|3 days ago||3 days ago|
|GNU General Public License v3.0 or later||MIT License|
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
What projects *didn't* you make in Rust?
19 projects | reddit.com/r/rust | 4 Dec 2021
3 projects | reddit.com/r/linux | 3 Dec 2021
GPU computing on Apple Silicon
7 projects | reddit.com/r/rust | 30 Nov 2021
Now, I wish I could recommend Rust itself as a shading language, but https://github.com/EmbarkStudios/rust-gpu is still in its infancy.
Rust-CUDA: writing and executing extremely fast GPU code fully in Rust
It turns out that when wgpu is combined with rust-gpu, it can run rust on gpus too
(on the "builder" directory it builds shaders with the spirv-builder crate)
The best way to do it is probably the way rust-gpu does it: https://github.com/EmbarkStudios/rust-gpu/blob/main/docs/src...
The entry point of the kernel would supply any objects that have special properties.
Blender 3.0 takes support for AMD GPUs to the next level
5 projects | news.ycombinator.com | 16 Nov 2021
I'd ask why wouldn't you want that?
I think this is more promising: https://github.com/EmbarkStudios/rust-gpu
Has anyone seriously considered C++AMP? Thoughts / Experiences?
2 projects | reddit.com/r/gpgpu | 12 Nov 2021
Yes, Vulkan GPU source is split, though technically in a way that makes it more similar to CUDA. Vulkan uses an intermediate format instead of consuming text code directly, meaning new features are easier to add and frontend code doesn't need to be passed to the vendors driver compiler. SPIR-V is like DXIL or PTX code for CUDA, basically LLVM IR for GPUs. The CUDA compiler compiles your device code into PTX code, and it's what enables you to have "non split" source code. There's even an option to have separate PTX code in CUDA. There are few projects that aim to bring Vulkan SPIR-V into source, including Rust GPU for rust (though it will still have to be in a separate file) and Circle C++ shader for C++.
3 projects | reddit.com/r/rust | 30 Oct 2021
This is the first release of autograph rebuilt on SPIR-V compute shaders that can be compiled from Rust source with rust-gpu!
Doing M1 MacBook Pro (M1 Max, 64GB) Compile Benchmarks!
37 projects | reddit.com/r/rust | 26 Oct 2021
Also I want to know about https://github.com/EmbarkStudios/rust-gpu/.
The AMD “Aldebaran” GPU That Won Exascale
5 projects | news.ycombinator.com | 15 Nov 2021
Incorrect. Vulkan has compute shaders, which are generally usable. Libraries like VkFFT demonstrate basic signal processing in Vulkan. This can certainly be expanded upon, & there are numerous other non-graphical uses.
There is a Vulkan ML TSG (Technical Subgroup), which is supposed to be working on ML. Even Nvidia is participating, with extensions like VK_NV_cooperative_matrix, which specifically target the ml tensor cores.
There's people could probably say this better/more specifically, but I'll give it a try: Vulkan is, above all, an general standard for dispatching & orchestrating work usually on a GPU. Right now that work is most of often graphics, but that is far from a limit.
SYCL is, imo, the opposite of where we need to go. It's the old historical legacy that CUDA has, of writing really dumb ignorant code & hoping the tools can make it run well on a GPU. Vulkan, on the other hand, asks us to consider deeply what near-to-the-metal resources we are going to need, and demands that we define, dispatch, & complete the actual processing engines on the GPU that will do the work. It's a much much much harder task, but it invites in fantastic levels of close optimization & tuning, allows for far more advanced pipelining & possibilities. If the future is good, it should abandon SYCL and CUDA, and bother to get good at Vulkan.
VkFFT - GPU Fast Fourier Transform library API guide release
1 project | reddit.com/r/programming | 22 Aug 20211 project | reddit.com/r/vulkan | 22 Aug 2021
Hello, I am the creator of the VkFFT - GPU Fast Fourier Transform library for Vulkan/CUDA/HIP and OpenCL. In the last update, I have released explicit 50-page documentation on how to use the VkFFT API. It describes all the necessary steps needed to set up the VkFFT library and explains the core design of the VkFFT. This is a very important update, as it should make VkFFT easier to use.
VkFFT can now perform Fast Fourier Transforms of arbitrary length
1 project | reddit.com/r/CUDA | 26 Jul 2021
For general optimized radix transforms, you might want to check the benchmark plots on GitHub or my Nvidia GTC talk where I describe them in detail.1 project | reddit.com/r/programming | 25 Jul 20211 project | reddit.com/r/vulkan | 25 Jul 2021
Performace-wise, VkFFT achieves up to half of the device bandwidth in Bluestein's FFTs, which is up to up to 4x faster on <1MB systems, similar in performance on 1MB-8MB systems and up to 2x faster on big systems than Nvidia's cuFFT. VkFFT is also >5x times faster than AMD's rocFFT for Bluestein's systems. You can check the benchmark and precision scripts on VkFFT's GitHub repo.
VkFFT now supports Discrete Cosine Transforms
2 projects | reddit.com/r/vulkan | 19 Jun 2021
VkFFT supports convolution calculations - see samples 7, 8 and 9 in the VkFFT repository.
VkFFT now supports Discrete Cosine Transforms on GPU
1 project | reddit.com/r/programming | 19 Jun 2021
Announcing neuronika 0.1.0, a deep learning framework in Rust
7 projects | reddit.com/r/rust | 15 Jun 2021
The Vulkan docs online are quite good; if i remember they have "hello world" type tutorials as well. An take a look at something like Vulkan fft that has both Vulkan and opencl backends. Looking at "real" projects will of course tens to obscure the main points, but can be helpful for comparison.
Vulkan and HPC
1 project | reddit.com/r/vulkan | 6 Jun 2021
It's still pretty niche, honestly. Though https://github.com/DTolm/VkFFT is a good counterexample.
What are some alternatives?
are-we-learning-yet - How ready is Rust for Machine Learning?
wgpu - Safe and portable GPU abstraction in Rust, implementing WebGPU API.
RustaCUDA - Rusty wrapper for the CUDA Driver API
wgpu-rs - Rust bindings to wgpu native library
hash-shader - SHA256 Compute Shader (Kernel) Written in Rust
makepad - Makepad is a creative software development platform for Rust that compiles to wasm/webGL, osx/metal, windows/dx11 linux/opengl
crates.io - Source code for crates.io
kompute - General purpose GPU compute framework for cross vendor graphics cards (AMD, Qualcomm, NVIDIA & friends). Blazing fast, mobile-enabled, asynchronous and optimized for advanced GPU data processing usecases. Backed by the Linux Foundation.
DiligentEngine - A modern cross-platform low-level graphics library and rendering framework
DirectXShaderCompiler - This repo hosts the source for the DirectX Shader Compiler which is based on LLVM/Clang.
Rust-CUDA - Ecosystem of libraries and tools for writing and executing fast GPU code fully in Rust.