Reproduce “FP64 Performance and the Limits of Ozaki Emulation on HPC Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Katarzyna, Swirydowicz, Nicholas Malaya, Noah Wolfe, Paul Bauman, Pieter Ghysels, Jonathan Belof
The authors evaluate whether Ozaki-II can use low-precision matrix engines for FP64 scientific workloads. They profile 47 configurations across 11 applications and benchmarks. Ten entries lack a dominant DGEMM kernel; from this sample, the authors infer less than 10% of HPC is eligible, so Ozaki is an option for some dense GEMMs rather than a general FP64 replacement.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.