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.

1. What happens when you run a CUDA kernel? (fergusfinn.com)
   294 points by mezark 9 days ago | 32 comments
2. Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs) (vosen.github.io)
   163 points by Tiberium 8 days ago | 14 comments
3. TorchCodec 0.14: HDR Video Decoding for CPU and CUDA, and Fast Wav Decoder (github.com)
   62 points by scott_s 28 days ago | 13 comments
4. Show HN: NanoEuler – GPT-2 scale model in pure C/CUDA from scratch (github.com)
   55 points by vforno 10 days ago | 26 comments
5. What about OpenCL and CUDA C++ alternatives? (www.modular.com)
   25 points by eatonphil 29 days ago | 4 comments
6. Show HN: Continuous Nvidia CUDA PC Sampling Profiler (www.polarsignals.com)
   14 points by gnurizen 23 days ago | 6 comments
7. Show HN: NanoEuler – GPT-2 scale model in pure C/CUDA from scratch (github.com)
   6 points by vforno 19 days ago | 3 comments
8. Show HN: CUDA Profiler for Production Inference (github.com)
   6 points by npgraph 15 days ago | discuss
9. CUDA Profiler for Production Inference (graphsignal.com)
   5 points by npgraph 16 days ago | discuss
10. Show HN: FlashQwen – A from-scratch CUDA inference engine for Qwen3 (github.com)
    5 points by langtang1996 23 days ago | discuss
11. Show HN: A free, GPU-accelerated Texas Hold'em GTO solver in C++/CUDA (bupticybee.github.io)
    4 points by bupticybee 1 day ago | 2 comments
12. Running a 35B MoE model on a 2017 AMD RX 580 8GB via Vulkan (no ROCm/CUDA) (github.com)
    4 points by aivisionslab 18 days ago | discuss
13. Fast Great-Circle Distance Calculation in CUDA C++ (developer.nvidia.com)
    4 points by Alien1Being 18 days ago | discuss
14. Zluda (CUDA-compatible runtime for AMD) loses funding again, gains 32bit compat (vosen.github.io)
    3 points by indrora 8 days ago | 1 comments
15. Optimizing a CUDA FSST decompression kernel (www.polarsignals.com)
    3 points by asubiotto 14 days ago | discuss
16. Running local AI on AMD RX 580 (2017 GPU) using Vulkan – no CUDA, no ROCm (setup-ia-local-rx580-vulkan.web.app)
    3 points by aivisionslab 19 days ago | discuss
17. Compiling Strassen-Like Matrix Multiplication Algorithms to Fast CUDA Kernels (dl.acm.org)
    3 points by matt_d 23 days ago | discuss
18. AutoMegaKernel: Compiling a LLM into a single CUDA kernel (arxiv.org)
    3 points by OsamaJaber 29 days ago | discuss
19. CUDA-like programming of Cerebras WSE (github.com)
    2 points by gdiamos 24 days ago | 1 comments
20. Nvidia CUDA Python 1.0 and CUDA 13.3 Release (developer.nvidia.com)
    2 points by ashvardanian 29 days ago | discuss
21. Show HN: Run Agent Skills with mistral.rs v0.8.10: /v1/skills support and more
    20 points by ericlbuehler 21 days ago | 2 comments
22. Ask HN: Are people generally interested using LLMs for learning purposes?
    7 points by iknownthing 15 days ago | 11 comments
23. Show HN: Sipp – Run small local LLMs in browser 3x faster (www.sipp.sh)
    5 points by jjhartmann 14 days ago | 3 comments
24. Show HN: PyTorch on Java (github.com)
    2 points by pdsminer 27 days ago | discuss
25. Reduce GVisor Cold Starts with GPU Snapshotting (cerebrium.ai)
    51 points by jono_irwin 7 days ago | 17 comments
26. Open-source Vulkan driver NVK gains experimental DLSS support (www.tomshardware.com)
    3 points by rndsignals 17 days ago | discuss
27. The U.S. Is Terrorizing Cuba to Make Rich Men Richer (www.currentaffairs.org)
    72 points by robtherobber 27 days ago | 60 comments
28. Spain's cadastre API is SOAP from 2003, so I built a JSON wrapper (+MCP) (prediohq.com)
    17 points by guillermofc 3 days ago | 10 comments
29. Cuba approves unprecedented free-market reforms to stave off economic collapse (www.cbsnews.com)
    8 points by spikels 19 days ago | discuss
30. Cuba to Privatize State Companies (www.miamiherald.com)
    6 points by paulpauper 18 days ago | discuss