Intel says its new Xeon 6 E-core CPUs feature up to 288 cores and will enable data centers to significantly reduce energy consumption and space. CRN goes through four important points about the new server CPUs, including how they’re different from past generations.
Intel used Computex 2024 to mark the launch of its Xeon 6 E-core processors and said the lineup has “tremendous density advantages” that will enable data centers to significantly consolidate racks running on older CPUs.
The processors, code-named Sierra Forest, are the first under an updated brand for Xeon, which is ditching the generation nomenclature and moving the generation number to the end of the name.
[Related: Intel Reveals 8-Chip Gaudi 3 Platform Price, Upending Industry Norm Of Secrecy]
They are also part of a new bifurcated server processor strategy, where Intel will release two types of Xeon processors from now on: those with performance cores (P-cores) to optimize for speed and those with efficient cores (E-cores) to optimize for efficiency.
While the company did not disclose OEMs or cloud service providers supporting Xeon 6 processors, some vendors revealed plans this week to ship new servers with the chips. These vendors included Supermicro, Lenovo, Tyan, QCT and Gigabyte.
Launched on Monday, the Xeon 6 processors represent the next step in Intel CEO Pat Gelsinger’s plan to restore the company to its former glory and compete with a growing crop of chip designers, chip manufacturers and major customers such as Amazon Web Services, Microsoft Azure and Google, which are building their own chip solutions.
The chips are arriving only several months after Intel debuted the fifth-generation Xeon processors in December, reflecting Gelsinger’s push to accelerate the company’s road map after it fell behind in cutting-edge manufacturing for several years.
In an interview with CRN, Ronak Singhal, chief architect of Xeon at Intel, said the company decided to pursue this strategy because workloads within data centers are becoming more diverse than ever before, resulting in increasingly different performance and power requirements that would be difficult to satisfy with a single architecture.
“You see things that are more heavyweight, things that are more lightweight, people looking for different levels of performance. And so trying to cover that entire span with a single solution becomes harder and harder,” said Singhal, who is also a senior fellow, the highest technical title that can be achieved at the company.
But while Xeon 6 lineup consists of processors with two different core types, Singhal emphasized that they are compatible with the same server socket and come with the same level of software support, apart from two AI features found only in P-core CPUs.
“It’s one platform, socket-compatible between the P-core version and the E-core version. So I invest in a single infrastructure, both at the hardware level and the software level, and then I can decide which CPU to populate that with,” he said.
What follows are four important points about Intel’s new Xeon 6 P-core processors. These points address software support, platform enhancements, specs, how the chips compare to previous generations and AMD’s EPYC CPUs, how they can reduce energy consumption, and Intel’s broader rollout plans for Xeon 6 processors.
Xeon 6: Software Support, Platform Enhancements
While the new lineup includes two different types of core architectures, they are compatible with the same server socket, rely on the same x86 instruction set architecture and support the same operating systems and hypervisors as well as the wide range of applications and libraries traditionally backed by Xeon CPUs.
The only difference in software support is that Xeon 6 E-core processors don’t come with support for Advanced Vector Extensions 512, also known as AVX-512, or Advanced Matrix Extensions, otherwise referred to as AMX. These features come with Xeon 6 P-core processors to support AI and vector operations.
Like fifth-generation Xeon processors that debuted in December, all Xeon 6 processors support DDR5 memory, but they provide a boost in memory bandwidth, with E-core processors improving bandwidth by 40 percent to 6400 megatransfers per second and P-core processors improving bandwidth by 2.3 times to 8000 megatransfers per second, the latter of which is made possible with memory using the new MCR DIMM standard.
The Xeon 6 processors support PCIe Gen 5, also found in fifth-gen Xeon, while bumping the number of lanes by up to 20 percent to a maximum of 96. They also expand upon fifth-gen Xeon’s support of Compute Express Link (CXL) 2.0, with all CPUs supporting Type C for memory bandwidth and capacity expansion, among use cases, in addition to Type A and Type A. that were supported by the previous generation.
Xeon 6: Intel’s Rollout Plan, E-Core And P-Core Specs
Intel is phasing the rollout of Xeon 6 processors over the next several months, and they are divided into different categories based on performance and other features. These categories consist of the Xeon 6900, 6700, 6500 and 6300 series as well as Xeon 6 system-on-chip series. Most categories include processors with P-cores and E-cores.
The Xeon 6700E series, with the “E” suffix indicating its use of E-cores, is launching this week for servers with one or two sockets. They feature up to 144 E-cores, a maximum thermal design power of 350 watts, eight channels of DDR5 memory with up to 6400 megatransfers per second, up to 88 PCIe 5.0 or CXL 2.0 lanes and four UPI 2.0 links with up to 24 gigatransfers per second.
The Xeon 6900P series, with the “P” suffix indicating its use of P-cores, will launch in the third quarter for servers with one or two sockets. They feature up to 128 P-cores, a maximum thermal design power of 500 watts, 12 channels of DDR5 memory with up to 8000 megatransfers per second, up to 96 PCIe 5.0 or CXL 2.0 lanes and six UPI 2.0 links with up to 24 gigatransfers per second.
Intel will then release a few more categories of Xeon 6 processors in the first quarter of next year. These include the Xeon 6900E series, which features up to 288 E-cores, a maximum thermal design power of 500 watts, 12 channels of DDR5 memory with up to 6400 megatransfers per second, to 96 PCIe 5.0 or CXL 2.0 lanes and six UPI 2.0 links with up to 24 gigatransfers per second.
The first quarter is also when Intel plans to launch the Xeon 6700P series, which supports servers with one, two, four or eight sockets. The series features up to 86 P-cores, a maximum thermal design power of 350 watts, eight channels of DDR5 memory with up to 8000 megatransfers per second, up to 88 PCIe 5.0 or CXL 2.0 lanes and four UPI 2.0 links with up to 24 gigatransfers per second.
How Xeon 6 E-Core Compares To Previous Generations, AMD EPYC
Intel said the Xeon 6 E-core processors provide major efficiency gains over previous generations, especially for CPUs that are more than a couple years old.
Compared to second-gen Xeon processors, for instance, the performance-per-watt of Xeon 6 E-core CPUs is 2.6 times higher for web and microservices workloads like Java, 3.4 times higher for networking workloads, 2.6 times higher for media transcoding and 2.7 times higher for data services workloads like MySQL OLTP, according to Intel.
The company said the Xeon 6 E-core processors also provide significant performance-per-watt gains over fifth-gen Xeon CPUs while delivering similar or, in some cases, better performance depending on the workload.
With web services workloads like server-side Java for instance, the 144-core Xeon 6780E has 20 percent higher performance-per-watt while reaching 94 percent of the performance level of a 64-core Xeon Platinum 8592+, according to the company. In multiple media workloads, however, Intel showed the Xeon 6-Ecore CPU providing an improvement in performance-per-watt and performance by double-digit percentages.
In terms of how Xeon 6 E-core processors compare to rival chips, Intel said the 128-core Xeon 6756E provides similar thread performance to the 64-core EPYC 9534 with improved performance-per-watt ranging from 3 percent to 34 percent.
Intel: Xeon 6 E-Core Can Save Data Centers Lots Of Energy
With Xeon 6 E-core processors, Intel said it can significantly reduce the number of servers with older processors needed to provide the same level of performance. This, in turn, will reduce the space and power needed for such server clusters.
For example, the chipmaker said the Xeon 6700E series can enable a data center to replace 200 server racks of second-gen Xeon processors with 66 racks of Xeon 6700E chips, resulting in 80,000 megawatt hours saved and 34,000 metric tons of CO2 emissions reduced over four years.
Early customers said they are finding significant reductions in energy consumption while maintaining a similar or higher level of performance in initial tests with Xeon 6 E-core processors, according to testimonials provided by Intel.
Swedish telecom giant Ericsson, for example, said test results showed that Xeon 6 E-core chips improved performance by 3.2 times and reduced energy by more than 40 percent for cloud infrastructure and 5G core applications compared to second-gen Xeon.
SAP, on the other hand, said initial testing showed that Xeon 6 E-core chips enabled it to reduce power consumption by up to 60 percent while achieving a similar level of performance for SAP Hana and SAP Hana Cloud workloads compared to fifth-gen Xeon.
