Skip to content

PiotrWolinski/mlx-performance-test

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MLX performance comparison

Overview

Sole purpose of this repository is to play around with MLX, to see what are the possible speed ups on the Apple silicon.
So far it only contains simple matrix multiplication, but I have plans to grow this collection further.

Setup

Those simple tests are run on the 2021 14" MacBook Pro with M1 Pro (8 cores) processor and 32GB of memory.

Results

Dot product

Dot product using mlx turns out to be on the average (over 10 runs) 12.5 times faster than numpy implementation.

Test it yourself

Simply install dependencies using poetry and run chosen test.

About

Playing around with MLX framework.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages