Intel is taking on Nvidia and AMD with its Max Series for high performance computing and artificial intelligence. The company unveiled two products under the Max umbrella: the Intel Xeon Max CPU and the Intel Data Center Max Series GPU. The Max GPU is Intel’s highest density processor, packing over 100 billion transistors into a 47-tile package with up to 128GB of high-bandwidth memory. The oneAPI open software ecosystem provides a single programming environment for both new processors, with Intel’s 2023 oneAPI and AI tools enabling the Intel Max Series products’ advanced features.

“To ensure no high-performance computing (HPC) workload is left behind, we need a solution that maximizes bandwidth, maximizes compute, maximizes developer productivity and ultimately maximizes impact,” Intel Super Compute Group corporate VP Jeff McVeigh said in an announcement. “The Intel Max Series product family brings high bandwidth memory to the broader market, along with oneAPI, making it easy to share code between CPUs and GPUs and solve the world’s biggest challenges faster.”

With the Max Series, Intel offers scalable, balanced CPUs and GPUs, incorporating memory bandwidth breakthroughs and united by oneAPI, an open, standards-based, cross-architecture programming framework.

Developed under the name Ponte Vecchio, the Intel Max Series GPU delivers:

• Up to 128 Xe-HPC cores.
• 408MB of L2 cache — which Intel says is the highest in the industry — and 64MB of L1 cache to increase throughput and performance.
• The only HPC/AI GPU with native ray tracing acceleration, designed to speed scientific visualization and animation.

The CPU Max, which was code-named Sapphire Rapids HBM, was developed as part of Intel’s “no workflow left behind” edict, and is “the first and only x86-based processor with high bandwidth memory” Intel says. Other highlights:

• 68 percent less power usage than an AMD Milan-X cluster for the same HPCG performance.
• AMX extensions boost AI performance and deliver 8x peak throughput over AVX-512 for INT8 with INT32 accumulation operations.
• Runs in different HBM and DDR memory configurations.
• Up to 56 performance cores on four tiles connected using Intel’s embedded multi-die interconnect bridge (EMIB) — all in a 350-watt envelope.
• 64GB of high bandwidth in-package memory.

The Max Series products are set for release in January 2023, and “Intel is shipping blades with Max Series GPUs to Argonne National Laboratory to power the Aurora supercomputer and will deliver Xeon Max CPUs to Los Alamos National Laboratory, Kyoto University and other supercomputing sites,” VentureBeat reports.

Intel’s top-performing Data Center Max Series GPU “will come in the form of OAM modules aimed at blade machines for supercomputing,” writes Tom’s Hardware, adding that “the company is also prepping humble Data Center GPU Max add-on cards that can go into mainstream servers and workstations.”

Intel’s Data Center Max 1100 GPU “employs a cut-down single-stack version of Intel’s Ponte Vecchio processor that features 56 Xe-HPC compute cores featuring 448 512-bit vector engines, 448 4096-bit matrix engines, 56 ray tracing cores, and 48GB of HBM2E memory.”