element-wise value comparison noticeably slower for large-scale dataframes

[leo4183]

Problem Description

element-wise value comparison noticeably slower for large-scale dataframes, compared to corresponding numpy array value comparison (although in older pandas version, the operation speed was acceptable).

import numpy as np
import pandas as pd
x = pd.DataFrame(np.random.rand(3000*4000).reshape((3000,-1)))

# pandas 0.25.1
%timeit x>0.0 
1.72 s ± 3.23 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

# pandas 0.17.1
%timeit x>0.0
115 ms ± 114 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

%timeit x.values>0.0
13.1 ms ± 148 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

INSTALLED VERSIONS

commit : None
python : 3.7.4.final.0
python-bits : 64
OS : Linux
OS-release : 4.18.0-80.7.2.el8_0.x86_64
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 0.25.1
numpy : 1.17.2

[TomAugspurger]

Duplicate of #24990. @jbrockmendel is working on this but it's a large project. Should be done for 1.0