Dell EMC is planning on tripling the amount of AMD-based servers in its portfolio, following the success of the chip manufacturer’s EPYC range.
AMD spent a long time in the wilderness, playing second fiddle to main rival Intel across both the desktop and server markets. Its Zen microarchitecture, however, has been met with widespread acclaim, with Zen-based chips offering a noticeably lower TCO than equivalent Intel parts. In our tests, EPYC-based servers from Dell EMC, Broadberry and HPE all showcased phenomenal per-core performance for an excellent price.
This has not gone unnoticed by Dell. The company currently offers three server platforms that use AMD chips but Dominique Vanhamme, the company’s EMEA vice president and general manager for storage and compute told IT Pro that the company is planning to triple the number of AMD-based platforms it offers by the end of the year.
“Out of, let’s say, 50 or so platforms that we have today,” he said, “three of them are AMD – we’ll probably triple that by the end of this year.”
He also confirmed that Dell EMC will be launching servers powered by AMD’s newest architecture – a 7nm architecture codenamed ‘Rome’ – in the second half of 2019.
While AMD will still be a minority among Dell’s server platforms, this planned expansion is in contrast to comments made by Dell EMC CTO John Roese last year, who told Cloud Pro that Intel was still “the big player” in the market and that the company had no plans to substantially increase its AMD offering, stating “don’t expect it to be a duopoly any time soon”.
A significant barrier to AMD’s growth in the server market, as Vanhamme pointed out, is that any workloads that currently run on Intel servers will need to be re-validated to run on AMD-powered hardware. Given Intel’s relative stranglehold on the market, this means that a full AMD migration is likely to be a major project for any sizeable company.
One of the primary driving factors behind this expansion of AMD platforms is a growing demand from customers, according to Vanhamme. The lower TCO offered by AMD’s EPYC chips is a large factor, he says; along with a cheaper list price, many EPYC chips use fewer cores and sockets to match the performance of equivalent Intel systems, which means that CIOs can save money on per-core and per-socket licensing costs. Lower power consumption is also attractive, he said.
One thing that surprised Vanhamme was the demand for EPYC servers from general-purpose customers. For example. high-performance computing was expected to be the biggest revenue driver, due to the per-core and per-socket performance advantages, but general demand has been surprisingly strong.
“So in the original plan, we were thinking that it will be a few first verticals that will pick up, like service providers,” he said. “We thought that maybe there are some hosters that may want to have that extra capacity when they provide IaaS services. We clearly see HPC, but we also see general customers for sure.”