AMD increased the pressure it’s been applying to archrival Intel with the unveiling of an AMD EPYC 7003 Series that includes an AMD EPYC 7763 processor. The company claims it is the fastest server processor available.
Each member of the AMD EPYC 7003 Series has 64 “Zen 3” cores per processor with up to 32MB of per-core cache memory. The processors also support the Peripheral Component Interconnect (PCI) Express 4.0 expansion bus standard, which promises to double the overall throughput available over the existing PCI Express 3.0 bus.
AMD claims its third-generation of AMD EPYC 7003 Series processors offer the highest core density and twice the integer performance compared to competitors, while also improving transactional database processing by up to 19% and Big Data analytic sorts by up to 60% to provide 61% better price/performance than its primary x86 rival.
The AMD EPYC 7003 Series processors also expands AMD Infinity Guard security capabilities to now include a Secure Encrypted Virtualization-Secure Nested Paging (SEV-SNP) to thwart malicious hypervisor-based attacks by creating an isolated execution environment in memory.
Amazon Web Services (AWS), Microsoft, Google and Oracle all pledged to provide cloud services based on the latest AMD EPYC processor while Dell Technologies, Hewlett-Packard Enterprise (HPE), Cisco, and Lenovo also launched on-premises platforms. Lenovo, for example, added 10 Lenovo ThinkSystem Servers and ThinkAgile Hyperconverged Infrastructure (HCI) platforms built on the latest EPYC processors that it claims achieve 25 new world records across a broad set of industry-standard benchmarks.
That capability not only enables Lenovo to improve overall price/performance the security capabilities also enable the company to help lower the total cost of ownership, noted Kamran Amini. “It’s easier from a lifecycle management perspective,” says Amini.
Also read: Tetrate and the Emergence of the Intercloud Hypervisor
AMD’s Enterprise Profile on the Rise
AMD overall is currently in a much better financial position than in recent memory. For fiscal 2020 it recently reported revenue of $9.76 billion with an operating income of $1.37 billion. It’s hard to say how much of that revenue is being driven by sales of enterprise servers and client systems, but the days when AMD was primarily used as only a stalking horse to encourage Intel to cut prices may be over.
Less apparent, of course, is to what degree AMD processors will be consumed in the cloud versus on-premises. The bulk of data may stay on-premises but it’s clear the rate at which new workloads are being deployed in the cloud is more rapid. It’s also worth noting the rivals such as Nvidia have made significant gains in terms of processing massive amounts of data. For example, Informatica recently announced it is making Nvidia processors available in the cloud via a serverless computing framework running on a public cloud.
Regardless of the path forward, it’s clear AMD is now a force to be reckoned with in the enterprise. The challenge now will be figuring out what the right mix of AMD-based servers running on-premises and in the cloud should be alongside existing x86 servers from Intel, GPUs from Nvidia, and eventually Arm-based platforms running at the edge and in the cloud.
Read next: NVIDIA, VMware Create the AI-Ready Enterprise Platform at Cloud Scale
