CUDA/HIP header-only library for low-precision (16 bit, 8 bit) and vectorized GPU kernel development

This is the open source version of HPL-MXP. The code performance has been verified on Frontier

mixed-precision GEMM library

Hybrid-Precision Analysis on CG Solver (H.A.C.S). Merging single and double precision to generate a fast yet accurate CG solver

Benchmarks for mixed-precision emulations

Simulate math functions in abitrary low precisions