Search
Homepage Rankings and Research Companies Channelcast Marketing Matters CRNtv Events WOTC Jobs Dell Technologies World Digital Experience 2020 HPE Zone Masergy Zenith Partner Program Newsroom Intel Partner Connect Digital Newsroom Dell Technologies Newsroom Fortinet Secure Network Hub IBM Newsroom Juniper Newsroom The IoT Integrator Lenovo Channel-First NetApp Data Fabric Intel Tech Provider Zone

Intel Set To Debut New Xeon CPUs That Widen Speed Gap Over AMD EPYC

'[Intel] genuinely [has] had better per-core performance over the years, and I think [the new processors] will accelerate that lead,' one Intel partner says of the new Xeon Gold processors that will push base frequencies to new highs for the second-generation Xeon Scalable server lineup.

Intel will soon launch its fastest Cascade Lake server processors yet in the form of two new Xeon Scalable chips that will give the semiconductor giant an extra edge in clock speed over AMD's EPYC processors.

The Santa Clara, Calif.-based company plans to announce the new Xeon Gold 6256 and Xeon Gold 6250 processors alongside the new Cascade Lake Refresh processors later this month, according to sources familiar with Intel's plans who asked to not be identified because they were not authorized to discuss the matter publicly.

Intel declined to comment.

The two new Xeon processors are "optimized for per-core performance" and will deliver "leadership frequency," providing benefits for demanding workloads and software cost optimization, according to a document seen by CRN. The upcoming Cascade Lake Refresh line, meanwhile, is meant to push core counts and base frequencies across a broad spectrum of Xeon Bronze, Silver and Gold processors for the mainstream and value portions of the market, as previously reported by CRN.

[Related : Why VMware's Licensing Change May Be Good News For AMD EPYC ]

The move is part of Intel's continued effort to narrow the gap with AMD's EPYC server processors, which the rival chipmaker has previously claimed can provide a 1.8x to 4x performance-per-dollar advantage over Intel's current stable of second-generation Xeon Scalable processors, code-named Cascade Lake.

With AMD's second-generation EPYC processors, the rival chipmaker has claimed performance leadership in part because of its higher core counts over Intel Xeon, but Intel has historically had an advantage when it comes to clock speed, according to one Intel channel partner.

"[Intel] genuinely [has] had better per-core performance over the years, and I think [the new Xeon processors] will accelerate that lead," said the Intel partner.

The highest base frequency of an AMD EPYC Rome processor is 3.2 GHz while the highest turbo frequency in the product line is 3.4 GHz — which are both lower than the base frequencies of Intel's two new Xeon processors.

The Xeon Gold 6250 comes with a 3.9 GHz base frequency — the highest seen in a processor from Intel’s Cascade Lake and Skylake SP lines — while sporting eight cores, a 37.5 MB cache and a 185W thermal design power, according to the document viewed by CRN. The Xeon Gold 6256, on the other hand, comes with a 3.6 GHz base frequency, the highest seen in a Xeon Scalable processor with 12 cores. The Xeon Gold 6256 has a 33 MB cache and 205W TDP.

Previously, the Cascade Lake processors with the highest base frequency were the Xeon Gold 5222 and Xeon Platinum 8256, both of which have a base clock speed of 3.8 GHz across four cores. For processors with 12 cores, the Cascade Lake part with the highest base frequency was the Xeon Gold 6246, which has a 3.3 GHz base clock speed. The Xeon Gold 6244, on the other hand, reaches a 3.6 GHz base frequency but only across eight cores.

The new Xeon processors will be good for software workloads like Oracle database applications that are based on per-core licensing, said a second Intel channel partner, who is already talking to customers about the forthcoming processors.

"Oracle's going to hit you on the number of cores that you're utilizing to run the Oracle database, but they're not hitting you on speed," said the Intel partner.

The Xeon Gold 6250 has already garnered a lot of customer interest, according to the Intel partner, because of its high frequency-to-core ratio, which will make it well suited for Oracle, SQL and steaming analytics applications, among others.

"The speeds in which we have to process data and get an answer back — yes or no, right away — It's incredibly important," the partner said. "The bar has been raised."

The two new processors could also benefit virtualization workloads where virtual machines are running demanding applications, according to the partner. That's an area where AMD's higher-core EPYC processors are set to take a hit after VMware last week announced a change to its per-CPU software licensing model that will soon charge an additional license for processors running more than 32 cores.

"Even at some point VMware could be a workload [that benefits from these processors]," the Intel partner said, "because customers aren't going to put in 48-core or 64-core processors for the sheer fact they're not going to want to apply two licenses per socket."

The new Xeon Gold 6250 and Xeon Gold 6256 were originally scheduled to debut alongside the new Cascade Lake Refresh processors on Feb. 23, but Intel's decision Tuesday to pull out of Mobile World Congress in Barcelona that week may delay the announcement by a day, according to one source. The organizers of Mobile World Congress have since called off the show due to concerns over the spread of coronavirus.

While the processors are new to the Cascade Lake family as part of the second-generation Xeon Scalable lineup, they are not considered Cascade Lake Refresh products and lack the "R" suffix as a result, sources said. The Xeon Gold 6250 and Xeon Gold 6256 have already appeared on some OEM websites, including Penguin Computing, Gigabyte and Fujitsu.

Sources who spoke to CRN could not shed light on other details such as turbo frequency, but they expect the new Xeon chips to share features with existing Cascade Lake processors, such as Deep Learning Boost instruction set for boosting inference workloads and Intel Optane DC persistent memory. It's also unclear how the addition of the two new processors will impact Intel's Xeon supply.

The new processors' performance-to-cost ratio will make for an enticing refresh opportunity with customers using Xeon Platinum processors from Intel's first-generation Xeon Scalable family, also known as Skylake SP, which launched in 2017, said the Intel partner who has been discussing the new parts with customers. The partner said the cost savings could be meaningful since Xeon Gold is a tier below Platinum but did not to disclose pricing details to CRN.

"The return on investment goes up, because you're going to get the processing power of that Platinum processor at a Gold cost," the partner said.

While that may lower the overall cost of the system at first, it could open up opportunities to add new kinds of components, like Intel Optane DC persistent memory, according to the partner. But more importantly, the partner said, it will help build trust with customers.

"You're showcasing to them you're not there after their dollar, you're there after their best interest which leads to a trusted advisorship we push from a [partner] standpoint," the partner said.

Back to Top

Video

 

sponsored resources