Version:  AMD_ZEN_HPL-MxP_2024-10-08

Dependencies:

  • The System: Zen4/Zen5 system with AVX512BF16 support.
  • OpenMPI 4/5:  This release was built with OpenMPI 5.0.3 and should run without issue if OpenMPI 5 or 4 is in the environment.
  • The binary was built on Red Hat® Enterprise Linux® 8.9 and tested on Red Hat® Enterprise Linux® 9, Ubuntu Linux 22.04.

Recommended Settings:

  • Boost : ON
  • Transparent Hugepages : always
  • SMT : OFF
  • NPS : 4
  • Determinism : Power

How to Run:

  • Ensure the above dependencies have been satisfied, and mpirun from OpenMPI is on your $PATH
  • Ensure OpenMPI is installed and loaded in your environment.
  • Invoke AMD HPL-MxP, supplying arguments for the size of the matrix (N) and block size (B)
  • Example Run Command for Single Node:
    • For AMD 4th Generation EPYC™ CPU, Dual Socket with 96 Cores/socket

      mpirun -np 24 --map-by l3cache:PE=8 --bind-to core -x OMP_NUM_THREADS=8 -x OMP_PROC_BIND=spread -x OMP_PLACES=cores ./hplMxP.x -N 65535 -B 1024
    • For AMD 5th Generation EPYC™ CPU, Dual Socket with 128 Cores/socket

      mpirun -np 32 --map-by l3cache:PE=8 --bind-to core -x OMP_NUM_THREADS=8 -x OMP_PROC_BIND=spread -x OMP_PLACES=cores ./hplMxP.x -N 65535 -B 1024