Let the benchmarking begin. Last week, AMD posted a YouTube video in which one of its EPYC-based systems outperformed a ‘comparable’ Intel Broadwell-based system on the STREAM benchmark and on a test case running ANSYS’s CFD application, Fluent. The intent was to showcase the new AMD chip’s (EPYC) strength on memory-bound HPC applications.
In the video, presenter Joshua Mora, senior manager field applications engineering, AMD, touts the EYPC’s memory controller and the memory bandwidth delivered. AMD has high hopes for its new EPYC line both in head-to-head competition with Intel as well as potentially creating a single socket market (see HPCwire article, AMD Charges Back into the Datacenter and HPC Workflows with EPYC Processor). Intel, of course, has been busy with its own introductions (see HPCwire article, Intel Unveils Xeon Scalable Processors). AMD EPYC will ultimately have to compete with Intel’s Skylake and IBM’s Power9 chip
The tested Intel system featured Xeon E5-2699 v4 processors (22 cores) and the AMD system featured EPYC 7601 (32 cores). Both were dual socket systems. “It is two clusters, tightly coupled with high speed, low latency InfiniBand interconnect running Windows OS,” according to Mora.
The AMD system was roughly 2X better on the STREAM benchmark which is intended to measure sustainable memory bandwidth. The dual socket Intel system ran at roughly 116 GB/s while the AMD system ran at roughly 266 GB/s. AMD says the STREAM performance is a good proxy for a range of HPC applications. Intel would no doubt offer a different view of the system’s set up and comparability of results.
AMD was roughly 78 percent faster running the Fluent simulation, which was a 14 million cell simulation of various aerodynamic effects on a jet. Mora cited AMD’s greater number of cores as well as its memory bandwidth as factors.
It seems clear AMD is ramping up its effort to win chunks of the datacenter and HPC landscaping following its absence from those markets in recent years.
At the time of EPYC’s launch, Scott Aylor, AMD corporate VP and GM of enterprise solutions business, said “It’s not enough to come back with one product, you’ve got to come back with a product cadence that moves as the market moves. So not only are we coming back with EPYC, we’re also [discussing follow-on products] so when customers move with us today on EPYC they know they have a safe home and a migration path with Rome.” AMD has committed to socket compatibility between EPYC 7000 line and Rome, code name of the next scheduled generation AMD processor aimed at the datacenter.
Based on the Zen core, EPYC is a line of system on a chip (SoC) devices designed with enhanced memory bandwidth and fast interconnect in mind. AMD also introduced a one-socket device, optimized for many workloads, which AMD says will invigorate a viable one-socket server market. With EPYC, “we can build a no compromise one-socket offering that will allow us to cover up to 50 percent of the two-socket market that is today held by the [Intel Broadwell] E5-2650 and below,” said Aylor.
Intel is the giant here, and not standing still. It will be interesting to watch the competition. EPYC seems to be a serious contender, though with a lot of mindshare ground to make up.