Amax Engineering: Scalable HPC and GPU Reference Design Solutions with 4th Gen Xeon Scalable Processors

Amax Engineering announces scalable HPC and GPU reference design solutions that support the 4th Gen Xeon Scalable processors with industry custom and proprietary cooling solutions.

The company’s reference designs take a streamlined approach toward customizing solution configurations based on performance requirements. The line of next-gen systems provides power users with flexibility when configuring GPU- and HPC-focused systems around 4th Gen Xeon Scalable processors.

Built-in accelerators improve performance across AI, data analytics, networking, storage, and HPC. AI training and inference performance is improved due to the new Intel Advanced Matrix Extensions (Intel AMX). Intel QuickAssist Technology (Intel QAT) and the Intel Dynamic Load Balancer (Intel DLB) also speed up data movement and compression for faster networking by offloading scheduling and queue management. These accelerators can result in more efficient utilization to improve power efficiency by optimizing CPU core resources.

Sample workload-optimized reference design solutions by the firm featuring 4th Gen Xeon Scalable processors include:

• AceleMax GPU reference design – 4th Gen Xeon CPU, 4U Air Cooled workstation with support for 10xdata center GPUs
• LiquidMax liquid cooled reference design – 4th Gen Xeon CPU, 2U 4Node Liquid Cooled server with 8x350W CPUs up to 9G High Bandwidth Memory (HBM)
• ServMax HPC reference design – 4th Gen Xeon CPU, 2U 4Node Air Cooled server with 4xhot-pluggable nodes supporting single CPU, OCP 3.0 slots and NVMe drives
• StorMax storage design – 2U high-performance SATA/SAS/NVMe storage platform, with drive options such as 28xE1.s, 24xU.2, 14xE3.s, 12xLFF HDD

Customers are guided through multiple form factors, cooling options, GPU requirements, and CPU family preferences to scale their performance demands. Cooling options such as traditional air, liquid cold plate, and immersion cooling systems are custom-engineered into each solution to optimize system performance and minimize energy consumption. From workstations to rackmount servers to rack-scalable deployments, the firm’s reference designs offer solutions that scale with growing needs.

“AI is the backbone of future High-Performance Computing workloads and Intel Xeon Scalable processors offer much-needed performance and power efficiency improvements through the use of innovative accelerators, all of which align with our philosophy of always being on the cutting edge,” said Bohr Tsao, VP, business development.

For further information on how the reference design solutions can improve performance and reduce energy consumption at scale, contact team of engineering experts.