89 points by matrixmagician 6 months ago flag hide 16 comments
wonder_woman 6 months ago next
Does the algorithm work for complex numbers and different data types? I'd be curious to know.
batman 6 months ago next
The algorithm should work as long as the data type being used supports the necessary arithmetic operations. It should be possible to modify the code for complex numbers.
john_doe 6 months ago prev next
Fascinating! I've been looking for faster matrix multiplication algorithms. Will need to try this in my project.
the_terminator 6 months ago next
Rust is an excellent choice for this, well done to the author.
sarah_connor 6 months ago prev next
I wonder if it's possible to further optimize this for GPUs? That could be very interesting.
john_doe 6 months ago next
That's a good thought. Might need to implement some CUDA code for that use case.
apollo11 6 months ago prev next
I've heard Rust is great for performance-critical code. This is a great example of that.
buzz_lightyear 6 months ago next
It really is! I've been using it for high-performance simulations, and it's been a game changer.
super_man 6 months ago prev next
What kind of speedup did you get compared to a standard matrix multiplication? I'm quite curious.
clark_kent 6 months ago next
I'd also like to know that. This is quite interesting!
spiderman 6 months ago prev next
I wonder if it would make sense to make a crate out of this and publish it on crates.io. This could help a lot of people.
hulk 6 months ago next
I agree, that would be a great idea. I'd love to give it a try!
black_widow 6 months ago prev next
Yes, this would be an incredibly useful contribution to the Rust ecosystem.
tony_stark 6 months ago prev next
Matrix multiplication is a very important operation in many areas, like machine learning. This could be a big help to the Rust ecosystem.
thor 6 months ago next
Indeed, I'm sure many developers in the community would find this to be very valuable.
iron_ma 6 months ago prev next
I'm impressed with how comprehensive and well-thought-out this implementation is.