Benchmark MATLAB & Julia for Matrix Operations

[RoyiAvital]

@RoyiAvital

Hi,

My first try with Julia was to compare its speed on Matrix Operations (Linear Algebra oriented) to MATLAB.

Benchmark MATLAB & Julia for Matrix Operations

As any test, it is far from being perfect, but it is telling something about the picture.

Take this benchmark in its context, it’s not about which is faster in general but which is faster in the context of Matrix and Linear Algebra operations.

Few remarks:

• This is my first time coding with Julia, I’d be happy to hear if I did something wrong and results could be much improved (This is the reason I can’t make any conclusions).
• Ideas of more Matrix / Linear Algebra oriented tests are welcome.

For me curiosity is the fuel for great voyage, so I really liked trying and experiencing new things.

P. S.
This all started here:
Why Julia can not provide MKL? · Issue #18374 · JuliaLang/julia · GitHub

No it is not.
It is optimized if you explicitly loop on its elements. Since it is one time use there is no gain in that as it is allocated in the assignment.
Whats happening is allocation and pointer naming. Exactly as if I did pre allocation and use fused assignment.

Moreover, it is the same as in MATLAB. Since the idea here is to compare to MATLAB on Linear Algebra related operations, there is no real issues here.