ASV: added benchamark tests for DataFrame.to_numpy() and .values

[hardikpnsp]

• closes ASV: asvs for DataFrame.values, to_numpy #35023
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

Adds benchmarks with single dtype and mixed frames of varying sizes.
xref #34999

[jreback]

can you run these so we can see (and post here) in dev mode is fine as there is nothing to compare

[hardikpnsp]

@jreback, here are the benchmark results produced by asv dev. hope this is what we are looking for.

[  6.25%] ··· frame_methods.ToNumpy.time_to_numpy_mixed_tall                                      18.5±0ms
[ 12.50%] ··· frame_methods.ToNumpy.time_to_numpy_mixed_wide                                      3.97±0ms
[ 18.75%] ··· frame_methods.ToNumpy.time_to_numpy_tall                                            40.9±0μs
[ 25.00%] ··· frame_methods.ToNumpy.time_to_numpy_wide                                            95.4±0μs
[ 31.25%] ··· frame_methods.ToNumpy.time_values_mixed_tall                                        21.6±0ms
[ 37.50%] ··· frame_methods.ToNumpy.time_values_mixed_wide                                        5.70±0ms
[ 43.75%] ··· frame_methods.ToNumpy.time_values_tall                                              56.4±0μs
[ 50.00%] ··· frame_methods.ToNumpy.time_values_wide                                              86.0±0μs

[hardikpnsp]

This would be my first opensource contribution. I was not sure in what section of whatsnew file should I do an entry for this. Or, if I need to do it at all for benchmark tests, would appreciate some notes in that direction. Thank you.

[jreback]

thanks @hardikpnsp no whatsnew necessary for this.