#nvidia
14pages tagged with "nvidia"
programming
- china ai hardware decoupling notes— ongoing collection of notes on the evolving china-us ai hardware decoupling
- cuda 13.0 release overview— comprehensive guide to nvidia cuda 13.0 features, changes, and migration considerations
- cuda docker containers— gpu acceleration in docker containers with nvidia container toolkit
- cuda gpu computing setup— overview and decision guide for cuda/gpu setup on linux systems
- cuda native installation— installing cuda toolkit directly on linux for development and compilation
- gh200 setup for pytorch training (october 2026)— a tested, reproducible pytorch training environment on an nvidia gh200 grace hopper box: uv, torch 2.13 cu129 on aarch64, torch.compile, document-masked attention without flash-attn, moe grouped gemm kernels, and the gotchas
- nvidia gpu compute capability reference— complete list of nvidia gpus and their cuda compute capabilities
- pytorch cuda setup for nvidia gh200 (arm64)— comprehensive guide to configuring pytorch with cuda support on nvidia gh200 grace hopper systems running arm64 architecture
- small encoder training throughput on consumer gpus— measured training throughput for 5m and 21m parameter transformer encoders across rtx 5070 ti, 4070 ti super, 4060 ti, and zen 4 cpus
data centers
- nvidia -- china export controls investigation— pre-litigation investigation into NVIDIA for statements about Chinese buyers circumventing U.S. export controls on AI chips, with $5.5B H20 chip charge
- nvidia -- crypto revenue misrepresentation— long-running securities class action against NVIDIA for understating cryptocurrency mining revenue, surviving Supreme Court review and remanded for discovery