Intel Launches 4th Gen Xeon Scalable Processors for Data Centers, CloudIntel Launches 4th Gen Xeon Scalable Processors for Data Centers, Cloud
Intel hopes to stem share loss after several delays in launching the new CPUs.
January 13, 2023
Following numerous delays, Intel has finally launched its 4th Gen Xeon Scalable processor (code-named Sapphire Rapids). At this week’s launch event, Intel described the new Xeon as one of its most significant product launches to date.
The launch included Intel’s new Xeon CPU Max Series (code-named Sapphire Rapids HBM), an x86-based processor with high bandwidth memory designed to accelerate HPC workloads without requiring changes. Intel also introduced its Data Center GPU Max Series (code-named Ponte Vecchio), which the company developed to provide enhanced security, AI, cloud, network and edge capabilities.
Partners that plan to embed Intel’s new processors include AWS, Cisco, Cloudera, CoreWeave, Dell Technologies, Dropbox, Ericsson, Fujitsu, Google Cloud, Hewlett Packard Enterprise, IBM Cloud, Inspur Information, IONOS, Lenovo, Los Alamos National Laboratory, Microsoft Azure, Nvidia, Oracle Cloud, OVHcloud, phoenixNAP, RedHat, SAP, SuperMicro, Telefonica and VMware.
Besides improving data center and cloud infrastructure performance, the new CPUs are important to Intel and its partners for other reasons. The delays, initially from manufacturing issues and ultimately to a flaw discovered last year, helped AMD expand its data center presence.
Intel Xeon processors, which once accounted for 99% of all data center infrastructure, could fall below 84% this year, DIgiTimes forecasts. Growth in demand for AMD’s EPYC processors is chipping away at Intel’s dominance with Xeon. According to DigiTimes, AMD EPYC processors will make up 16.3% of servers this year and even more growth in 2024.
At this week’s launch event, Intel executive VP and general manager Sandra Rivera emphasized the CPU improvements. Rivera said the new CPUs demonstrated a 2.9x performance per watt improvement over its predecessor when using all accelerators.
Customers can expect to see TCO improvements ranging from 52% to 66% and a 10x boost in the performance of AI workloads, she said. And with networking applications, customers can encrypt data using Intel’s Quick Assist tool to achieve up to 47% fewer cores at the same performance level.
Intel’s Sandra Rivera
“This represents the future of purpose-built workload acceleration,” Rivera said. “These step function performance gains are only possible because of the differentiated features we built into fourth Gen Z Xeon.”
The Xeon CPU Max Series
The Xeon CPU Max Series provides 64 GB of high bandwidth memory (HBM2e), which Intel said provides a substantial boost in data throughput for high-performance computing (HPC) and AI workloads. Intel claims that apps running on the Xeon CPU Max Series can deliver up to 3.7 times higher performance than Intel’s 3rd Gen Xeon Scalable processors.
Intel embedded 100 billion transistors into a 47-tile package on the Xeon CPU Max. The company designed it to run workloads such as physics, financial services and life sciences.
Major cloud providers that have already deployed Intel’s new 4th Gen Xeon Scalable processor include AWS, Microsoft, Google Cloud, IBM and Oracle. AWS has already deployed the latest processors in its EC2 service, said Amazon VP David Brown.
“Amazon EC2 R7iz instances already available today use the fourth generation Intel Xeon scalable processors and have the highest performance per vCPU among our x86-based EC2 instances,” Brown said. “They also deliver up to 20% higher performance than comparable high-frequency instances and are ideal for electronic design automation, relational databases, data analytics and other workloads requiring a combination of high compute performance and high memory.”
Intel officials also touted the security enhancements introduced to the 4th Gen Xeon Scalable processor. It supports Intel’s Software Guard Extensions (SGX), which Intel introduced with its 3rd Gen Xeon Scalable processor. Intel SGX provides confidential computing by reducing the attack surface in private, public and cloud-to-edge scenarios.
The new 4th Gen Xeon adds virtual-machine isolation technology called Trust Domain Extensions (TDX), which Intel developed to facilitate the migration of applications into confidential computing environments.
“We look forward to being one of the first cloud providers to offer confidential services based on Intel 4th generation Xeon scalable processors with Intel Trust Domain Extensions later this year, enabling organizations to achieve confidentiality by seamlessly lifting and shifting their workloads without requiring any code changes,” said Mark Russinovich, Microsoft Azure CTO.
Besides Microsoft, confidential computing services with the new Xeon CPUs will debut in Alibaba Cloud, Google Cloud and IBM Cloud.
Read more about:VARs/SIs
About the Author(s)
You May Also Like