Intel officially introduces its newest CPU and GPU in the Intel Xeon Max Series line. This new family of Intel products will be showcased at the Supercomputing '22 event in Dallas.
These two components were released to support high performance computing (HPC) workloads, for example for supercomputers, servers or data centers, as well as artificial intelligence (AI).
“The Intel Max Series family of products is bringing high-bandwidth memory to a wider market, together with oneAPI (will) make it easier to practice code sharing between CPUs and GPUs and solve the world's biggest challenges more quickly,” said Jeff McVeigh, Corporate Vice President and General Manager, Intel Super Compute Group, Monday (14/11).
For Intel® Xeon® CPU Max Series, which comes with the code name Sapphire Rapids HBM, comes with the first high-bandwidth x86-based processor configuration. This performance can accelerate many HPC workloads without having to change the program.
Meanwhile, the Intel® Data Center GPU Max Series, codenamed Ponte Vecchio, is Intel's highest-density processor, containing more than 100 billion transistors in a 47-tile package with high-bandwidth memory up to 128 gigabit (GB).
The oneAPI open software ecosystem provides a single programming environment for the two new processors. Where these new products will power the Aurora supercomputer at Argonne National Laboratory.
“To ensure no HPC workloads are left behind, we need solutions that maximize bandwidth, maximize compute, maximize developer productivity and ultimately maximize impact,” he explained.
This support allows these CPUs to process many HPC workloads without the need to modify code. In practice, the CPU is also supported by Intel's oneAPI and AI 2023 software so that it can integrate code between the CPU and GPU and process computational data faster.
Technically the Intel Xeon Max CPU offers up to 56 cores or cores built with four tiles supported by embedded multi-die interconnect bridge (EMIB) technology with a 350 watt envelope. This CPU is also equipped with 64 GB of memory along with a high bandwidth package and PCI Express 5.0 and CXL1.1 I/O.
“The Max Series products meet the needs of this community with scalable and balanced CPUs and GPUs that integrate breakthroughs in memory bandwidth and are brought together by oneAPI, an open standards-based, cross-architecture programming framework.”
According to Intel, the CPU in the future will support the supercomputer runway of Los Alamos National Laboratory, Kyoto University and other supercomputers. Intel claims the Xeon Max Series CPUs will be used by 12 vendors including HP, Dell, Lenovo and Supermicro in more than 30 system designs.
The GPU Max Series will support the US Department of Energy's Aurora supercomputer whose launch has been delayed for quite a while. Aurora itself will be the second supercomputer in the US that offers high speed.