https://sankethbk.github.io/blog/tags/occupancy/Recent content in Occupancy on Sanketh's BlogHugo -- 0.166.0en-usTue, 01 Sep 2026 00:00:00 +0000https://sankethbk.github.io/blog/posts/gpu/2026-09-01-2-cuda-occupancy/Tue, 01 Sep 2026 00:00:00 +0000https://sankethbk.github.io/blog/posts/gpu/2026-09-01-2-cuda-occupancy/<h1 id="cuda-occupancy-filling-the-sm-to-hide-latency">CUDA Occupancy: Filling the SM to Hide Latency</h1> <p>The <a href="https://sankethbk.github.io/blog/gpu/2026-09-01-1-cuda-shared-memory-tiling">previous note</a> showed how to use shared memory to fix uncoalesced access patterns. This note covers <strong>occupancy</strong>: how many warps can live on an SM at the same time, and why that number determines whether the GPU can hide memory latency.</p> <p>Occupancy is the first CUDA performance topic that is not about memory access at all. It is about keeping the warp schedulers busy.</p>