This repository is Python benchmark for common usage in data science.
Environment: OS Ubuntu 18.04, CPU Intel i5-8400, Anaconda Python 3.6.5
Data: n row x 4 columns random double dataframe.
Remark:
- time or size (Y axis) in log scale.
- Maximum row for excel output is 1048575 rows, so file size and read speed of 10^7 are under estimated.
Installation please refer to my blog post: Chinese English
Result (run time in seconds):
| OpenBLAS | MKL | |
|---|---|---|
| Vector Multiply | 4.637 | 5.022 |
| Matrix Multiply | 4.629 | 5.224 |
| Eigen decomposition | 4.695 | 3.449 |
Remark: vector size 8192, matrix size for multiply 8192x8192, matrix size for decomposition 2048x2048