Sathvik Swaminathan
RSS FeedI'm Sathvik Swaminathan, a Silicon Design Engineer at AMD, previously a Research Scientist at Intel Labs. My background is CPU performance analysis and memory management in tiered memory systems. I'm now learning GPUs, and documenting it here.
Social Links:
Featured
Tiling Hierarchy on GPUs
GPUs are equipped with a tremendous amount of compute and a tiered memory architecture. How data is placed and accessed across these tiers determines how much of that compute is actually used.
Recent Posts
GPU Architecture
How GPUs trade latency for throughput, and why the SIMT execution model, the register file, and the memory system all follow from that one decision.