Compares matrix multiplication implementations with timeit

numpy has a very fast implementation of matrix multiplication. There are many ways to be slower. The following uses timeit to compare implementations.

Compared implementations:

• multiply_matrix code

• c_multiply_matrix code

• c_multiply_matrix_parallel code

• c_multiply_matrix_parallel_transposed code

Preparation

import timeit
import numpy

from td3a_cpp.tutorial.td_mul_cython import (
    multiply_matrix, c_multiply_matrix,
    c_multiply_matrix_parallel,
    c_multiply_matrix_parallel_transposed as cmulparamtr)


va = numpy.random.randn(150, 100).astype(numpy.float64)
vb = numpy.random.randn(100, 100).astype(numpy.float64)
ctx = {
    'va': va, 'vb': vb, 'c_multiply_matrix': c_multiply_matrix,
    'multiply_matrix': multiply_matrix,
    'c_multiply_matrix_parallel': c_multiply_matrix_parallel,
    'c_multiply_matrix_parallel_transposed': cmulparamtr}

Measures

numpy

res0 = timeit.timeit('va @ vb', number=100, globals=ctx)
print("numpy time", res0)

numpy time 0.029423824977129698

python implementation

res1 = timeit.timeit(
    'multiply_matrix(va, vb)', number=10, globals=ctx)
print('python implementation', res1)

python implementation 36.37305003963411

cython implementation

res2 = timeit.timeit(
    'c_multiply_matrix(va, vb)', number=100, globals=ctx)
print('cython implementation', res2)

cython implementation 0.73594726389274

cython implementation parallelized

res3 = timeit.timeit(
    'c_multiply_matrix_parallel(va, vb)', number=100, globals=ctx)
print('cython implementation parallelized', res3)

cython implementation parallelized 0.10048561217263341

cython implementation parallelized, AVX + transposed

res4 = timeit.timeit(
    'c_multiply_matrix_parallel_transposed(va, vb)', number=100, globals=ctx)
print('cython implementation parallelized avx', res4)

cython implementation parallelized avx 0.042188897263258696

Speed up…

print(f"numpy is {res1 / res0:f} faster than pure python.")
print(f"numpy is {res2 / res0:f} faster than cython.")
print(f"numpy is {res3 / res0:f} faster than parallelized cython.")
print(f"numpy is {res4 / res0:f} faster than avx parallelized cython.")

numpy is 1236.176808 faster than pure python.
numpy is 25.011951 faster than cython.
numpy is 3.415110 faster than parallelized cython.
numpy is 1.433835 faster than avx parallelized cython.