Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Rust SIMD on the GPU

vectorware.com

Post reply on HN
Print view View original

31–40 of 130 posts

Re: Rust SIMD on the GPU

#31

Post by LegNeato »

Eridrus wrote: post #29
LegNeato wrote: post #4

Author here, AMA.

Given the massive demand for GPUs for LLMs, what sorts of work do you expect to economically benefit from utilizing GPUs more?

Part of our thesis is that decent GPUs are in every shipping device and most software doesn't use them and should.

Re: Rust SIMD on the GPU

#32

Post by O3marchnative »

The author mentions Rust's portable SIMD library [0]. The only issue with portable SIMD is it's only available on nightly. I used it in my FFT crate, but we had to switch to the fearless_simd crate in order to get a portable SIMD solution that works on stable [1].

[0] https://doc.rust-lang.org/std/simd/index.html

[1] https://github.com/linebender/fearless_simd

Re: Rust SIMD on the GPU

#33

Post by nynx »

Do you have examples of complex algorithms running on the gpu with rust with competative performance? Radix sort might be a good one to start with

Re: Rust SIMD on the GPU

#34

Post by LegNeato »

lbhdc wrote: post #18
LegNeato wrote: post #4

Author here, AMA.

This is really cool! It sounds like y'all have a compiler fork that you are using to make this work. I wanna tinker with this, is your compiler available?

It is not currently available but we intend to make it available after we launch our products.

Re: Rust SIMD on the GPU

#35

Post by Eridrus »

LegNeato wrote: post #31
Eridrus wrote: post #29

Earlier quoted context omitted.

Given the massive demand for GPUs for LLMs, what sorts of work do you expect to economically benefit from utilizing GPUs more?

Part of our thesis is that decent GPUs are in every shipping device and most software doesn't use them and should.

I guess you're looking at consumer hardware then since servers have exactly what you pay for.

Can you say more about the application space you're targeting?

Re: Rust SIMD on the GPU

#36

Post by lbhdc »

LegNeato wrote: post #28
lbhdc wrote: post #21

Earlier quoted context omitted.

What is vectorware's business model? Are you planning to sell support/consulting to companies using your stack? Or are you looking to sell licenses to your tool? Or something else?

The tentative plan is to open source all the compiler and `std` bits with our products built on top (compilers are not good businesses). More about our products coming in the next couple of months!

Looking forward to reading more about it. Good luck on the launch :)

Re: Rust SIMD on the GPU

#37

Post by camel-cdr »

I love how ever example of portable SIMD isn't portable.

They specifies a constant SIMD width so it's non-portable. Well, not performance portable, but why are we using SIMD again?

Re: Rust SIMD on the GPU

#38

Post by ismailmaj »

6r17 wrote: post #15

My heard hurts - i was stupid enough to think that SIMD was a CPU only thing - I don't understand why it would be ported to GPU - huge kudos to managing to surprise me

There is something very SIMD-coded in GPU programming which is coalesced stores/loads, if a warp (32 threads) handles contiguous memory, it will create ~4 transactions instead of 32.
Post reply on HN