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
Welcome to the lucky 10,000! SIMD is actually a pretty integral part of how GPUs are able to work efficiently, it's part of why there's such a strong focus on branchless programming in the field.
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!
The post is kind of vague on the IR you're targeting. Can you give some examples of what the SIMD-ized IR looks like, and how it maps to the target PTX?
I'm confused too. How does this fit between these approaches for paraellization:
- CUDA kernels and Tiles (e.g. Cudarc, cuda-oxide, rust-gpu etc) - SIMD on the GPU. (E.g. as in the title...)
- CPU SIMD using avx or SSE instructions (And probably thin wrappers for vectors so you can have sane syntax). Or the maybe-upcoming core simd which should abstract over architecture-specific instructions. Magic floats etc which do 4-16 computations at once, but are a bit clumsy to work with
- Rayon thread pools - arbitrary parallel computations, including SIMD, one per CPU core.
It looks like from the code samples like maybe a cleaner syntax for writing code on the GPU than CUDA kernels? E.g. without mucking with serialization, host and device by abstracting over it? And inspired by core::simd. (Good choice if so, in the interest of standardizing on syntax; I did this for my x86 SIMD vector/quaternion lib as well)
Each family of operations is a trait parameterized by the operation itself:
pub trait EvaluateReduction<Operation, T>: LaneEvaluator {
/// Reduce one distributed definition to an ordinary uniform scalar.
fn evaluate_reduction(&self, value: LaneValue<Self, role::Distributed, T>) -> T;
}
Call sites name the operation:
let one = evaluator.splat::<Splat, _>(1_u32);
let two = evaluator.splat::<Splat, _>(2_u32);
let three = evaluator.binary::<Add, _>(one, two);
let total = evaluator.reduce::<Sum, u32>(three); // a uniform u32
let running = <Executor as EvaluateScan<Scan<Sum, Exclusive>, u32>>::scan(&evaluator, three);
Operations like Sum, Max, ReduceXor, Inclusive, and Exclusive are all distinct types.
As mentioned in the post, execution shape is typed too. A static shuffle takes its control as a type-level constant, and the shuffle mode constrains which controls are expressible:
// Shift down one lane, keeping our own value where the source is inactive.
let down = <Executor as EvaluateShuffle<Shuffle<Down>, DownOrSelf<1>, u32>>::shuffle(&ev, v);
// Broadcast from lane zero.
let bcast = <Executor as EvaluateShuffle<Shuffle<Broadcast>, WarpLane<0>, u32>>::shuffle(&ev, down);
// Butterfly exchange with the neighbor one bit away.
let bfly = <Executor as EvaluateShuffle<Shuffle<Xor>, Butterfly<1>, u32>>::shuffle(&ev, bcast);
For an example of errors caught, a warp-scoped executor for a device-scoped barrier is a compile error:
<ScopedWarpExecutor<'_, WarpUniform> as EvaluateBarrier<Barrier<Device>>>::barrier(evaluator)
// error[E0277]: the trait bound `Device: NvptxBarrierScope` is not satisfied
// help: the trait `NvptxBarrierScope` is implemented for `Warp`
Strip mining is typed on the amount of work and the lane capacity, and it hands back one chunk at a time along with the predicate saying which lanes live in that chunk:
// Six work items across four active lanes: two chunks, based at 0 and 4.
<Executor as EvaluateStripMine<StripMine, (WorkItems, ActiveLanes<StripMined<4>>), i32>>::
for_each_strip_mined(
&evaluator,
(WorkItems::new(6)?, ActiveLanes::new(4)?),
|index, active| {
// ...
},
);
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?
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
GPUs work on vectors and matrices very often, that's what they are good at, so it makes a lot of sense that they can operate with SIMD I think!
Welcome to the lucky 10,000! SIMD is actually a pretty integral part of how GPUs are able to work efficiently, it's part of why there's such a strong focus on branchless programming in the field.
Author here, AMA.
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!
The post is kind of vague on the IR you're targeting. Can you give some examples of what the SIMD-ized IR looks like, and how it maps to the target PTX?
I'm confused too. How does this fit between these approaches for paraellization:
It looks like from the code samples like maybe a cleaner syntax for writing code on the GPU than CUDA kernels? E.g. without mucking with serialization, host and device by abstracting over it? And inspired by core::simd. (Good choice if so, in the interest of standardizing on syntax; I did this for my x86 SIMD vector/quaternion lib as well)Didn't want to go into crazy detail in the post.
Each family of operations is a trait parameterized by the operation itself:
Call sites name the operation: Operations like Sum, Max, ReduceXor, Inclusive, and Exclusive are all distinct types.As mentioned in the post, execution shape is typed too. A static shuffle takes its control as a type-level constant, and the shuffle mode constrains which controls are expressible:
For an example of errors caught, a warp-scoped executor for a device-scoped barrier is a compile error: Strip mining is typed on the amount of work and the lane capacity, and it hands back one chunk at a time along with the predicate saying which lanes live in that chunk: Hopefully that gives the flavor of it.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.
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?
Congrats to the Rust-GPU folks! Nice to see the good work flowing.
Good job!
Hey - this is probably off-topic/meta, but what is going on with the comments here? Is it bots?
No idea, but it seems HN needs POW challenges.
Could also just be trolls attracted by the Rust topic.