Upstream UpdateNullstack contributes Spectrum diffusion & CDNA3 attention upstream to SGLangRead PR log
Nullstackby NullStack
Back to all posts
AMDResearchLiteAttention

A Fast Attention Kernel for MI300X, Written in HIP, Not Assembly

July 24, 2026โ€ขBy NullStack Research

How we achieved near-peak memory bandwidth on AMD CDNA3 with clean C++ HIP code, beating hand-tuned assembly through careful memory hierarchy scheduling.

MM

NullStack Research

@nullstack_in

๐•

Low-level algorithms and hardware-aware systems power private, high-throughput AI inference for coding agents. Stay tuned as we release further benchmarks.