Best Practice Guide - AMD EPYC

The EPYC processors are the latest generation of processors from AMD Inc. While they not yet show large adaptation on the top-500 list their performance might change this in the future. The processors are based on x86-64 architecture and provide vector units for a range of different data types, the most relevant being 64-bits floating point. Vector units are 256 bits wide and can operate on four double precision (64-bits) numbers at a time. The processors feature a high number of memory controllers, 8 in the EPYC 7601 model (see [6] for details) that was used for evaluation in the writing of this guide. They also provide 128 PCIe version 3.0 lanes. This guide provides information about how to use the AMD EPYC processors in an HPC environment and it describes some experiences with the use of some common tools for this processor, in addition to a general overview of the architecture and memory system. Being a NUMA type architecture information about the nature of the NUMA is relevant. In addition some tuning and optimization techniques as well as debugging are covered also. In this mini guide we cover the following tools: Compilers, performance libraries, threading libraries (OpenMP), Message passing libraries (MPI), memory access and allocation libraries, debuggers, performance profilers, etc. Some benchmarks, in which we compare compilers and libraries have been performed and some recommendations and hints about how to use the Intel tools with this processor are presented. While the AMD EPYC is a x86-64 architecture it's not fully compatible with Intel processors when it comes to the new features found on the latest generations of the Intel processors. Issues might be present when using highly optimized versions of the Intel libraries. In contrast to the Intel tools the GNU tools and tools from other independent vendors have full support for EPYC. A set of compilers and development tools have been tested with satisfactory results