# Outperformed by Matlab

**URL:** <https://discourse.julialang.org/t/outperformed-by-matlab/71603>\
**Category:** Performance\
**Tags:** matlab, multithreading, linearalgebra, tullio, loopvectorization\
**Created:** [November 16, 2021, 5:58pm UTC](https://discourse.julialang.org/t/outperformed-by-matlab/71603 "2021-11-16T17:58:00Z")\
**Posts on this page:** 1\
**Showing post:** 56

<div class="post-metadata">

**Author:** ![goerch](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/goerch/32/29122_2.png) [@goerch](https://discourse.julialang.org/u/goerch)\
**Post date:** [November 23, 2021, 7:28pm UTC](https://discourse.julialang.org/t/outperformed-by-matlab/71603/56 "2021-11-23T19:28:46Z")

</div>

Lessons learned:

1. Regarding `kron`: we had [this](https://discourse.julialang.org/t/micro-benchmarks-of-kron-for-vectors/71689) discussion, which shows room to improve performance for [Kronecker](https://en.wikipedia.org/wiki/Kronecker_product) products and their kind like Hadamard and Khatri-Rao, probably.
2. Regarding basic linear algebra: we had [this](https://discourse.julialang.org/t/mul-dispatch-to-blas-incomplete/71831) and [that](https://github.com/JuliaLang/julia/issues/43175) discussion about how to dispatch high level linear algebra operations to OpenBLAS or MKL.

For now I’d recommend users coming from MATLAB and seeking high performance to look into the lower level BLAS interfaces. Would it be sensible to update the documentation to reflect this experience?

P.S.: a word on the differences w.r.t. the selected BLAS implementation wouldn’t hurt, also.

---

_[View the full topic](https://discourse.julialang.org/t/outperformed-by-matlab/71603)._