Hackernews posts about CUDA

CUDA is a parallel computing platform and programming model developed by NVIDIA that allows developers to leverage the processing power of graphics processing units (GPUs) for general-purpose computation.

RSS feed for CUDA

Related:Nvidia WSL ROCm

61
newCUDA for AMD on Windowsgithub.com
▲ 175chiassedu80 22d 💬 109
62
newShow HN: Agentic CUDA Kernel Optimizergithub.com
▲ 37bertaye 11d 💬 11
63
newCUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 12xiaoyu2006 26d discuss
64
newIntroducing CUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 12ubj 27d discuss
65
newKernelBench: Can LLMs Write GPU Kernels? – Benchmark and Toolkit, Torch –> CUDAgithub.com
▲ 6peter_d_sherman 4d discuss
66
newShow HN: CUDA/graphics in QEMU-KVM VMs without passing the Nvidia card to themgithub.com
▲ 5reindertpelsma 28d 💬 2
67
newCUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 5thomasfromcdnjs 19d discuss
68
newCUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 5Jhsto 27d discuss
69
new"The CUDA Moat is Gone" – TechTechPotato with AI& founder David Bennet [video]www.youtube.com
▲ 4olvy0 24d discuss
70
newCUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 4vertigoruntime 28d discuss
71
newShow HN: CUDA-verkle – GPU-accelerated 256-MSM for Verkle tree commitmentsgithub.com
▲ 3grobat79 13d discuss
72
newZLUDA – Drop-In Replacement for CUDA on Non-Nvidia GPUsgithub.com
▲ 2peter_d_sherman 22d 💬 1
73
newPTXBench: What about just CUDA-PTX?zhang677.github.io
▲ 2matt_d 6d discuss
74
newCUDA Device Max Connectionsleimao.github.io
▲ 2eigenBasis 9d discuss
75
newNvidia Expands Open Source CUDA-Q Platform for Fault-Tolerant Quantum Computingnvidianews.nvidia.com
▲ 2noobplus 20d discuss
76
newInto the CUDA Multiverse of Runtime Compilation: GPU DBMS Kernel Fusion [pdf]www.vldb.org
▲ 2matt_d 20d discuss
77
newCUDA Rust: Two Tracks for Writing GPU Kernelsdeveloper.nvidia.com
▲ 2CoderLim110 22d discuss
78
newDeepJIT: Header-Only C++20 JIT Runtime for Nvidia CUDA and Huawei Ascendgithub.com
▲ 2mehmetoguzderin 26d discuss
79
newShow HN: Reflex – a GGUF/CUDA inference engine tuned for cold-start latencygithub.com
▲ 1leochong 12d 💬 1
80
newMost CUDA tutorials introduce too much too earlycomputejunkie.substack.com
▲ 1ell-hol 3h discuss
81
newCUDA Tile (cuTile) first working port in Javagithub.com
▲ 1birdculture 22d discuss
82
newLaunch HN: Magnitude (YC S25) – Self-optimizing inference engine for agentsgithub.com
▲ 194anerli 5d 💬 99
83
newShow HN: TurboGPT: train 22KiB transformer in 13sgithub.com
▲ 56lostmsu 6d 💬 10
84
newShow HN: InstinctFlash – High-Performance Serving Runtime for Robotics Modelsgithub.com
▲ 27guanming0717 13d 💬 4
85
newShow HN: Jinfer – AI inference engine for the JVM. AI in a jarqxotic.ai
▲ 6mukel 21d 💬 1
86
newShow HN: Optimized runtimes for three VLAs on Jetson Thorgithub.com
▲ 2hhuytho 13d discuss
87
newFlashRec – Wide-beam generative recommendation. Every SID is a real item
▲ 1vect0ry 7d discuss
88
newShow HN: PicoLM v1.0-rc2 ("Yura Kana"). Run an LLM on Digital Unixgithub.com
▲ 1gabucino 12d discuss
89
newNvidia announces native GPU programming in Rustdeveloper.nvidia.com
▲ 970nonmaskable 20d 💬 404
90
newDeepSeek open sourced their Huawei Ascend programming stackaistockwire.com
▲ 5ashish01 3d discuss

More stories →