#Cuda

Run CUDA-targeted Windows applications on AMD GPUs with ZLUDA + ROCm/HIP. - Speedstu/CUDA-for-AMD-Windows

A humble exploratory PoC for a hardware-native, optical timing-frozen control plane engine. Fuses inline CUDA PTX, RAII memory tunnels, and 4D JAX shard_map structures to investigate 0ns-overhead f...

Use Nix and NixOS on your DGX Spark! Contribute to graham33/nixos-dgx-spark development by creating an account on GitHub.

Tracing one vector-add kernel from nvcc all the way down to the warps that execute it.

cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions. -...

Nvidia's RTX Spark is competing in a space that AMD kickstarted over a year ago.

TL;DR: We profiled a LiDAR object detector expecting the 3D backbone to dominate. It didn't....

The N1X reportedly comes in two SKUs: a top-end 20-core option with 6,144 CUDA cores matching the desktop RTX 5070, and a cut-down 18-core option with 5,120 CUDA cores.…

NVIDIA CUDA 13.3 brings new capabilities and performance optimizations to developers across the CUDA ecosystem. The launch of NVIDIA CUDA Tile programming in…

Developers can now use NVIDIA CUDA Tile programming within large existing C++ GPU codebases to develop highly optimized GPU kernels using tile-based…

TL;DR After del tensor; torch.cuda.empty_cache(), PyTorch's caching allocator still...

Introduction You've written model.to('cuda') a hundred times. You've celebrated when...

From Dev Community: RTX 5090, LLaMA.cpp TurboQuant, & Blackwell CUDA Scheduling Boosts GPU Performance

A practical guide to debugging silent memory corruption in CUDA kernels, with compute-sanitizer workflows and a look at Rust-on-GPU tooling.

From Dev.to - tutorial: How to Deploy Llama 3.2 90B with vLLM + Speculative Decoding on a $16/Month DigitalOcean GPU Droplet: 2.5x Faster Inference at 1/110th Claude Cost

Nvidia's real AI moat isn't "a piece of hardware," writes Wired's Sheon Han. It's CUDA: a mature, deeply optimized software ecosystem that keeps machine-learning workloads tied to Nvidia GPUs.…

From Dev RSS Feed: RTX 5080 Launched, Rust for CUDA, & LLM GPU Scheduling Deep Dive

NVlabs publica el primer compilador oficial que convierte Rust estándar en kernels CUDA. Mismo archivo .rs para CPU y GPU, sin C++ de por medio.

cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign language bindings, just Rust.