pybench

This project provides a single file bench.py that is inspired to the bench.m function available in MATLAB®.

Seven tasks are tested:

• LU factorization
• FFT Fast Fourier Transform
• ODE Ordinary diff. eqn.
• SPARSE Solve sparse system
• 2-D plot involving the use if "imshow"
• 2-D plot involving the use if "drawnow"
• 3-D plot involving the use if "mplot3d"

(disclaimer: the gif is accelerated by a factor of 4 with respect to the real execution time)

1. Installation

Clone the repository:

git clone https://github.com/giaccone/pybench.git

and run the script bench.py with a python interpreter having this modules installed:

• py-cpuinfo
• numpy
• scipy
• matplotlib
• drawnow
• rich

2. Usage

Simply run the script

python bench.py <options>

available options:

• n: an positive integer representing the number of times each task is executed (default is 1)
• export: boolean 0 or 1. If set to 1 the script exports the best results of each task on a file called results.txt (default is 0)
• import: boolean 0 or 1. If set to 1 the script import the results of other machines collected in the file other-results.txt (default is 0)

3. Examples

3.1 Single run

python bench.py

3.2 Multiple run

With optional parameter n it is possible to set the number of times each task is executed. The following example makes use of n=3.

python bench.py n=3

3.3 Add other results for comparison

With optional parameter import it is possible include in the output results of other machine previously tested and saved on the file other-results.txt. The repository currently include a file with the results of three different machines.

python bench.py n=3 import=1

3.4 Export results

With optional parameter export it is possible export the (best) results of the machine under test in a file called results.txt. The content of this file can be copied in the file other-results.txt for future comparisons.

python bench.py n=3 import=1