When using pyjulia to compute a stack of inverse matrices, I find it’s slower than directly using numpy.linalg.inv. But this shows using julia is faster than using python. This makes me confused. Here is my julia script file named “jlinv.jl”: using StaticArrays, Einsum function jlinv(m) sizem_…

A stack of inverse matrices calculation when comparing to python

Fesh November 11, 2021, 5:00pm 4

I already ran getjlinv more than once before I posted cost-time screenshot. The reason might not be the compilation. Maybe I should use NumPy directly instead of pyjulia.

Topic		Replies	Views
Slow matrix multiplication in Julia compared to Python numpy New to Julia question	17	5623	May 19, 2018
Numpy 10x faster than Julia ?! What am I doing wrong ?! [solved - julia faster now] Performance question	37	11135	October 15, 2019
LinearAlgebra.inv matrix inversion slower than MATLAB? Performance performance	21	2232	October 25, 2021
Computing Inverse of a stack of matrices General Usage performance , linearalgebra	19	3684	August 29, 2020
Pseudo-inverse of large matrix very slow New to Julia question , performance , linearalgebra , python	20	4321	January 12, 2022

A stack of inverse matrices calculation when comparing to python

Related topics