NVIDIA Nsight Compute Now Profiles CUDA Tile Kernels — Two Changes Cut Kernel Time 81%
NVIDIA Developer · youtube · 2026-09-05
An official NVIDIA Developer tutorial: Nsight Compute can now profile CUDA Tile kernels directly. CUDA Tile lets you write CUDA at a higher level using multidimensional Tiles while the compiler handles hardware mapping — raising the question of how to know whether the kernel is fast when the compiler decides how it runs. Nsight Compute closes that visibility gap.
It connects GPU performance data back to your CUDA Tile kernel, identifies the limiting factor, and points to concrete optimizations, while keeping visibility into occupancy, memory access behavior, generated SASS and source-level performance. The video shows measurable results: two targeted changes (occupancy hints and adjusted tile block dimensions) cut kernel duration by 81%.
Also covered: mapping metrics and stalls to individual Python/C++ source lines and compiled SASS, and using comparison tools to trace gains and monitor uncoalesced memory accesses.
More from Infra
- llama.cpp team launches Llama: one-click local AI on your Mac menu bar — Hugging Face · 2026-10-23
- China's daily token use hits 500T, up 4x as average price doubles to ¥13.46 — teortaxesTex · 2026-10-23
- New all-reduce kernels deliver 12% decode speedup for DeepSeek-V4.1 on PCIe-connected GPUs — AlpinDale · 2026-10-23
- Offshore oil rigs could be converted into AI data centers powered by stranded gas — clemnt · 2026-10-23
- Same native UI, two stacks: Swift build is ~80MB Mac-only, Zig build is ~10MB cross-platform — jedisct1 · 2026-10-23
- SurfSense: an open-source local NotebookLM with MCP agents and a prompt-caching lesson — FurtiveMirth · 2026-10-23