Hackernews posts about Blackwell

  1. Guess-Verify-Refine: Data-Aware Top-K for Sparse-Attention Decoding on Blackwell (arxiv.org)
    4 points by matt_d 20 days ago | 1 comments
  2. Pyptx – A Python DSL to Write Nvidia PTX for Hopper and Blackwell (github.com)
    3 points by jxmorris12 20 days ago | discuss
  3. Dynamic Persistent Tile Scheduling w/ Cluster Launch Control (CLC) on Blackwell (research.colfax-intl.com)
    2 points by matt_d 6 days ago | discuss
  4. Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs (arxiv.org)
    2 points by matt_d 18 days ago | discuss
  5. Claude Code Opus 4.7 keeps checking on malware
    69 points by decide1000 29 days ago | 63 comments
  6. Show HN: A web-based replacement for Nvidia's CUDA occupancy spreadsheet (toolbelt.widgita.xyz)
    1 points by fairlight1337 27 days ago | 1 comments
  7. SonicMoE: A HW-Efficient and SW-Extensible Blueprint for Fine-Grained MoEs (dao-lab.ai)
    2 points by matt_d 24 days ago | discuss
  8. 15 years of FP64 segmentation, and why the Blackwell Ultra breaks the pattern (nicolasdickenmann.com)
    202 points by fp64enjoyer 3 months ago | 80 comments
  9. Nvidia has produced the first Blackwell wafer on US soil (www.xda-developers.com)
    164 points by kristianp 7 months ago | 64 comments
  10. Blackwell: Nvidia's GPU (chipsandcheese.com)
    112 points by pella 11 months ago | 31 comments
  11. Nvidia Blackwell GeForce RTX 50 Series Opens New World of AI Computer Graphics (nvidianews.nvidia.com)
    76 points by OnlineInference over 1 year ago | 23 comments
  12. Nvidia's Blackwell Reworked – Shipment Delays and GB200A Reworked Platforms (www.semianalysis.com)
    67 points by charleshn almost 2 years ago | 6 comments
  13. Writing high-performance matrix multiplication kernels for Blackwell (docs.jax.dev)
    63 points by lairv 8 months ago | 6 comments
  14. SVDQuant+NVFP4: 4× Smaller, 3× Faster FLUX with 16-bit Quality on Blackwell GPUs (hanlab.mit.edu)
    52 points by lmxyy about 1 year ago | 10 comments
  15. Matmul on Blackwell: Part 2 – Using Hardware Features to Optimize Matmul (www.modular.com)
    23 points by robertvc 8 months ago | 13 comments
  16. AMD launches AI chip to rival Nvidia's Blackwell (www.cnbc.com)
    20 points by teleforce over 1 year ago | discuss
  17. Nvidia Puts Grace Blackwell on Every Desk and at Every AI Developer's Fingertips (nvidianews.nvidia.com)
    11 points by TiredOfLife over 1 year ago | 4 comments
  18. Nvidia Unveils Rubin CPX Amidst Chart-Topping Blackwell Ultra MLPerf Results (hothardware.com)
    11 points by rbanffy 8 months ago | discuss
  19. Nvidia RTX Blackwell GPU Architecture [pdf] (images.nvidia.com)
    10 points by ydnyshhh over 1 year ago | discuss
  20. Nvidia DGX Station – The GB300 Grace Blackwell 748 GB VRAM Workstation (www.nvidia.com)
    9 points by car about 2 months ago | 4 comments
  21. Nvidia and TSMC Celebrate First Nvidia Blackwell Wafer Produced in the US (blogs.nvidia.com)
    9 points by jonbaer 7 months ago | 1 comments
  22. Blackwell: Nvidia's GPU (old.chipsandcheese.com)
    9 points by mfiguiere 11 months ago | 1 comments
  23. Nvidia Blackwell Ultra (GB300): -97% INT8/FP64,+50% FP4 Dense,+55% VRAM,+114% At (resources.nvidia.com)
    8 points by Sweepi 9 months ago | 3 comments
  24. Nvidia Blackwell Delivers World-Record DeepSeek-R1 Inference Performance (developer.nvidia.com)
    8 points by daviducolo about 1 year ago | discuss
  25. AI inference costs dropped up to 10x on Nvidia's Blackwell (venturebeat.com)
    7 points by CrankyBear 3 months ago | 3 comments
  26. Nvidia Tensor Core Evolution: From Volta to Blackwell (semianalysis.com)
    6 points by pella 11 months ago | discuss
  27. NVIDIA to Manufacture Blackwell AI Supercomputers in US for First Time (blogs.nvidia.com)
    6 points by walterbell about 1 year ago | discuss
  28. Nvidia Reportedly Delays Next-Generation Blackwell AI Chip Due to Design Flaw (www.investors.com)
    6 points by ComputerGuru almost 2 years ago | discuss
  29. Nvidia announces $3000 128gb Blackwell desktop GPU/CPU mini super computer (nvidianews.nvidia.com)
    5 points by h_tbob over 1 year ago | 2 comments
  30. Nvidia GPU: Discussing Blackwell's Limitations and Predicting Rubin (github.com)
    5 points by yankcrime 4 months ago | discuss
  31. We rebuilt MXFP8 MoE kernels from scratch to run 3.5x faster on Blackwell (cursor.com)
    5 points by ecz 9 months ago | discuss
  32. 384GB Personal AI Workstation with Four Nvidia RTX 6000 Pro Blackwell Max-Q GPUs (a16z.com)
    4 points by transpute 9 months ago | 2 comments
  33. Nvidia Blackwell Ultra: The Chip Powering the AI Factory Era (developer.nvidia.com)
    4 points by ksec 9 months ago | 1 comments
  34. NVIDIA confirms Blackwell support for OCP microscaling formats MXFP8, MXFP4 (developer.nvidia.com)
    4 points by opcode84 over 1 year ago | 1 comments
  35. InferenceX v2: Nvidia Blackwell vs AMD vs. Hopper – SemiAnalysis (newsletter.semianalysis.com)
    4 points by randomgermanguy 3 months ago | discuss
  36. Inference Providers Cut AI Costs by 10x with Open Source Models on Blackwell (blogs.nvidia.com)
    4 points by gmays 3 months ago | discuss
  37. Grace Blackwell Desktop Supercomputer: First Impressions (jasoneckert.github.io)
    4 points by speckx 6 months ago | discuss