HPL Linpack with Intel MKL -- Single Node, Two Processors

Mare__Johnny — Wed, 29 Aug 2018 19:17:44 GMT

Hi everyone,

I've compiled HPL with Intel's MPI and Intel MKL but I'm confused about the best way to run it on a single node with two processors (12 cores each). Should I be using mpirun even though I only have a single node? If I try to set my PxQ grid to be 24 then I get an error saying that there are not enough processes. This is alleviated when I use mpirun. But I'm worried about communication overhead. Is it really necessary to use MPI when I've got a single node but with two processors? One of the other things I was confused about was what I should set the mkl option as when linking the MKL Library (cluster vs parallel). I've chosen parallel for the time being.

Any thoughts?

Thanks for Reading

Hi,

Kazushige_G_Intel — Thu, 30 Aug 2018 20:22:33 GMT

Hi,

This HPL requires 1 MPI process per NUMA node to utilize full bandwidth for main memory and performance will be higher with large problem sizes. Please set PxQ=2x1 and launch two MPI processes with script (runme_intel64_{static, dynamic}). If problem size is small, communication cost will be major bottleneck and performance will be lower. In this case, you could try 1 MPI process instead of two MPI processes.

topic Hi, in Intel® oneAPI Math Kernel Library

HPL Linpack with Intel MKL -- Single Node, Two Processors

Hi,