Workshops Archive

FP64 Performance and the Limits of Ozaki Emulation on HPC Systems


Workshop: PMBS26: The 17th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computer Systems

Authors: Katarzyna (Kasia) Swirydowicz, Nicholas Malaya, Noah Wolfe, Paul Bauman, Pieter Ghysels, and Jonathan Belof (Advanced Micro Devices (AMD))

Abstract: A growing industry narrative holds that native double precision can be deprioritized in HPC accelerators, on the hypothesis that low-precision tensor cores can emulate FP64 on demand, and that most workloads are bandwidth-bound. This paper evaluates those claims and the importance of FP64 in scientific computing. We show roofline profiling data of representative HPC workloads across common computational motifs, to evaluate the arithmetic-intensity and accuracy. We also consider FP64 emulation via the Ozaki method to ascertain some of the possibilities and limitations.

These results indicate that emulation should not be evaluated as a replacement for FP64, but as an optimization for the subset of workloads whose dominant computation is sufficiently large, dense, and GEMM-like: this is less than 10\% of the HPC ecosystem. Even in this favorable regime, GEMM speedup is not sufficient: numerical behavior, workspace requirements, matrix shape, and integration overhead determine whether the benefit survives at application scale.


Back to PMBS26: The 17th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computer Systems Archive Listing

Back to Full Workshop Archive Listing