0.6ms          136±1ms     0.93  rolling.Apply.time_rolling('DataFrame', 3, 'float', <function Apply.<lambda> at 0x7f277d8de160>, False)
-       146±0.8ms        136±0.9ms     0.93  rolling.Apply.time_rolling('DataFrame', 3, 'int', <function Apply.<lambda> at 0x7f277d8de160>, False)
-       145±0.3ms          135±1ms     0.93  rolling.Apply.time_rolling('Series', 3, 'int', <function Apply.<lambda> at 0x7f277d8de160>, False)
-     1.62±0.01ms      1.51±0.01ms     0.93  series_methods.ValueCounts.time_value_counts('int')
-      10.8±0.1μs       10.1±0.2μs     0.93  tslibs.resolution.TimeResolution.time_get_resolution('s', 1, datetime.timezone(datetime.timedelta(seconds=3600)))
-      37.9±0.6ms       35.4±0.2ms     0.93  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0.5, 'midpoint')
-         184±3ms          172±1ms     0.93  io.json.ToJSON.time_to_json('index', 'df_int_floats')
-     1.81±0.01ms      1.69±0.03ms     0.93  frame_methods.GetDtypeCounts.time_frame_get_dtype_counts
-      9.19±0.2μs      8.57±0.09μs     0.93  tslibs.normalize.Normalize.time_is_date_array_normalized(1, datetime.timezone(datetime.timedelta(seconds=3600)))
-      29.2±0.3μs       27.2±0.3μs     0.93  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<CustomBusinessDay>)
-      44.3±0.1ms       41.3±0.2ms     0.93  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'median')
-      10.9±0.2μs      10.2±0.09μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 6000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-       150±0.8ms          140±1ms     0.93  groupby.TransformEngine.time_dataframe_cython(True)
-      21.5±0.3μs       20.1±0.6μs     0.93  series_methods.SearchSorted.time_searchsorted('uint16')
-      16.4±0.1ms       15.2±0.3ms     0.93  reshape.Explode.time_explode(10000, 3)
-     9.54±0.05μs       8.89±0.1μs     0.93  tslibs.resolution.TimeResolution.time_get_resolution('D', 1, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     10.1±0.07μs       9.41±0.2μs     0.93  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('timestamp', 0, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     11.0±0.09μs       10.2±0.2μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 1000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     9.04±0.09μs       8.41±0.1μs     0.93  tslibs.resolution.TimeResolution.time_get_resolution('s', 1, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      7.06±0.2ms      6.56±0.03ms     0.93  stat_ops.FrameOps.time_op('median', 'Int64', 0)
-     8.97±0.07μs       8.34±0.2μs     0.93  tslibs.resolution.TimeResolution.time_get_resolution('us', 1, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-     3.15±0.02μs      2.93±0.08μs     0.93  tslibs.tz_convert.TimeTZConvert.time_tz_convert_from_utc(100, datetime.timezone.utc)
-     5.34±0.05ms      4.96±0.08ms     0.93  ctors.SeriesConstructors.time_series_constructor(<function gen_of_str at 0x7f277fe8a310>, False, 'float')
-         103±1ms       96.0±0.5ms     0.93  rolling.Apply.time_rolling('Series', 300, 'int', <function Apply.<lambda> at 0x7f277d8de160>, False)
-      35.4±0.2ms       32.9±0.2ms     0.93  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0.5, 'higher')
-      9.97±0.1ms      9.25±0.04ms     0.93  series_methods.NanOps.time_func('std', 1000000, 'Int64')
-      4.19±0.6ms      3.88±0.02ms     0.93  rolling.ExpandingMethods.time_expanding('DataFrame', 'float', 'min')
-     11.0±0.09μs       10.2±0.2μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 1011, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-      21.4±0.2μs       19.8±0.4μs     0.93  series_methods.SearchSorted.time_searchsorted('uint8')
-      10.4±0.1μs       9.67±0.1μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 9000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      58.3±0.2ms       54.0±0.3ms     0.93  rolling.Quantile.time_quantile('Series', 1000, 'int', 0.5, 'higher')
-     2.24±0.01ms      2.08±0.01ms     0.93  series_methods.NSort.time_nlargest('last')
-      11.0±0.2μs       10.2±0.1μs     0.93  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(1, 4006, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-         349±3μs          323±2μs     0.93  join_merge.Concat.time_concat_empty_left(0)
-     6.83±0.06ms      6.32±0.05ms     0.93  rolling.Apply.time_rolling('Series', 3, 'float', <function Apply.<lambda> at 0x7f277d8de160>, True)
-     1.62±0.01ms      1.50±0.01ms     0.93  series_methods.NSort.time_nsmallest('last')
-      10.6±0.1μs       9.78±0.1μs     0.93  tslibs.timestamp.TimestampOps.time_tz_convert(tzfile('/usr/share/zoneinfo/US/Central'))
-       105±0.5ms         96.7±1ms     0.93  rolling.Apply.time_rolling('DataFrame', 300, 'float', <function Apply.<lambda> at 0x7f277d8de160>, False)
-       143±0.3ms          133±1ms     0.92  rolling.Apply.time_rolling('DataFrame', 3, 'int', <function sum at 0x7f2790e97430>, False)
-     25.4±0.09ms       23.4±0.1ms     0.92  io.csv.ReadCSVSkipRows.time_skipprows(None, 'c')
-      10.4±0.1μs       9.60±0.2μs     0.92  index_object.Indexing.time_get_loc_sorted('Float')
-      10.5±0.3μs       9.72±0.1μs     0.92  tslibs.period.TimeDT64ArrToPeriodArr.time_dt64arr_to_periodarr(0, 2000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      4.39±0.2ms      4.06±0.03ms     0.92  index_cached_properties.IndexCache.time_is_unique('IntervalIndex')
-      68.5±0.3ms       63.3±0.4ms     0.92  rolling.Quantile.time_quantile('Series', 1000, 'int', 0.5, 'linear')
-       150±0.8ms          138±1ms     0.92  groupby.TransformEngine.time_dataframe_cython(False)
-      50.2±0.4μs       46.3±0.5μs     0.92  array.IntegerArray.time_from_integer_array
-       103±0.6ms         95.1±1ms     0.92  rolling.Apply.time_rolling('DataFrame', 300, 'int', <function sum at 0x7f2790e97430>, False)
-      68.8±0.5ms       63.4±0.3ms     0.92  rolling.Quantile.time_quantile('Series', 1000, 'int', 0.5, 'midpoint')
-      97.8±0.8μs         90.2±1μs     0.92  series_methods.Any.time_any(1000000, 'slow', 'bool')
-      58.4±0.2ms       53.8±0.3ms     0.92  rolling.Quantile.time_quantile('Series', 1000, 'int', 0.5, 'nearest')
-      6.63±0.1μs       6.11±0.1μs     0.92  tslibs.timedelta.TimedeltaConstructor.time_from_datetime_timedelta
-     10.5±0.07μs       9.70±0.1μs     0.92  index_object.Indexing.time_get_loc('Float')
-         757±5μs          697±3μs     0.92  io.parsers.ConcatDateCols.time_check_concat('AAAA', 1)
-      15.4±0.4μs       14.2±0.1μs     0.92  tslibs.offsets.OffestDatetimeArithmetic.time_apply(<BusinessMonthBegin>)
-      44.3±0.5ms       40.7±0.2ms     0.92  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'median')
-       168±0.9ms          155±2ms     0.92  io.json.ToJSON.time_to_json_wide('columns', 'df_date_idx')
-      58.4±0.2ms       53.7±0.2ms     0.92  rolling.Quantile.time_quantile('Series', 1000, 'int', 0.5, 'lower')
-        49.7±2ms         45.6±1ms     0.92  stat_ops.FrameOps.time_op('mean', 'Int64', 1)
-       144±0.7ms        132±0.5ms     0.92  rolling.Apply.time_rolling('DataFrame', 3, 'float', <function sum at 0x7f2790e97430>, False)
-      9.29±0.1μs      8.52±0.05μs     0.92  tslibs.normalize.Normalize.time_is_date_array_normalized(0, datetime.timezone(datetime.timedelta(seconds=3600)))
-      56.0±0.1ms       51.3±0.4ms     0.92  frame_methods.Dropna.time_dropna_axis_mixed_dtypes('all', 1)
-     1.96±0.02μs      1.80±0.01μs     0.92  timedelta.TimedeltaIndexing.time_shallow_copy
-     3.13±0.01ms      2.86±0.04ms     0.92  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'int', 'mean')
-     3.11±0.05μs       2.85±0.1μs     0.91  tslibs.tz_convert.TimeTZConvert.time_tz_convert_from_utc(1, datetime.timezone.utc)
-      21.0±0.3ms       19.2±0.3ms     0.91  timeseries.Iteration.time_iter_preexit(<function period_range at 0x7f2782af4700>)
-      10.9±0.2μs       9.96±0.2μs     0.91  tslibs.tslib.TimeIntsToPydatetime.time_ints_to_pydatetime('datetime', 0, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-       170±0.5μs        155±0.9μs     0.91  period.Algorithms.time_drop_duplicates('index')
-      44.3±0.4ms       40.4±0.3ms     0.91  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'median')
-       106±0.8ms       96.2±0.4ms     0.91  rolling.Apply.time_rolling('DataFrame', 300, 'int', <function Apply.<lambda> at 0x7f277d8de160>, False)
-         353±2μs        321±0.9μs     0.91  join_merge.Concat.time_concat_empty_right(0)
-     1.92±0.01ms      1.75±0.02ms     0.91  series_methods.NSort.time_nsmallest('first')
-     20.5±0.06ms      18.7±0.07ms     0.91  groupby.MultiColumn.time_col_select_numpy_sum
-         202±4ms          183±3ms     0.91  io.json.ToJSON.time_to_json_wide('records', 'df_int_floats')
-      90.2±0.3ms       82.0±0.1ms     0.91  hash_functions.IsinWithArange.time_isin(<class 'object'>, 1000, 2)
-      22.0±0.2ms       19.9±0.1ms     0.91  rolling.Pairwise.time_pairwise(10, 'corr', True)
-      58.2±0.5μs       52.8±0.7μs     0.91  series_methods.NanOps.time_func('prod', 1000, 'float64')
-     2.98±0.01ms      2.71±0.02ms     0.91  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'float', 'mean')
-     57.5±0.08μs       52.1±0.6μs     0.91  array.BooleanArray.time_from_integer_array
-      15.2±0.2ms       13.8±0.1ms     0.91  frame_methods.Apply.time_apply_lambda_mean
-      45.9±0.4ms       41.6±0.3ms     0.91  rolling.Quantile.time_quantile('Series', 10, 'float', 0.5, 'linear')
-      45.1±0.2ms       40.8±0.3ms     0.91  rolling.Methods.time_rolling('Series', 10, 'float', 'median')
-      4.28±0.6ms      3.87±0.02ms     0.90  rolling.ExpandingMethods.time_expanding('DataFrame', 'float', 'max')
-      52.6±0.1ms       47.6±0.1ms     0.90  strings.Methods.time_title
-      7.28±0.1ms       6.59±0.1ms     0.90  rolling.Apply.time_rolling('DataFrame', 3, 'float', <function Apply.<lambda> at 0x7f277d8de160>, True)
-      45.5±0.3ms       41.1±0.5ms     0.90  rolling.Methods.time_rolling('Series', 10, 'int', 'median')
-     2.17±0.01ms         1.96±0ms     0.90  series_methods.NSort.time_nlargest('first')
-      7.21±0.1ms      6.51±0.07ms     0.90  rolling.Engine.time_rolling_apply('DataFrame', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'cython')
-      16.7±0.2ms       15.1±0.2ms     0.90  frame_methods.Apply.time_apply_np_mean
-         151±2ms          137±3ms     0.90  io.json.ToJSON.time_to_json_wide('values', 'df')
-     3.00±0.02ms      2.71±0.01ms     0.90  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'mean')
-       387±0.7μs          349±2μs     0.90  timedelta.TimedeltaIndexing.time_intersection
-     3.03±0.08ms      2.74±0.03ms     0.90  arithmetic.FrameWithFrameWide.time_op_different_blocks(<built-in function add>)
-     7.56±0.06ms      6.82±0.03ms     0.90  rolling.Engine.time_expanding_apply('DataFrame', 'int', <function Engine.<lambda> at 0x7f277d8de310>, 'cython')
-     37.6±0.06ms       33.9±0.3ms     0.90  strings.Methods.time_lower
-     6.62±0.05ms      5.96±0.05ms     0.90  rolling.Engine.time_expanding_apply('DataFrame', 'float', <function sum at 0x7f2790e97430>, 'cython')
-     1.81±0.02ms      1.63±0.01ms     0.90  rolling.Engine.time_expanding_apply('Series', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'numba')
-      49.6±0.5ms       44.6±0.2ms     0.90  rolling.GroupbyLargeGroups.time_rolling_multiindex_creation
-     43.2±0.07ms       38.8±0.3ms     0.90  rolling.Quantile.time_quantile('Series', 10, 'float', 0.5, 'higher')
-      1.97±0.01s          1.77±0s     0.90  timeseries.Iteration.time_iter(<function period_range at 0x7f2782af4700>)
-        218±10μs          196±1μs     0.90  tslibs.tz_convert.TimeTZConvert.time_tz_convert_from_utc(10000, <DstTzInfo 'US/Pacific' LMT-1 day, 16:07:00 STD>)
-      42.6±0.2ms       38.3±0.3ms     0.90  rolling.Quantile.time_quantile('Series', 10, 'float', 0.5, 'nearest')
-     6.59±0.04ms      5.91±0.05ms     0.90  rolling.Engine.time_expanding_apply('DataFrame', 'int', <function sum at 0x7f2790e97430>, 'cython')
-     6.24±0.05ms      5.59±0.03ms     0.90  rolling.Engine.time_rolling_apply('DataFrame', 'float', <function sum at 0x7f2790e97430>, 'cython')
-     46.2±0.05ms       41.4±0.3ms     0.90  rolling.Quantile.time_quantile('Series', 10, 'float', 0.5, 'midpoint')
-     7.60±0.07ms      6.80±0.03ms     0.90  rolling.Engine.time_expanding_apply('DataFrame', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'cython')
-     2.82±0.01ms      2.52±0.01ms     0.89  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'int', 'sum')
-      42.7±0.1ms       38.1±0.3ms     0.89  rolling.Quantile.time_quantile('Series', 10, 'float', 0.5, 'lower')
-      6.24±0.1ms      5.57±0.05ms     0.89  rolling.Apply.time_rolling('DataFrame', 3, 'float', <function sum at 0x7f2790e97430>, True)
-      11.3±0.4μs       10.1±0.2μs     0.89  timeseries.TzLocalize.time_infer_dst('UTC')
-     7.69±0.08ms      6.84±0.02ms     0.89  rolling.Pairwise.time_pairwise(10, 'corr', False)
-     7.66±0.04ms      6.81±0.03ms     0.89  rolling.Pairwise.time_pairwise(1000, 'corr', False)
-       154±0.7μs        137±0.8μs     0.89  tslibs.tz_convert.TimeTZConvert.time_tz_convert_from_utc(10000, tzfile('/usr/share/zoneinfo/Asia/Tokyo'))
-     2.69±0.01ms      2.39±0.01ms     0.89  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'sum')
-     2.68±0.02ms      2.38±0.01ms     0.89  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'float', 'sum')
-      25.3±0.2ms       22.5±0.2ms     0.89  frame_methods.ToNumpy.time_to_numpy_mixed_tall
-     14.7±0.07ms       13.1±0.2ms     0.89  groupby.Categories.time_groupby_nosort
-         204±5ms          181±4ms     0.89  io.json.ToJSON.time_to_json_wide('index', 'df_int_floats')
-         177±3ms          157±1ms     0.89  io.csv.ToCSV.time_frame('wide')
-      39.7±0.2ms       35.2±0.4ms     0.89  rolling.Quantile.time_quantile('Series', 10, 'int', 0.5, 'linear')
-     3.29±0.01ms      2.91±0.05ms     0.89  rolling.EWMMethods.time_ewm('DataFrame', 1000, 'float', 'std')
-     1.83±0.05ms      1.62±0.04ms     0.89  rolling.Engine.time_expanding_apply('Series', 'float', <function sum at 0x7f2790e97430>, 'numba')
-     22.1±0.09ms      19.6±0.06ms     0.89  rolling.Pairwise.time_pairwise(None, 'corr', True)
-      39.8±0.3ms       35.2±0.2ms     0.88  rolling.Quantile.time_quantile('Series', 10, 'int', 0.5, 'midpoint')
-      79.3±0.8μs       70.2±0.4μs     0.88  series_methods.NanOps.time_func('mean', 1000, 'int64')
-         227±2ms          201±4ms     0.88  io.json.ToJSON.time_to_json_wide('index', 'df_int_float_str')
-     7.33±0.08ms      6.48±0.03ms     0.88  rolling.Engine.time_rolling_apply('DataFrame', 'int', <function Engine.<lambda> at 0x7f277d8de310>, 'cython')
-     4.86±0.01ms      4.29±0.04ms     0.88  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'float', 'min')
-         525±6μs         464±10μs     0.88  join_merge.Append.time_append_homogenous
-         170±1ms          150±1ms     0.88  io.json.ToJSON.time_to_json_wide('columns', 'df')
-     3.71±0.01ms      3.27±0.01ms     0.88  series_methods.NanOps.time_func('mean', 1000000, 'boolean')
-         181±3ms          160±2ms     0.88  io.json.ToJSON.time_to_json_wide('split', 'df_int_floats')
-     19.5±0.07ms       17.2±0.1ms     0.88  rolling.Pairwise.time_pairwise(None, 'cov', True)
-       125±0.3ms        111±0.2ms     0.88  gil.ParallelKth.time_kth_smallest
-      24.1±0.4ms       21.2±0.2ms     0.88  categoricals.Indexing.time_reindex
-     1.85±0.01ms      1.63±0.03ms     0.88  rolling.Engine.time_expanding_apply('Series', 'int', <function sum at 0x7f2790e97430>, 'numba')
-      79.3±0.9μs       69.8±0.7μs     0.88  series_methods.NanOps.time_func('mean', 1000, 'int32')
-      7.38±0.1ms      6.49±0.04ms     0.88  rolling.Apply.time_rolling('DataFrame', 3, 'int', <function Apply.<lambda> at 0x7f277d8de160>, True)
-         972±3μs          855±4μs     0.88  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 14)
-     4.77±0.05ms      4.20±0.05ms     0.88  rolling.Apply.time_rolling('DataFrame', 300, 'float', <function sum at 0x7f2790e97430>, True)
-         182±2ms          160±3ms     0.88  io.json.ToJSON.time_to_json_wide('values', 'df_int_floats')
-      36.7±0.4ms       32.2±0.2ms     0.88  rolling.Quantile.time_quantile('Series', 10, 'int', 0.5, 'nearest')
-      36.6±0.3ms       32.2±0.2ms     0.88  rolling.Quantile.time_quantile('Series', 10, 'int', 0.5, 'lower')
-      78.9±0.5μs       69.2±0.6μs     0.88  series_methods.NanOps.time_func('mean', 1000, 'int8')
-         556±3μs          487±3μs     0.88  join_merge.Concat.time_concat_mixed_ndims(0)
-     4.94±0.01ms      4.33±0.02ms     0.88  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'min')
-      37.0±0.2ms       32.4±0.2ms     0.88  rolling.Quantile.time_quantile('Series', 10, 'int', 0.5, 'higher')
-       120±0.6ms        105±0.1ms     0.88  replace.Convert.time_replace('DataFrame', 'Timedelta')
-         120±1ms        105±0.7ms     0.87  replace.Convert.time_replace('DataFrame', 'Timestamp')
-     5.50±0.05ms      4.81±0.02ms     0.87  rolling.Apply.time_rolling('DataFrame', 300, 'float', <function Apply.<lambda> at 0x7f277d8de160>, True)
-     19.6±0.06ms       17.1±0.1ms     0.87  rolling.Pairwise.time_pairwise(1000, 'cov', True)
-      66.5±0.2μs       58.1±0.6μs     0.87  series_methods.NanOps.time_func('min', 1000, 'int32')
-      6.29±0.1ms      5.48±0.04ms     0.87  rolling.Engine.time_rolling_apply('DataFrame', 'int', <function sum at 0x7f2790e97430>, 'cython')
-         230±6ms          200±3ms     0.87  io.json.ToJSON.time_to_json_wide('records', 'df_int_float_str')
-         208±4ms          181±2ms     0.87  io.json.ToJSON.time_to_json_wide('values', 'df_int_float_str')
-       113±0.7ms         97.9±1ms     0.87  index_object.IndexAppend.time_append_obj_list
-     5.03±0.01ms      4.36±0.01ms     0.87  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'float', 'max')
-         815±1ms          706±3ms     0.87  hash_functions.NumericSeriesIndexingShuffled.time_loc_slice(<class 'pandas.core.indexes.numeric.Float64Index'>, 5000000)
-     5.53±0.06ms      4.78±0.02ms     0.87  rolling.Apply.time_rolling('DataFrame', 300, 'int', <function Apply.<lambda> at 0x7f277d8de160>, True)
-         573±8μs          495±3μs     0.86  period.Indexing.time_intersection
-     5.49±0.04ms      4.75±0.03ms     0.86  stat_ops.FrameOps.time_op('std', 'Int64', 0)
-     4.81±0.07ms      4.15±0.04ms     0.86  rolling.Apply.time_rolling('DataFrame', 300, 'int', <function sum at 0x7f2790e97430>, True)
-       210±0.5ms          181±5ms     0.86  io.json.ToJSON.time_to_json_wide('columns', 'df_int_floats')
-     26.3±0.08ms      22.7±0.05ms     0.86  join_merge.Concat.time_concat_series(0)
-      66.6±0.4μs       57.3±0.4μs     0.86  series_methods.NanOps.time_func('min', 1000, 'int64')
-      5.98±0.7ms      5.14±0.03ms     0.86  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'kurt')
-      66.8±0.1μs       57.5±0.6μs     0.86  series_methods.NanOps.time_func('min', 1000, 'int8')
-      3.09±0.2ms      2.65±0.01ms     0.86  rolling.ExpandingMethods.time_expanding('DataFrame', 'float', 'std')
-     5.03±0.02ms      4.31±0.03ms     0.86  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'float', 'max')
-      68.0±0.7μs       58.1±0.2μs     0.85  replace.ReplaceList.time_replace_list(True)
-      55.4±0.5μs       47.3±0.3μs     0.85  series_methods.NanOps.time_func('sum', 1000, 'int64')
-      10.6±0.2μs      9.05±0.04μs     0.85  timeseries.TzLocalize.time_infer_dst(tzutc())
-      38.1±0.5ms       32.5±0.1ms     0.85  frame_methods.Dropna.time_dropna_axis_mixed_dtypes('any', 1)
-         188±6ms          160±5ms     0.85  indexing.CategoricalIndexIndexing.time_get_indexer_list('monotonic_incr')
-      6.01±0.7ms      5.12±0.04ms     0.85  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'kurt')
-     3.50±0.02ms      2.98±0.02ms     0.85  series_methods.IsIn.time_isin('object')
-     2.61±0.02ms      2.22±0.02ms     0.85  stat_ops.FrameOps.time_op('mean', 'Int64', 0)
-        734±30μs          623±4μs     0.85  indexing.NonNumericSeriesIndexing.time_getitem_list_like('datetime', 'non_monotonic')
-     6.20±0.06ms      5.26±0.03ms     0.85  gil.ParallelRolling.time_rolling('std')
-     5.57±0.04ms      4.71±0.01ms     0.85  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'int', 'min')
-     5.62±0.02ms      4.75±0.02ms     0.85  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'int', 'min')
-       218±0.4ms          184±2ms     0.84  io.json.ToJSON.time_to_json_wide('records', 'df_td_int_ts')
-     5.57±0.01ms      4.70±0.02ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'min')
-         204±2ms          172±4ms     0.84  io.json.ToJSON.time_to_json_wide('columns', 'df_td_int_ts')
-         218±1ms          183±2ms     0.84  io.json.ToJSON.time_to_json_wide('index', 'df_td_int_ts')
-      5.89±0.7ms         4.96±0ms     0.84  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'kurt')
-     3.74±0.05ms      3.15±0.02ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'mean')
-         187±1ms          157±1ms     0.84  io.json.ToJSON.time_to_json_wide('values', 'df_td_int_ts')
-     5.76±0.09ms      4.84±0.01ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'int', 'min')
-         210±2ms        177±0.8ms     0.84  io.json.ToJSON.time_to_json_wide('split', 'df_int_float_str')
-     5.67±0.09ms      4.76±0.01ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'min')
-     5.82±0.05ms      4.89±0.01ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'int', 'max')
-     5.47±0.02ms      4.60±0.01ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'min')
-     3.70±0.03ms      3.10±0.01ms     0.84  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'mean')
-         227±2ms          190±4ms     0.84  io.json.ToJSON.time_to_json_wide('columns', 'df_int_float_str')
-     22.3±0.08ms      18.6±0.07ms     0.84  join_merge.Merge.time_merge_dataframe_integer_2key(True)
-     3.95±0.02ms      3.29±0.02ms     0.83  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 16)
-      6.32±0.7ms      5.26±0.03ms     0.83  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'kurt')
-        750±60μs          625±6μs     0.83  indexing.NonNumericSeriesIndexing.time_getitem_list_like('datetime', 'unique_monotonic_inc')
-     5.73±0.06ms      4.78±0.02ms     0.83  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'max')
-         188±3ms          157±1ms     0.83  io.json.ToJSON.time_to_json_wide('split', 'df_td_int_ts')
-     3.66±0.01ms      3.05±0.01ms     0.83  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'mean')
-         144±1μs          120±1μs     0.83  series_methods.NanOps.time_func('std', 1000, 'int32')
-     21.4±0.04ms      17.8±0.06ms     0.83  stat_ops.Correlation.time_corr_wide_nans('pearson')
-     3.35±0.02ms      2.78±0.01ms     0.83  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'sum')
-       145±0.9μs          119±1μs     0.83  series_methods.NanOps.time_func('std', 1000, 'int8')
-     5.74±0.03ms      4.74±0.06ms     0.83  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'min')
-     7.73±0.07ms      6.39±0.05ms     0.83  gil.ParallelRolling.time_rolling('kurt')
-     3.38±0.02ms      2.79±0.01ms     0.82  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'sum')
-         194±7ms          160±5ms     0.82  indexing.CategoricalIndexIndexing.time_get_indexer_list('monotonic_decr')
-         146±1μs          121±1μs     0.82  series_methods.NanOps.time_func('std', 1000, 'int64')
-        95.4±1μs         78.4±1μs     0.82  series_methods.NanOps.time_func('mean', 1000, 'boolean')
-     1.96±0.04ms      1.61±0.02ms     0.82  arithmetic.FrameWithFrameWide.time_op_same_blocks(<built-in function add>)
-        82.9±1μs       68.1±0.6μs     0.82  series_methods.NanOps.time_func('mean', 1000, 'Int64')
-     5.77±0.03ms      4.74±0.01ms     0.82  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'max')
-      58.8±0.5μs       48.3±0.6μs     0.82  series_methods.NanOps.time_func('sum', 1000, 'int8')
-     3.41±0.01ms      2.79±0.02ms     0.82  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'sum')
-     5.74±0.03ms      4.70±0.01ms     0.82  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'int', 'max')
-     5.71±0.01ms         4.66±0ms     0.82  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'max')
-     6.73±0.08ms      5.48±0.06ms     0.82  ctors.SeriesConstructors.time_series_constructor(<function arr_dict at 0x7f277fe8a160>, True, 'int')
-     7.95±0.05ms      6.48±0.03ms     0.82  rolling.Pairwise.time_pairwise(None, 'corr', False)
-     5.79±0.03ms      4.72±0.01ms     0.82  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'int', 'max')
-     6.54±0.07ms      5.32±0.07ms     0.81  ctors.SeriesConstructors.time_series_constructor(<function arr_dict at 0x7f277fe8a160>, False, 'int')
-     5.27±0.02ms      4.29±0.03ms     0.81  rolling.Pairwise.time_pairwise(10, 'cov', False)
-        48.9±3μs       39.7±0.7μs     0.81  series_methods.Any.time_any(1000000, 'fast', 'bool')
-      68.5±0.6ms       55.6±0.4ms     0.81  groupby.TransformEngine.time_series_cython(False)
-         192±4ms          156±6ms     0.81  indexing.CategoricalIndexIndexing.time_get_indexer_list('non_monotonic')
-      6.27±0.8ms      5.09±0.03ms     0.81  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'kurt')
-      68.7±0.6ms       55.7±0.4ms     0.81  groupby.TransformEngine.time_series_cython(True)
-     5.32±0.08ms      4.31±0.08ms     0.81  rolling.Pairwise.time_pairwise(1000, 'cov', False)
-     3.12±0.04ms       2.52±0.1ms     0.81  rolling.Apply.time_rolling('DataFrame', 3, 'int', <built-in function sum>, True)
-     5.65±0.03ms      4.57±0.01ms     0.81  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'max')
-      10.1±0.1ms      8.15±0.03ms     0.81  stat_ops.FrameOps.time_op('median', 'float', 0)
-     3.12±0.06ms       2.52±0.1ms     0.81  rolling.Apply.time_rolling('DataFrame', 3, 'float', <built-in function sum>, True)
-     6.21±0.01ms      5.01±0.01ms     0.81  rolling.Methods.time_rolling('Series', 1000, 'float', 'kurt')
-         410±5μs          330±3μs     0.80  stat_ops.SeriesOps.time_op('mean', 'float')
-      49.0±0.9μs       39.4±0.3μs     0.80  series_methods.All.time_all(1000000, 'fast', 'bool')
-      66.6±0.6μs       53.2±0.7μs     0.80  series_methods.NanOps.time_func('max', 1000, 'int32')
-       100±0.9μs         79.8±1μs     0.79  series_methods.NanOps.time_func('mean', 1000, 'float64')
-         177±1ms        140±0.8ms     0.79  categoricals.Rank.time_rank_string
-        85.5±1ms       67.8±0.8ms     0.79  rolling.Groupby.time_rolling_offset('sum')
-         182±2μs          144±1μs     0.79  series_methods.NanOps.time_func('std', 1000, 'Int64')
-      66.6±0.6μs       52.8±0.1μs     0.79  series_methods.NanOps.time_func('max', 1000, 'int64')
-       177±0.7μs          140±1μs     0.79  series_methods.NanOps.time_func('std', 1000, 'float64')
-      1.03±0.02s          815±5ms     0.79  groupby.Apply.time_copy_overhead_single_col
-      66.6±0.7μs       52.7±0.3μs     0.79  series_methods.NanOps.time_func('max', 1000, 'int8')
-      85.5±0.3ms       67.6±0.1ms     0.79  rolling.Groupby.time_rolling_offset('kurt')
-      84.9±0.5ms       67.1±0.6ms     0.79  rolling.Groupby.time_rolling_offset('max')
-     9.05±0.02ms      7.14±0.03ms     0.79  join_merge.Merge.time_merge_dataframe_integer_2key(False)
-      6.64±0.7ms      5.24±0.03ms     0.79  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'kurt')
-         183±1μs        145±0.8μs     0.79  series_methods.NanOps.time_func('std', 1000, 'boolean')
-     2.31±0.04ms      1.82±0.07ms     0.79  rolling.Engine.time_expanding_apply('DataFrame', 'int', <function sum at 0x7f2790e97430>, 'numba')
-     2.11±0.01ms      1.66±0.03ms     0.79  rolling.EWMMethods.time_ewm('DataFrame', 1000, 'float', 'mean')
-        85.0±1ms       67.0±0.6ms     0.79  rolling.Groupby.time_rolling_offset('mean')
-      5.58±0.7ms      4.40±0.02ms     0.79  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'skew')
-     2.24±0.02ms      1.76±0.02ms     0.79  rolling.EWMMethods.time_ewm_times('DataFrame', 10, 'int', 'std')
-     2.12±0.01ms      1.66±0.02ms     0.79  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 15)
-      5.59±0.7ms      4.39±0.02ms     0.79  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'skew')
-        88.9±1ms       69.7±0.6ms     0.78  rolling.Groupby.time_rolling_offset('median')
-     3.92±0.02ms      3.07±0.05ms     0.78  series_methods.NanOps.time_func('mean', 1000000, 'Int64')
-      3.27±0.4ms      2.56±0.01ms     0.78  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'int', 'min')
-     5.50±0.05ms      4.31±0.02ms     0.78  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'std')
-      86.0±0.6ms       67.4±0.4ms     0.78  rolling.Groupby.time_rolling_offset('min')
-     2.46±0.01ms      1.92±0.03ms     0.78  groupby.RankWithTies.time_rank_ties('int64', 'average')
-     2.43±0.01ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('datetime64', 'average')
-     2.25±0.01ms      1.76±0.03ms     0.78  rolling.EWMMethods.time_ewm_times('DataFrame', 1000, 'int', 'mean')
-         151±2μs          118±1μs     0.78  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('period', 'unique_monotonic_inc')
-     2.44±0.02ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('datetime64', 'dense')
-     2.44±0.02ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('datetime64', 'min')
-     2.30±0.09ms      1.79±0.02ms     0.78  rolling.Engine.time_expanding_apply('DataFrame', 'int', <function Engine.<lambda> at 0x7f277d8de310>, 'numba')
-     2.45±0.01ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('int64', 'min')
-     2.25±0.01ms      1.75±0.02ms     0.78  rolling.EWMMethods.time_ewm_times('DataFrame', 1000, 'int', 'std')
-     2.44±0.02ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('datetime64', 'first')
-      34.6±0.3μs       26.9±0.6μs     0.78  series_methods.Any.time_any(1000, 'slow', 'bool')
-     2.44±0.01ms      1.90±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('int64', 'dense')
-     2.24±0.02ms      1.74±0.01ms     0.78  rolling.EWMMethods.time_ewm('DataFrame', 10, 'int', 'mean')
-     2.24±0.01ms      1.74±0.02ms     0.78  rolling.EWMMethods.time_ewm_times('DataFrame', 10, 'int', 'mean')
-     2.43±0.02ms      1.89±0.02ms     0.78  groupby.RankWithTies.time_rank_ties('float32', 'max')
-     2.41±0.02ms      1.86±0.01ms     0.78  groupby.RankWithTies.time_rank_ties('float64', 'dense')
-     2.12±0.03ms      1.64±0.01ms     0.78  rolling.EWMMethods.time_ewm_times('DataFrame', 10, 'float', 'mean')
-     2.44±0.03ms      1.89±0.02ms     0.77  groupby.RankWithTies.time_rank_ties('float32', 'min')
-     2.12±0.01ms      1.64±0.02ms     0.77  rolling.EWMMethods.time_ewm_times('DataFrame', 1000, 'float', 'std')
-     3.38±0.02ms      2.61±0.02ms     0.77  rolling.ForwardWindowMethods.time_rolling('DataFrame', 1000, 'int', 'max')
-     2.45±0.01ms      1.90±0.01ms     0.77  groupby.RankWithTies.time_rank_ties('int64', 'first')
-     2.32±0.07ms      1.79±0.03ms     0.77  rolling.Engine.time_expanding_apply('DataFrame', 'float', <function sum at 0x7f2790e97430>, 'numba')
-      11.0±0.4μs       8.51±0.1μs     0.77  timeseries.TzLocalize.time_infer_dst(None)
-        75.4±1ms       58.3±0.2ms     0.77  reshape.Cut.time_qcut_timedelta(1000)
-     2.48±0.04ms      1.91±0.02ms     0.77  groupby.RankWithTies.time_rank_ties('datetime64', 'max')
-     2.46±0.02ms      1.89±0.02ms     0.77  groupby.RankWithTies.time_rank_ties('float32', 'dense')
-     2.12±0.01ms      1.63±0.01ms     0.77  rolling.EWMMethods.time_ewm_times('DataFrame', 10, 'float', 'std')
-     5.44±0.03ms      4.18±0.04ms     0.77  rolling.Pairwise.time_pairwise(None, 'cov', False)
-     2.32±0.05ms      1.78±0.03ms     0.77  rolling.Engine.time_expanding_apply('DataFrame', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'numba')
-     2.42±0.02ms      1.86±0.01ms     0.77  groupby.RankWithTies.time_rank_ties('float64', 'average')
-     11.5±0.06μs       8.86±0.1μs     0.77  indexing.CategoricalIndexIndexing.time_get_loc_scalar('monotonic_incr')
-        86.4±1ms         66.4±1ms     0.77  series_methods.SeriesConstructor.time_constructor('dict')
-      5.87±0.8ms      4.50±0.01ms     0.77  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'skew')
-     2.45±0.02ms      1.88±0.02ms     0.77  groupby.RankWithTies.time_rank_ties('float32', 'average')
-     2.42±0.02ms         1.86±0ms     0.77  groupby.RankWithTies.time_rank_ties('float64', 'min')
-     2.46±0.02ms      1.88±0.01ms     0.77  groupby.RankWithTies.time_rank_ties('float32', 'first')
-        1.03±0ms          789±4μs     0.77  period.Algorithms.time_value_counts('index')
-     6.71±0.01ms      5.13±0.01ms     0.77  rolling.Methods.time_rolling('Series', 1000, 'int', 'kurt')
-     2.49±0.05ms      1.90±0.02ms     0.76  groupby.RankWithTies.time_rank_ties('int64', 'max')
-     2.43±0.02ms      1.86±0.01ms     0.76  groupby.RankWithTies.time_rank_ties('float64', 'max')
-     3.94±0.01ms      3.01±0.02ms     0.76  series_methods.NanOps.time_func('max', 1000000, 'float64')
-     2.45±0.01ms      1.87±0.01ms     0.76  groupby.RankWithTies.time_rank_ties('float64', 'first')
-      35.4±0.3μs       27.0±0.3μs     0.76  series_methods.All.time_all(1000, 'slow', 'bool')
-      3.39±0.4ms      2.58±0.02ms     0.76  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'int', 'max')
-      5.64±0.7ms      4.30±0.01ms     0.76  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'float', 'std')
-      34.6±0.5μs       26.4±0.2μs     0.76  series_methods.Any.time_any(1000, 'fast', 'bool')
-      4.42±0.1μs       3.36±0.1μs     0.76  index_cached_properties.IndexCache.time_shape('DatetimeIndex')
-      33.4±0.3ms       25.4±0.5ms     0.76  categoricals.Indexing.time_intersection
-       72.3±10ms       54.9±0.3ms     0.76  reshape.Cut.time_qcut_datetime(1000)
-      35.0±0.2μs       26.6±0.4μs     0.76  series_methods.All.time_all(1000, 'fast', 'bool')
-      8.87±0.2μs       6.73±0.1μs     0.76  index_cached_properties.IndexCache.time_engine('DatetimeIndex')
-      85.0±0.8μs       64.4±0.9μs     0.76  series_methods.NanOps.time_func('sum', 1000, 'float64')
-      5.94±0.8ms      4.50±0.01ms     0.76  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'skew')
-      3.41±0.02s       2.57±0.03s     0.76  groupby.Apply.time_copy_function_multi_col
-      3.35±0.8ms      2.53±0.01ms     0.76  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'int', 'sum')
-     3.92±0.02ms      2.95±0.04ms     0.75  series_methods.NanOps.time_func('mean', 1000000, 'float64')
-      5.84±0.7ms      4.39±0.01ms     0.75  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'skew')
-     6.66±0.01ms      5.01±0.01ms     0.75  rolling.Methods.time_rolling('Series', 10, 'float', 'kurt')
-      14.3±0.4μs       10.7±0.4μs     0.75  index_cached_properties.IndexCache.time_engine('TimedeltaIndex')
-         958±9μs          717±3μs     0.75  period.Algorithms.time_value_counts('series')
-      8.93±0.5μs       6.68±0.2μs     0.75  index_cached_properties.IndexCache.time_engine('PeriodIndex')
-      5.81±0.7ms      4.34±0.02ms     0.75  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'float', 'std')
-      2.19±0.1ms      1.64±0.02ms     0.75  rolling.EWMMethods.time_ewm_times('DataFrame', 1000, 'float', 'mean')
-      5.52±0.7ms      4.12±0.01ms     0.75  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'float', 'std')
-     4.20±0.05ms       3.12±0.2ms     0.74  stat_ops.FrameOps.time_op('prod', 'float', 0)
-      4.67±0.1μs       3.46±0.2μs     0.74  index_cached_properties.IndexCache.time_shape('PeriodIndex')
-         134±2μs       99.4±0.4μs     0.74  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('period', 'non_monotonic')
-         303±2μs          223±1μs     0.74  index_cached_properties.IndexCache.time_is_monotonic('Float64Index')
-     4.86±0.07ms      3.58±0.06ms     0.74  ctors.SeriesConstructors.time_series_constructor(<function arr_dict at 0x7f277fe8a160>, True, 'float')
-     4.68±0.06ms      3.43±0.05ms     0.73  ctors.SeriesConstructors.time_series_constructor(<function arr_dict at 0x7f277fe8a160>, False, 'float')
-      5.87±0.7ms      4.30±0.02ms     0.73  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'skew')
-         666±3μs        487±0.8μs     0.73  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 13)
-     7.84±0.08μs      5.73±0.04μs     0.73  categoricals.SearchSorted.time_categorical_index_contains
-      57.8±0.7ms       42.2±0.4ms     0.73  reshape.Cut.time_cut_datetime(1000)
-         304±1μs          222±2μs     0.73  index_cached_properties.IndexCache.time_is_monotonic_increasing('Float64Index')
-      3.92±0.6ms      2.86±0.05ms     0.73  rolling.ForwardWindowMethods.time_rolling('DataFrame', 10, 'int', 'mean')
-         303±1μs          221±1μs     0.73  index_cached_properties.IndexCache.time_is_monotonic_decreasing('Float64Index')
-      5.39±0.2ms       3.92±0.2ms     0.73  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'count')
-      5.51±0.1ms      4.00±0.03ms     0.73  rolling.ExpandingMethods.time_expanding('DataFrame', 'int', 'max')
-      5.44±0.4ms       3.95±0.2ms     0.73  rolling.Methods.time_rolling('DataFrame', 1000, 'float', 'count')
-      87.0±0.5μs       63.2±0.6μs     0.73  series_methods.NanOps.time_func('max', 1000, 'float64')
-      4.45±0.7ms      3.23±0.01ms     0.73  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'int', 'mean')
-      7.14±0.2μs      5.18±0.02μs     0.72  categoricals.SearchSorted.time_categorical_contains
-      4.38±0.7ms      3.17±0.01ms     0.72  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'int', 'mean')
-      87.6±0.4μs       63.3±0.7μs     0.72  series_methods.NanOps.time_func('min', 1000, 'float64')
-      4.48±0.7ms      3.23±0.01ms     0.72  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'int', 'mean')
-         112±1μs       80.5±0.7μs     0.72  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('period', 'non_monotonic')
-      5.43±0.1ms       3.91±0.2ms     0.72  rolling.Methods.time_rolling('DataFrame', 10, 'float', 'count')
-         299±1μs          215±1μs     0.72  index_cached_properties.IndexCache.time_is_monotonic_increasing('UInt64Index')
-         299±2μs          215±2μs     0.72  index_cached_properties.IndexCache.time_is_monotonic('UInt64Index')
-         112±2μs       80.2±0.5μs     0.72  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('period', 'nonunique_monotonic_inc')
-         299±2μs          214±1μs     0.72  index_cached_properties.IndexCache.time_is_monotonic_decreasing('UInt64Index')
-     7.18±0.01ms      5.14±0.02ms     0.72  rolling.Methods.time_rolling('Series', 10, 'int', 'kurt')
-      5.41±0.4ms       3.87±0.2ms     0.72  rolling.Methods.time_rolling('DataFrame', 1000, 'int', 'count')
-         111±2μs       79.3±0.2μs     0.71  indexing.NonNumericSeriesIndexing.time_getitem_pos_slice('period', 'unique_monotonic_inc')
-        63.6±1ms       45.3±0.1ms     0.71  reshape.Cut.time_cut_timedelta(1000)
-      4.12±0.7ms      2.92±0.01ms     0.71  rolling.VariableWindowMethods.time_rolling('DataFrame', '1h', 'int', 'sum')
-            196M             139M     0.71  rolling.PeakMemFixedWindowMinMax.peakmem_fixed('max')
-            196M             139M     0.71  rolling.PeakMemFixedWindowMinMax.peakmem_fixed('min')
-     6.04±0.02ms      4.28±0.02ms     0.71  rolling.Methods.time_rolling('Series', 1000, 'float', 'skew')
-      32.9±0.5ms         23.2±1ms     0.71  hash_functions.UniqueAndFactorizeArange.time_factorize(7)
-      4.10±0.7ms      2.90±0.02ms     0.71  rolling.VariableWindowMethods.time_rolling('DataFrame', '1d', 'int', 'sum')
-     6.75±0.02ms      4.75±0.01ms     0.70  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'float', 'kurt')
-        5.37±1μs       3.78±0.1μs     0.70  index_cached_properties.IndexCache.time_values('TimedeltaIndex')
-     6.76±0.03ms      4.75±0.01ms     0.70  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'kurt')
-        163±10μs          114±1μs     0.70  indexing.NonNumericSeriesIndexing.time_getitem_label_slice('period', 'nonunique_monotonic_inc')
-        104±10ms         72.7±4ms     0.70  frame_methods.Equals.time_frame_object_equal
-     6.21±0.03ms      4.34±0.01ms     0.70  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'min')
-        33.4±1ms         23.3±1ms     0.70  hash_functions.UniqueAndFactorizeArange.time_factorize(12)
-      6.99±0.1ms      4.86±0.03ms     0.70  gil.ParallelRolling.time_rolling('var')
-     6.12±0.03ms      4.25±0.04ms     0.70  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'min')
-      5.65±0.4ms      3.93±0.02ms     0.69  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'float', 'min')
-     6.20±0.03ms      4.28±0.01ms     0.69  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'min')
-      4.32±0.7ms      2.95±0.03ms     0.68  rolling.VariableWindowMethods.time_rolling('DataFrame', '50s', 'int', 'sum')
-     6.35±0.03ms      4.33±0.01ms     0.68  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'max')
-      45.5±0.2ms      30.9±0.03ms     0.68  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 8000, -2)
-     5.79±0.03ms      3.92±0.03ms     0.68  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'float', 'max')
-        2.75±0ms      1.86±0.02ms     0.68  series_methods.NanOps.time_func('argmax', 1000000, 'float64')
-      7.78±0.2ms       5.25±0.7ms     0.67  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'std')
-        8.70±1μs       5.87±0.2μs     0.67  index_cached_properties.IndexCache.time_shape('TimedeltaIndex')
-     1.78±0.04ms      1.20±0.06ms     0.67  replace.FillNa.time_replace(True)
-     6.31±0.02ms      4.24±0.02ms     0.67  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'max')
-         519±3ms          348±2ms     0.67  series_methods.IsInLongSeriesLookUpDominates.time_isin('object', 5, 'random_misses')
-     6.56±0.01ms      4.39±0.02ms     0.67  rolling.Methods.time_rolling('Series', 1000, 'int', 'skew')
-     4.67±0.03ms      3.11±0.01ms     0.67  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'mean')
-     5.72±0.03ms      3.79±0.01ms     0.66  rolling.Methods.time_rolling('Series', 1000, 'float', 'std')
-        40.9±1μs       27.1±0.2μs     0.66  series_methods.NanOps.time_func('argmax', 1000, 'int32')
-      73.1±0.6μs       48.3±0.5μs     0.66  series_methods.NanOps.time_func('argmax', 1000, 'float64')
-      41.0±0.8μs       27.1±0.2μs     0.66  series_methods.NanOps.time_func('argmax', 1000, 'int8')
-      17.4±0.2ms         11.5±2ms     0.66  stat_ops.FrameOps.time_op('skew', 'float', 0)
-     6.50±0.03ms      4.28±0.01ms     0.66  rolling.Methods.time_rolling('Series', 10, 'float', 'skew')
-     1.09±0.05ms         717±60μs     0.66  rolling.Engine.time_rolling_apply('DataFrame', 'int', <function sum at 0x7f2790e97430>, 'numba')
-     2.67±0.05μs      1.75±0.03μs     0.66  period.Indexing.time_shallow_copy
-     2.75±0.01ms      1.80±0.01ms     0.65  series_methods.NanOps.time_func('prod', 1000000, 'int8')
-        1.35±0ms          879±3μs     0.65  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 14)
-         234±2μs          153±1μs     0.65  stat_ops.SeriesOps.time_op('sum', 'float')
-      41.4±0.4μs       27.0±0.1μs     0.65  series_methods.NanOps.time_func('argmax', 1000, 'int64')
-     6.50±0.02ms      4.19±0.02ms     0.64  rolling.Methods.time_rolling('Series', 1000, 'float', 'min')
-        13.3±1ms      8.57±0.05ms     0.64  series_methods.NanOps.time_func('std', 1000000, 'float64')
-     6.51±0.01ms      4.18±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 0, 'midpoint')
-     6.55±0.01ms      4.20±0.03ms     0.64  rolling.Methods.time_rolling('Series', 1000, 'float', 'max')
-     6.51±0.02ms      4.17±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 0, 'linear')
-     6.50±0.03ms      4.16±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 0, 'nearest')
-     6.50±0.03ms      4.16±0.02ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 0, 'higher')
-      15.7±0.9ms       10.1±0.9ms     0.64  stat_ops.FrameOps.time_op('sem', 'int', 0)
-         359±1μs          229±1μs     0.64  index_cached_properties.IndexCache.time_is_monotonic('CategoricalIndex')
-       359±0.9μs          229±1μs     0.64  index_cached_properties.IndexCache.time_is_monotonic_decreasing('CategoricalIndex')
-         482±2μs          308±2μs     0.64  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 12)
-     6.53±0.03ms      4.17±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 0, 'lower')
-     6.55±0.01ms      4.18±0.02ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 1, 'linear')
-     6.57±0.02ms      4.19±0.02ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 1, 'nearest')
-         359±1μs          229±1μs     0.64  index_cached_properties.IndexCache.time_is_monotonic_increasing('CategoricalIndex')
-     6.54±0.01ms      4.17±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 1, 'higher')
-     4.38±0.02ms      2.79±0.01ms     0.64  rolling.Methods.time_rolling('DataFrame', 10, 'int', 'sum')
-     6.54±0.02ms      4.16±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 1, 'midpoint')
-      6.14±0.3ms      3.90±0.03ms     0.64  gil.ParallelRolling.time_rolling('mean')
-     6.57±0.01ms      4.18±0.01ms     0.64  rolling.Quantile.time_quantile('Series', 1000, 'float', 1, 'lower')
-      7.69±0.5ms      4.87±0.02ms     0.63  stat_ops.FrameOps.time_op('mad', 'int', 0)
-     7.02±0.03ms      4.43±0.03ms     0.63  rolling.Methods.time_rolling('Series', 10, 'int', 'skew')
-     4.49±0.01ms      2.82±0.01ms     0.63  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0, 'nearest')
-      6.94±0.7ms      4.36±0.01ms     0.63  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'min')
-     4.49±0.01ms      2.81±0.01ms     0.63  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0, 'lower')
-        97.8±8ms         61.3±1ms     0.63  frame_methods.Equals.time_frame_object_unequal
-     4.50±0.01ms      2.82±0.01ms     0.63  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0, 'higher')
-     4.50±0.01ms      2.82±0.02ms     0.63  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0, 'midpoint')
-     4.57±0.01ms      2.86±0.03ms     0.63  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 1, 'linear')
-     6.24±0.04ms      3.90±0.02ms     0.63  rolling.Methods.time_rolling('Series', 1000, 'int', 'std')
-     6.82±0.02ms      4.26±0.02ms     0.63  rolling.Methods.time_rolling('Series', 10, 'float', 'min')
-     2.04±0.02ms         1.27±0ms     0.62  series_methods.NanOps.time_func('sum', 1000000, 'int8')
-      7.19±0.8ms      4.49±0.03ms     0.62  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'min')
-     6.81±0.01ms      4.25±0.01ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 0, 'linear')
-      7.02±0.8ms      4.38±0.01ms     0.62  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'max')
-      7.04±0.8ms      4.38±0.03ms     0.62  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'min')
-     4.49±0.01ms      2.80±0.02ms     0.62  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 0, 'linear')
-     6.86±0.01ms      4.26±0.01ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 1, 'linear')
-     6.85±0.03ms      4.26±0.01ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 1, 'nearest')
-     6.88±0.02ms      4.27±0.02ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 1, 'lower')
-     6.83±0.02ms      4.24±0.02ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 0, 'midpoint')
-     6.84±0.02ms      4.24±0.02ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 0, 'nearest')
-     5.81±0.02ms      3.60±0.05ms     0.62  rolling.ExpandingMethods.time_expanding('Series', 'float', 'kurt')
-     4.57±0.01ms      2.83±0.02ms     0.62  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 1, 'midpoint')
-     4.59±0.02ms      2.84±0.01ms     0.62  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 1, 'lower')
-     6.90±0.03ms      4.27±0.02ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 1, 'midpoint')
-     6.85±0.02ms      4.24±0.01ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 0, 'lower')
-     6.89±0.02ms      4.27±0.01ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 1, 'higher')
-     6.85±0.03ms      4.24±0.02ms     0.62  rolling.Quantile.time_quantile('Series', 10, 'float', 0, 'higher')
-     6.64±0.09ms      4.10±0.03ms     0.62  period.DataFramePeriodColumn.time_set_index
-     6.91±0.08ms      4.26±0.01ms     0.62  rolling.Methods.time_rolling('Series', 10, 'float', 'max')
-     4.59±0.03ms      2.83±0.03ms     0.62  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 1, 'higher')
-     4.58±0.02ms      2.82±0.01ms     0.62  rolling.Quantile.time_quantile('DataFrame', 10, 'int', 1, 'nearest')
-      7.00±0.4ms         4.29±0ms     0.61  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'max')
-      7.38±0.8ms      4.51±0.01ms     0.61  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'max')
-      6.45±0.8ms      3.94±0.01ms     0.61  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'std')
-     6.96±0.02ms      4.25±0.02ms     0.61  rolling.Methods.time_rolling('Series', 1000, 'int', 'min')
-      18.0±0.4ms         10.9±1ms     0.61  stat_ops.FrameOps.time_op('kurt', 'float', 0)
-      6.46±0.8ms      3.92±0.01ms     0.61  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'std')
-      6.25±0.8ms      3.78±0.04ms     0.61  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'std')
-      7.21±0.8ms      4.36±0.01ms     0.60  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'max')
-        14.0±1ms      8.42±0.07ms     0.60  series_methods.NanOps.time_func('median', 1000000, 'float64')
-     1.30±0.01ms          779±7μs     0.60  groupby.GroupByMethods.time_dtype_as_group('int', 'rank', 'transformation')
-     1.04±0.02ms         625±20μs     0.60  rolling.Engine.time_rolling_apply('DataFrame', 'int', <function Engine.<lambda> at 0x7f277d8de310>, 'numba')
-     1.29±0.01ms          774±4μs     0.60  groupby.GroupByMethods.time_dtype_as_field('int', 'rank', 'direct')
-     7.07±0.03ms      4.24±0.01ms     0.60  rolling.Methods.time_rolling('Series', 1000, 'int', 'max')
-      7.92±0.2ms       4.73±0.7ms     0.60  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'std')
-        1.29±0ms          773±3μs     0.60  groupby.GroupByMethods.time_dtype_as_field('int', 'rank', 'transformation')
-       210±0.8ms        125±0.7ms     0.60  groupby.DateAttributes.time_len_groupby_object
-     1.30±0.01ms          775±2μs     0.59  groupby.GroupByMethods.time_dtype_as_group('int', 'rank', 'direct')
-     1.31±0.01ms          780±5μs     0.59  groupby.GroupByMethods.time_dtype_as_group('float', 'rank', 'direct')
-     7.33±0.03ms      4.35±0.01ms     0.59  rolling.Methods.time_rolling('Series', 10, 'int', 'min')
-     1.27±0.01ms          755±4μs     0.59  groupby.GroupByMethods.time_dtype_as_field('float', 'rank', 'transformation')
-     7.40±0.03ms      4.38±0.01ms     0.59  rolling.Methods.time_rolling('Series', 10, 'int', 'max')
-     2.32±0.01ms      1.37±0.01ms     0.59  series_methods.NanOps.time_func('sum', 1000000, 'float64')
-     6.35±0.01ms      3.76±0.06ms     0.59  rolling.ExpandingMethods.time_expanding('Series', 'int', 'kurt')
-      7.98±0.2ms       4.72±0.7ms     0.59  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'std')
-     1.32±0.01ms          778±4μs     0.59  groupby.GroupByMethods.time_dtype_as_group('float', 'rank', 'transformation')
-     1.06±0.03ms         624±60μs     0.59  rolling.Engine.time_rolling_apply('DataFrame', 'float', <function sum at 0x7f2790e97430>, 'numba')
-        29.1±1ms         17.1±2ms     0.59  hash_functions.UniqueAndFactorizeArange.time_unique(7)
-     1.03±0.05ms         605±10μs     0.59  rolling.Engine.time_rolling_apply('DataFrame', 'float', <function Engine.<lambda> at 0x7f277d8de310>, 'numba')
-     1.28±0.01ms          752±2μs     0.59  groupby.GroupByMethods.time_dtype_as_field('float', 'rank', 'direct')
-        29.4±1ms         17.2±2ms     0.58  hash_functions.UniqueAndFactorizeArange.time_unique(12)
-     4.45±0.01ms      2.59±0.02ms     0.58  rolling.EWMMethods.time_ewm('Series', 1000, 'int', 'std')
-     4.44±0.02ms      2.58±0.03ms     0.58  rolling.EWMMethods.time_ewm('Series', 10, 'int', 'std')
-     6.53±0.02ms      3.79±0.01ms     0.58  rolling.Methods.time_rolling('Series', 10, 'float', 'std')
-     5.45±0.01ms       3.15±0.1ms     0.58  rolling.ExpandingMethods.time_expanding('DataFrame', 'int', 'count')
-     1.33±0.01ms          765±3μs     0.58  groupby.GroupByMethods.time_dtype_as_group('datetime', 'rank', 'direct')
-     1.33±0.01ms          762±2μs     0.57  groupby.GroupByMethods.time_dtype_as_group('datetime', 'rank', 'transformation')
-         384±2μs          218±2μs     0.57  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 11)
-     4.36±0.01ms      2.45±0.03ms     0.56  rolling.EWMMethods.time_ewm('Series', 10, 'float', 'std')
-     4.36±0.01ms      2.44±0.01ms     0.56  rolling.EWMMethods.time_ewm('Series', 1000, 'float', 'std')
-     6.32±0.03ms       3.52±0.1ms     0.56  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'kurt')
-     6.32±0.01ms      3.53±0.03ms     0.56  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'kurt')
-     7.07±0.05ms      3.94±0.01ms     0.56  rolling.Methods.time_rolling('Series', 10, 'int', 'std')
-     6.31±0.01ms      3.50±0.02ms     0.55  rolling.ExpandingMethods.time_expanding('Series', 'float', 'min')
-      4.99±0.7ms      2.77±0.01ms     0.55  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'mean')
-      4.95±0.7ms      2.74±0.03ms     0.55  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'mean')
-     6.39±0.01ms      3.50±0.03ms     0.55  rolling.ExpandingMethods.time_expanding('Series', 'float', 'max')
-      4.94±0.7ms      2.70±0.03ms     0.55  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'mean')
-      5.24±0.8ms      2.87±0.02ms     0.55  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'mean')
-     3.94±0.05ms      2.16±0.01ms     0.55  series_methods.ValueCounts.time_value_counts('float')
-      56.0±0.5ms      30.6±0.05ms     0.55  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 2000, 2)
-      5.29±0.8ms      2.88±0.01ms     0.55  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'mean')
-        1.12±0ms          611±5μs     0.54  groupby.GroupByMethods.time_dtype_as_field('datetime', 'rank', 'transformation')
-        1.13±0ms          612±3μs     0.54  groupby.GroupByMethods.time_dtype_as_field('datetime', 'rank', 'direct')
-      5.21±0.8ms      2.81±0.01ms     0.54  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'mean')
-      43.0±0.1ms       23.0±0.1ms     0.54  join_merge.Concat.time_concat_small_frames(0)
-        88.6±1ms      47.4±0.09ms     0.54  frame_methods.Equals.time_frame_nonunique_equal
-     6.80±0.01ms      3.63±0.02ms     0.53  rolling.ExpandingMethods.time_expanding('Series', 'int', 'min')
-        89.8±1ms       48.0±0.2ms     0.53  frame_methods.Equals.time_frame_nonunique_unequal
-     5.89±0.03ms      3.15±0.04ms     0.53  rolling.ExpandingMethods.time_expanding('Series', 'float', 'skew')
-         956±4μs          507±2μs     0.53  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 13)
-     6.89±0.02ms      3.63±0.02ms     0.53  rolling.ExpandingMethods.time_expanding('Series', 'int', 'max')
-     1.14±0.01ms          595±6μs     0.52  groupby.GroupByMethods.time_dtype_as_group('float', 'cumprod', 'transformation')
-      67.4±0.3ms       35.3±0.3ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 5, 'random_hits')
-      68.9±0.4ms       36.1±0.3ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 5, 'random_hits')
-      68.3±0.3ms       35.7±0.4ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 5, 'monotone_hits')
-     1.14±0.01ms          594±5μs     0.52  groupby.GroupByMethods.time_dtype_as_group('float', 'cumprod', 'direct')
-        1.11±0ms          580±2μs     0.52  groupby.GroupByMethods.time_dtype_as_group('int', 'cumprod', 'transformation')
-      68.3±0.4ms       35.5±0.4ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 5, 'random_misses')
-      68.0±0.3ms       35.3±0.3ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 5, 'monotone_misses')
-      66.7±0.4ms       34.6±0.4ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 5, 'random_misses')
-      4.66±0.7ms      2.42±0.01ms     0.52  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'float', 'sum')
-      67.2±0.3ms       34.8±0.3ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 5, 'monotone_hits')
-      66.9±0.2ms       34.6±0.3ms     0.52  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 5, 'monotone_misses')
-     6.41±0.01ms      3.31±0.06ms     0.52  rolling.ExpandingMethods.time_expanding('Series', 'int', 'skew')
-     1.13±0.03ms          582±8μs     0.52  groupby.GroupByMethods.time_dtype_as_group('int', 'cumprod', 'direct')
-      4.95±0.8ms      2.55±0.02ms     0.51  rolling.VariableWindowMethods.time_rolling('Series', '1h', 'int', 'sum')
-      4.65±0.7ms      2.39±0.01ms     0.51  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'float', 'sum')
-      4.71±0.7ms      2.42±0.01ms     0.51  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'float', 'sum')
-     1.08±0.01ms          556±2μs     0.51  groupby.GroupByMethods.time_dtype_as_field('int', 'cumprod', 'transformation')
-     1.09±0.01ms          557±2μs     0.51  groupby.GroupByMethods.time_dtype_as_field('int', 'cumprod', 'direct')
-     7.78±0.02ms      3.97±0.01ms     0.51  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'min')
-     7.86±0.02ms      4.01±0.02ms     0.51  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'max')
-      5.89±0.2μs      2.93±0.03μs     0.50  categoricals.Indexing.time_get_loc
-         338±2μs        167±0.9μs     0.50  hash_functions.IsinAlmostFullWithRandomInt.time_isin_outside(<class 'numpy.float64'>, 10)
-     5.30±0.03ms      2.62±0.01ms     0.49  rolling.Methods.time_rolling('Series', 1000, 'float', 'mean')
-     5.53±0.01ms      2.73±0.02ms     0.49  rolling.ExpandingMethods.time_expanding('DataFrame', 'float', 'count')
-        1.59±0ms          782±2μs     0.49  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 8000)
-      5.29±0.8ms      2.57±0.01ms     0.49  rolling.VariableWindowMethods.time_rolling('Series', '50s', 'int', 'sum')
-         621±1ms        302±0.3ms     0.49  series_methods.IsInLongSeriesLookUpDominates.time_isin('float64', 1000, 'monotone_misses')
-      4.64±0.3ms      2.23±0.01ms     0.48  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'min')
-      4.70±0.3ms      2.25±0.01ms     0.48  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'max')
-     5.82±0.02ms      2.74±0.04ms     0.47  rolling.Methods.time_rolling('Series', 1000, 'int', 'mean')
-     5.01±0.01ms      2.31±0.02ms     0.46  rolling.Methods.time_rolling('Series', 1000, 'float', 'sum')
-     5.73±0.01ms      2.63±0.01ms     0.46  rolling.Methods.time_rolling('Series', 10, 'float', 'mean')
-      11.4±0.8ms      5.20±0.02ms     0.46  stat_ops.FrameOps.time_op('mad', 'int', 1)
-         1.58±0s          713±5ms     0.45  series_methods.IsInLongSeriesValuesDominate.time_isin('float64', 'random')
-      5.65±0.8ms      2.53±0.01ms     0.45  rolling.VariableWindowMethods.time_rolling('Series', '1d', 'int', 'sum')
-     5.72±0.02ms      2.54±0.03ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 1, 'midpoint')
-     5.71±0.02ms      2.53±0.02ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 1, 'linear')
-     6.24±0.01ms      2.77±0.02ms     0.44  rolling.Methods.time_rolling('Series', 10, 'int', 'mean')
-     5.69±0.02ms      2.52±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 0, 'lower')
-      41.1±0.3ms      18.1±0.05ms     0.44  index_object.Range.time_iter_dec
-     5.74±0.02ms      2.52±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 1, 'lower')
-     6.08±0.03ms       2.67±0.2ms     0.44  rolling.Quantile.time_quantile('Series', 10, 'int', 0, 'nearest')
-     5.69±0.04ms      2.50±0.04ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 0, 'midpoint')
-     5.74±0.03ms      2.51±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 1, 'nearest')
-     5.74±0.02ms      2.51±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 1, 'higher')
-     5.70±0.02ms      2.49±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 0, 'linear')
-     5.72±0.01ms      2.50±0.02ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 0, 'nearest')
-      41.5±0.9ms       18.1±0.1ms     0.44  index_object.Range.time_iter_inc
-     5.58±0.02ms      2.44±0.02ms     0.44  rolling.Methods.time_rolling('Series', 1000, 'int', 'sum')
-     6.08±0.02ms       2.65±0.2ms     0.44  rolling.Quantile.time_quantile('Series', 10, 'int', 0, 'higher')
-     5.71±0.04ms      2.48±0.01ms     0.44  rolling.Quantile.time_quantile('Series', 1000, 'int', 0, 'higher')
-       639±0.5ms        278±0.7ms     0.44  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 1000, 'monotone_misses')
-         915±5μs          394±3μs     0.43  groupby.GroupByMethods.time_dtype_as_field('int', 'cummax', 'transformation')
-         920±9μs          393±4μs     0.43  groupby.GroupByMethods.time_dtype_as_group('int', 'cummax', 'transformation')
-         913±3μs          389±4μs     0.43  groupby.GroupByMethods.time_dtype_as_group('int', 'cummin', 'transformation')
-         939±4μs          400±4μs     0.43  groupby.GroupByMethods.time_dtype_as_group('float', 'cumsum', 'direct')
-         914±4μs          389±2μs     0.43  groupby.GroupByMethods.time_dtype_as_group('int', 'cummax', 'direct')
-         912±8μs          388±3μs     0.43  groupby.GroupByMethods.time_dtype_as_field('int', 'cummin', 'direct')
-         925±9μs          393±4μs     0.42  groupby.GroupByMethods.time_dtype_as_field('int', 'cummax', 'direct')
-     5.45±0.02ms         2.31±0ms     0.42  rolling.Methods.time_rolling('Series', 10, 'float', 'sum')
-     4.40±0.04ms      1.87±0.02ms     0.42  stat_ops.FrameOps.time_op('mean', 'float', 0)
-         913±3μs          387±2μs     0.42  groupby.GroupByMethods.time_dtype_as_field('int', 'cummin', 'transformation')
-         909±4μs          383±4μs     0.42  groupby.GroupByMethods.time_dtype_as_field('int', 'cumsum', 'direct')
-         915±4μs          385±3μs     0.42  groupby.GroupByMethods.time_dtype_as_group('int', 'cummin', 'direct')
-         912±3μs          383±3μs     0.42  groupby.GroupByMethods.time_dtype_as_group('int', 'cumsum', 'direct')
-     4.38±0.04ms      1.83±0.02ms     0.42  stat_ops.FrameOps.time_op('sum', 'float', 0)
-         943±7μs          395±4μs     0.42  groupby.GroupByMethods.time_dtype_as_group('float', 'cumsum', 'transformation')
-         931±4μs          390±1μs     0.42  groupby.GroupByMethods.time_dtype_as_group('float', 'cummin', 'transformation')
-         759±5μs          318±4μs     0.42  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 12)
-         907±9μs          379±3μs     0.42  groupby.GroupByMethods.time_dtype_as_field('int', 'cumsum', 'transformation')
-         936±8μs          391±3μs     0.42  groupby.GroupByMethods.time_dtype_as_group('float', 'cummax', 'direct')
-      16.8±0.9ms       7.03±0.3ms     0.42  stat_ops.FrameOps.time_op('mad', 'float', 1)
-        932±20μs          388±9μs     0.42  groupby.GroupByMethods.time_dtype_as_group('int', 'cumsum', 'transformation')
-        934±20μs          389±3μs     0.42  groupby.GroupByMethods.time_dtype_as_group('float', 'cummax', 'transformation')
-         935±2μs          388±2μs     0.42  groupby.GroupByMethods.time_dtype_as_group('float', 'cummin', 'direct')
-         1.50±0s          621±2ms     0.42  series_methods.IsInLongSeriesValuesDominate.time_isin('int64', 'random')
-         1.61±0s          667±5ms     0.41  series_methods.IsInLongSeriesValuesDominate.time_isin('float32', 'random')
-      3.24±0.2ms      1.34±0.01ms     0.41  rolling.EWMMethods.time_ewm_times('Series', 1000, 'int', 'std')
-     6.17±0.03ms      2.53±0.02ms     0.41  rolling.Quantile.time_quantile('Series', 10, 'int', 1, 'nearest')
-     6.07±0.01ms      2.48±0.02ms     0.41  rolling.Quantile.time_quantile('Series', 10, 'int', 0, 'lower')
-      3.28±0.2ms      1.34±0.01ms     0.41  rolling.EWMMethods.time_ewm_times('Series', 10, 'int', 'std')
-      3.28±0.2ms         1.34±0ms     0.41  rolling.EWMMethods.time_ewm('Series', 1000, 'int', 'mean')
-     5.99±0.02ms      2.45±0.01ms     0.41  rolling.Methods.time_rolling('Series', 10, 'int', 'sum')
-      14.0±0.9ms         5.70±1ms     0.41  stat_ops.FrameOps.time_op('mad', 'float', 0)
-     6.06±0.02ms      2.47±0.01ms     0.41  rolling.Quantile.time_quantile('Series', 10, 'int', 0, 'linear')
-      3.28±0.2ms      1.34±0.01ms     0.41  rolling.EWMMethods.time_ewm_times('Series', 10, 'int', 'mean')
-     6.06±0.02ms      2.47±0.01ms     0.41  rolling.Quantile.time_quantile('Series', 10, 'int', 0, 'midpoint')
-     5.83±0.03ms      2.37±0.04ms     0.41  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'mean')
-      3.30±0.2ms      1.34±0.01ms     0.41  rolling.EWMMethods.time_ewm_times('Series', 1000, 'int', 'mean')
-     6.16±0.03ms      2.49±0.01ms     0.40  rolling.Quantile.time_quantile('Series', 10, 'int', 1, 'lower')
-     6.16±0.01ms      2.49±0.01ms     0.40  rolling.Quantile.time_quantile('Series', 10, 'int', 1, 'higher')
-     6.17±0.03ms      2.49±0.02ms     0.40  rolling.Quantile.time_quantile('Series', 10, 'int', 1, 'linear')
-     5.82±0.01ms      2.35±0.01ms     0.40  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'float', 'mean')
-         949±8μs          383±4μs     0.40  groupby.GroupByMethods.time_dtype_as_group('datetime', 'cummin', 'transformation')
-      6.17±0.1ms      2.48±0.01ms     0.40  rolling.Quantile.time_quantile('Series', 10, 'int', 1, 'midpoint')
-         950±5μs          382±2μs     0.40  groupby.GroupByMethods.time_dtype_as_group('datetime', 'cummin', 'direct')
-     52.1±0.09ms      20.8±0.06ms     0.40  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 8000, 0)
-     6.34±0.01ms      2.49±0.01ms     0.39  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'mean')
-         859±3μs          336±3μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cumprod', 'transformation')
-     3.14±0.01ms      1.23±0.02ms     0.39  rolling.EWMMethods.time_ewm('Series', 1000, 'float', 'mean')
-     3.13±0.01ms      1.22±0.01ms     0.39  rolling.EWMMethods.time_ewm_times('Series', 10, 'float', 'std')
-     6.34±0.01ms         2.47±0ms     0.39  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'mean')
-         866±5μs          337±2μs     0.39  groupby.GroupByMethods.time_dtype_as_field('datetime', 'cummin', 'transformation')
-     3.14±0.01ms         1.22±0ms     0.39  rolling.EWMMethods.time_ewm_times('Series', 1000, 'float', 'mean')
-         859±2μs          335±2μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cumsum', 'direct')
-     3.14±0.02ms      1.22±0.01ms     0.39  rolling.EWMMethods.time_ewm_times('Series', 10, 'float', 'mean')
-         866±3μs          336±3μs     0.39  groupby.GroupByMethods.time_dtype_as_field('datetime', 'cummin', 'direct')
-     3.14±0.02ms      1.22±0.01ms     0.39  rolling.EWMMethods.time_ewm_times('Series', 1000, 'float', 'std')
-         867±5μs          336±4μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cummin', 'transformation')
-         870±4μs          337±1μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cummax', 'transformation')
-         871±7μs        337±0.9μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cummax', 'direct')
-         858±7μs          332±2μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cumsum', 'transformation')
-         865±3μs          334±3μs     0.39  groupby.GroupByMethods.time_dtype_as_field('float', 'cumprod', 'direct')
-     3.15±0.02ms      1.22±0.01ms     0.39  rolling.EWMMethods.time_ewm('Series', 10, 'float', 'mean')
-     5.23±0.05ms         2.01±0ms     0.38  stat_ops.FrameOps.time_op('mean', 'float', 1)
-         768±2μs          294±7μs     0.38  indexing_engines.NumericEngineIndexing.time_get_loc((<class 'pandas._libs.index.Int16Engine'>, <class 'numpy.int16'>), 'monotonic_incr')
-         873±2μs          334±2μs     0.38  groupby.GroupByMethods.time_dtype_as_field('float', 'cummin', 'direct')
-      3.51±0.2ms         1.33±0ms     0.38  rolling.EWMMethods.time_ewm('Series', 10, 'int', 'mean')
-        1.02±0ms          388±1μs     0.38  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 1300)
-     1.70±0.01ms          640±3μs     0.38  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 2000)
-     1.70±0.01ms          638±6μs     0.37  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 2000)
-     5.50±0.01ms      2.06±0.04ms     0.37  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'float', 'sum')
-     5.51±0.02ms      2.06±0.01ms     0.37  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'float', 'sum')
-     5.06±0.02ms      1.89±0.03ms     0.37  stat_ops.FrameOps.time_op('sum', 'float', 1)
-        27.2±2ms         10.0±3ms     0.37  stat_ops.FrameOps.time_op('sem', 'float', 0)
-     5.01±0.01ms      1.84±0.03ms     0.37  rolling.ExpandingMethods.time_expanding('Series', 'float', 'mean')
-      13.2±0.8ms         4.83±1ms     0.37  stat_ops.FrameOps.time_op('var', 'float', 0)
-     6.22±0.03ms      2.27±0.01ms     0.37  rolling.ExpandingMethods.time_expanding('Series', 'float', 'std')
-        1.12±0ms          409±4μs     0.36  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 1300)
-     6.03±0.02ms      2.19±0.01ms     0.36  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'sum')
-     6.05±0.05ms      2.18±0.01ms     0.36  rolling.ForwardWindowMethods.time_rolling('Series', 1000, 'int', 'sum')
-     4.72±0.02ms      1.69±0.02ms     0.36  rolling.ExpandingMethods.time_expanding('Series', 'float', 'sum')
-     6.75±0.03ms      2.41±0.05ms     0.36  rolling.ExpandingMethods.time_expanding('Series', 'int', 'std')
-     5.59±0.01ms      1.96±0.02ms     0.35  rolling.ExpandingMethods.time_expanding('Series', 'int', 'mean')
-        26.2±2ms       9.02±0.3ms     0.34  frame_methods.Equals.time_frame_float_unequal
-     5.26±0.03ms      1.81±0.04ms     0.34  rolling.ExpandingMethods.time_expanding('Series', 'int', 'sum')
-     5.93±0.02ms      2.03±0.02ms     0.34  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 7000)
-     6.51±0.02ms      2.21±0.01ms     0.34  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'min')
-         772±6μs          260±2μs     0.34  series_methods.IsInDatetime64.time_isin_empty
-      89.2±0.2ms      30.0±0.04ms     0.34  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 1000, 2)
-     6.60±0.01ms      2.22±0.01ms     0.34  rolling.ForwardWindowMethods.time_rolling('Series', 10, 'int', 'max')
-     9.09±0.02ms       3.02±0.1ms     0.33  rolling.Methods.time_rolling('Series', 1000, 'float', 'count')
-         646±7μs          214±2μs     0.33  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 11)
-     8.99±0.01ms       2.97±0.1ms     0.33  rolling.Methods.time_rolling('Series', 1000, 'int', 'count')
-        812±40μs          268±3μs     0.33  indexing_engines.NumericEngineIndexing.time_get_loc((<class 'pandas._libs.index.Int8Engine'>, <class 'numpy.int8'>), 'monotonic_incr')
-     7.00±0.03ms      2.30±0.02ms     0.33  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 8000)
-         1.51±0s          488±1ms     0.32  series_methods.IsInLongSeriesValuesDominate.time_isin('int32', 'random')
-     9.52±0.04ms       3.06±0.1ms     0.32  rolling.Methods.time_rolling('Series', 10, 'float', 'count')
-     9.45±0.07ms       2.99±0.1ms     0.32  rolling.Methods.time_rolling('Series', 10, 'int', 'count')
-      6.10±0.1ms         1.92±0ms     0.32  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 7000)
-      98.7±0.3ms       30.5±0.1ms     0.31  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 2000, -2)
-         586±1ms        179±0.2ms     0.31  series_methods.IsInLongSeriesLookUpDominates.time_isin('int64', 1000, 'monotone_misses')
-         529±6μs          162±1μs     0.31  hash_functions.IsinAlmostFullWithRandomInt.time_isin(<class 'numpy.float64'>, 10)
-     7.26±0.08ms      2.19±0.01ms     0.30  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 8000)
-      14.3±0.2ms         4.25±2ms     0.30  stat_ops.FrameOps.time_op('std', 'float', 0)
-     1.02±0.01ms          297±1μs     0.29  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 2000)
-      8.86±0.2ms      2.56±0.02ms     0.29  arithmetic.Timeseries.time_timestamp_ops_diff(None)
-     16.3±0.05ms      4.67±0.04ms     0.29  indexing.InsertColumns.time_assign_list_like_with_setitem
-         939±8μs          263±3μs     0.28  series_methods.IsInDatetime64.time_isin_mismatched_dtype
-        78.2±3ms       21.5±0.4ms     0.28  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 70000)
-         601±1ms        160±0.1ms     0.27  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 1000, 'monotone_misses')
-     1.71±0.01ms          437±2μs     0.26  series_methods.Dir.time_dir_strings
-       109±0.9ms       27.0±0.3ms     0.25  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 80000)
-       495±0.9ms        120±0.8ms     0.24  series_methods.IsInLongSeriesLookUpDominates.time_isin('float64', 1000, 'monotone_hits')
-     8.81±0.06ms       2.09±0.1ms     0.24  rolling.ExpandingMethods.time_expanding('Series', 'float', 'count')
-        864±10μs        205±0.8μs     0.24  hash_functions.IsinWithArangeSorted.time_isin(<class 'numpy.float64'>, 1000)
-      8.98±0.2ms       2.04±0.1ms     0.23  rolling.ExpandingMethods.time_expanding('Series', 'int', 'count')
-      3.55±0.02s          800±3ms     0.23  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 900000)
-      3.56±0.01s          733±4ms     0.21  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 900000)
-         513±1ms        104±0.2ms     0.20  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 1000, 'monotone_hits')
-        99.5±4ms       19.8±0.3ms     0.20  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 70000)
-       165±0.4ms      30.8±0.03ms     0.19  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 1000, -2)
-        140±10ms       25.6±0.2ms     0.18  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 80000)
-      3.14±0.02s          539±4ms     0.17  hash_functions.IsinWithRandomFloat.time_isin(<class 'object'>, 750000)
-       118±0.1ms      19.9±0.04ms     0.17  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 2000, 0)
-       388±0.7ms       64.5±0.8ms     0.17  series_methods.IsInLongSeriesValuesDominate.time_isin('int32', 'monotone')
-         1.89±0s        309±0.3ms     0.16  series_methods.IsInLongSeriesLookUpDominates.time_isin('float64', 1000, 'random_misses')
-     4.28±0.01ms          678±6μs     0.16  index_cached_properties.IndexCache.time_is_unique('CategoricalIndex')
-       3.97±0.1s          600±2ms     0.15  hash_functions.IsinWithRandomFloat.time_isin_outside(<class 'object'>, 750000)
-         1.90±0s        284±0.2ms     0.15  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 1000, 'random_misses')
-     4.57±0.08ms          670±4μs     0.15  indexing.NonNumericSeriesIndexing.time_getitem_list_like('period', 'nonunique_monotonic_inc')
-         465±1ms       59.1±0.2ms     0.13  series_methods.IsInLongSeriesLookUpDominates.time_isin('int64', 1000, 'monotone_hits')
-         184±6ms         23.1±1ms     0.13  hash_functions.UniqueAndFactorizeArange.time_factorize(11)
-        194±10ms         23.2±1ms     0.12  hash_functions.UniqueAndFactorizeArange.time_factorize(8)
-       483±0.6ms       57.1±0.1ms     0.12  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 1000, 'monotone_hits')
-         1.89±0s        187±0.7ms     0.10  series_methods.IsInLongSeriesLookUpDominates.time_isin('float64', 1000, 'random_hits')
-         178±6ms         17.1±2ms     0.10  hash_functions.UniqueAndFactorizeArange.time_unique(11)
-       200±0.1ms       19.0±0.1ms     0.10  hash_functions.IsinWithArange.time_isin(<class 'numpy.float64'>, 1000, 0)
-         588±6ms       54.0±0.5ms     0.09  categoricals.Indexing.time_reindex_missing
-         187±9ms         17.1±2ms     0.09  hash_functions.UniqueAndFactorizeArange.time_unique(8)
-        27.8±3ms      2.50±0.06ms     0.09  frame_methods.Equals.time_frame_float_equal
-         1.90±0s        171±0.3ms     0.09  series_methods.IsInLongSeriesLookUpDominates.time_isin('float32', 1000, 'random_hits')
-      10.6±0.3ms         924±10μs     0.09  frame_methods.Shift.time_shift(1)
-       128±0.8ms      5.47±0.05ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'sum')
-       129±0.7ms      5.49±0.06ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'sum')
-       129±0.9ms      5.45±0.05ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'min')
-       128±0.6ms      5.44±0.04ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'mean')
-       129±0.9ms      5.45±0.02ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'min')
-         129±1ms      5.44±0.03ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'mean')
-         132±1ms      5.54±0.06ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'median')
-       132±0.8ms      5.48±0.03ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'median')
-       129±0.7ms      5.30±0.04ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'max')
-       128±0.8ms      5.27±0.02ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'max')
-       137±0.7ms      5.53±0.02ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'std')
-         137±1ms      5.48±0.05ms     0.04  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'std')
-       182±0.6ms      5.79±0.05ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'median')
-       180±0.9ms      5.71±0.09ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'sum')
-         180±1ms       5.70±0.1ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'sum')
-      1.87±0.02s       59.1±0.5ms     0.03  series_methods.IsInLongSeriesLookUpDominates.time_isin('int64', 1000, 'random_hits')
-       183±0.9ms      5.78±0.05ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'median')
-         180±1ms      5.68±0.04ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'mean')
-         181±1ms      5.68±0.04ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'min')
-         181±1ms      5.66±0.05ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'mean')
-         181±1ms      5.66±0.08ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'max')
-         181±1ms      5.64±0.02ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'max')
-         181±1ms      5.62±0.06ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'min')
-       189±0.7ms      5.83±0.07ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'std')
-       189±0.9ms      5.81±0.07ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'std')
-         1.88±0s       57.4±0.1ms     0.03  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 1000, 'random_hits')
-      1.86±0.06s       54.8±0.1ms     0.03  series_methods.IsInLongSeriesLookUpDominates.time_isin('int64', 1000, 'random_misses')
-         205±2ms      5.76±0.02ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('Series', 'int', 'count')
-         205±2ms      5.74±0.04ms     0.03  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'int', 'count')
-        869±40ms         23.0±1ms     0.03  hash_functions.UniqueAndFactorizeArange.time_factorize(10)
-      84.1±0.5μs      2.05±0.02μs     0.02  period.Indexing.time_unique
-         1.88±0s       42.8±0.1ms     0.02  series_methods.IsInLongSeriesLookUpDominates.time_isin('int32', 1000, 'random_misses')
-       114±0.4ms       2.56±0.5ms     0.02  timeseries.ToDatetimeFromIntsFloats.time_nanosec_float64
-         103±1μs      2.06±0.01μs     0.02  timedelta.TimedeltaIndexing.time_unique
-        862±50ms         17.1±2ms     0.02  hash_functions.UniqueAndFactorizeArange.time_unique(10)
-         348±2ms      6.14±0.05ms     0.02  rolling.ExpandingMethods.time_expanding_groupby('DataFrame', 'float', 'count')
-         347±2ms      6.09±0.02ms     0.02  rolling.ExpandingMethods.time_expanding_groupby('Series', 'float', 'count')
-       2.30±0.1s       36.8±0.7ms     0.02  hash_functions.Float64GroupIndex.time_groupby
-      1.64±0.09s         23.5±1ms     0.01  hash_functions.UniqueAndFactorizeArange.time_factorize(9)
-       171±0.6μs      2.32±0.01μs     0.01  timeseries.DatetimeIndex.time_unique('dst')
-      1.63±0.09s         17.2±2ms     0.01  hash_functions.UniqueAndFactorizeArange.time_unique(9)
-      4.31±0.04s       9.19±0.7ms     0.00  timeseries.ToDatetimeFromIntsFloats.time_sec_float64
-     3.34±0.01ms      3.16±0.02μs     0.00  timeseries.DatetimeIndex.time_unique('tz_naive')
-     3.40±0.03ms      3.15±0.01μs     0.00  timeseries.DatetimeIndex.time_unique('tz_local')
-     3.43±0.03ms      3.13±0.06μs     0.00  timeseries.DatetimeIndex.time_unique('tz_aware')

SOME BENCHMARKS HAVE CHANGED SIGNIFICANTLY.
PERFORMANCE DECREASED.

The text was updated successfully, but these errors were encountered:

jorisvandenbossche · 2020-12-20T16:49:25Z

@mroeschke possibly relevant to look into:

+         336±2ms          1.22±0s     3.65  groupby.TransformEngine.time_series_numba(True)
+         287±2ms          1.02±0s     3.55  groupby.AggEngine.time_series_numba(True)
+         289±2ms          1.02±0s     3.52  groupby.AggEngine.time_dataframe_numba(True)

(the benchmarks were only added or were changed after 1.1 release (#36240), so the online benchmarks at https://pandas.pydata.org/speed/pandas/ don't show a regression)

stuartarchibald · 2021-02-01T14:36:05Z

@jorisvandenbossche Was the Numba version the same across these two test runs?

mroeschke · 2023-10-11T18:53:56Z

Given this issue is with an unsupported version of pandas, going to close for now

jorisvandenbossche added the Performance Memory or execution speed performance label Dec 20, 2020

mroeschke closed this as completed Oct 11, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

PERF: performance regressions in 1.2.0rc #38591

PERF: performance regressions in 1.2.0rc #38591

jorisvandenbossche commented Dec 20, 2020 •

edited

Loading

jorisvandenbossche commented Dec 20, 2020

stuartarchibald commented Feb 1, 2021

mroeschke commented Oct 11, 2023

PERF: performance regressions in 1.2.0rc #38591

PERF: performance regressions in 1.2.0rc #38591

Comments

jorisvandenbossche commented Dec 20, 2020 • edited Loading

jorisvandenbossche commented Dec 20, 2020

stuartarchibald commented Feb 1, 2021

mroeschke commented Oct 11, 2023

jorisvandenbossche commented Dec 20, 2020 •

edited

Loading