HPE servers and AMD EPYC™ 9004X CPUs accelerate technical compute workloads

[ComputeExperts]

HPE and AMD have teamed up to tackle some of today’s most complex technical problems and workloads with the world’s highest performance x86 server CPU for technical computing.[i]

Are you looking for a new way to increase the performance of technical software solutions to support more complex simulations and rapid designs? Searching for a high-performance technical computing solution that can accelerate performance and flexibility, while boosting productivity and enhancing security?

Technical professionals working on computational fluid dynamics (CFD), electronic design automation (EDA), finite element analysis (FEA) – plus open-source high-performance computing apps that range from weather and ocean flow simulation to computational chemistry and other life sciences – can see and appreciate the benefits of the EPYC 9004X processors’ enhanced L3 cache size.

“Out-of-the-box” performance

Built on the ground-breaking AMD 3D Chiplet architecture, and leveraging advanced 5nm process technology, AMD EPYC™ 9004X processors were made to deliver high technical computing performance. These processors are designed with up to 96 “Zen 4” Cores and up to 1.152GB of L3 cache for breakthrough performance on targeted applications. For many challenging workloads, having access to this large L3 cache can improve performance by continuously feeding the processor with data – without having to access the main memory outside of the CPU.

Hewlett Packard Enterprise and AMD provide the answer to your questions with HPE ProLiant Gen11 and HPE Cray servers powered by new AMD EPYC™ 9004 Series processors with 3D V-Cache™ technology. These recently introduced powerhouse systems accelerate critical product design and technical computing workloads.

These next-generation CPUs raise the bar for breakthrough performance on targeted critical product design and technical computing workloads, optimizing performance and accelerating the development of new products and technologies. The out-of-the-box performance uplift is delivered with no application software changes. You simply upgrade your existing HPE ProLiant Gen11 or HPE Cray servers with AMD EPYC™ 9004X processors[ii] to receive the performance and productivity benefits you need to drive better business outcomes.

Examples of AMD EPYC performance benefits include:

• Computation Fluid Dynamics, with up to ~ 2.5x the performance vs. the competition[iii]
• Explicit Finite Element Analysis, delivering up to ~ 2.1x the performance vs. the competition[iv]
• Superlinear Scaling, enabling up to ~14 nodes of CFD performance on an 8-node cluster[v]

AMD EPYC™ 9004X processors utilize the same memory architecture as the rest of the 4th Generation EPYC family but have up to three times as much L3 cache. This means you can take advantage of more than 1GB of L3 cache per core without sacrificing performance, productivity, or choice. With HPE ProLiant Gen11 or HPE Cray servers powered by AMD EPYC 9004 Series processors with 3D V-Cache, you receive up to 1.15GB L3 cache per socket to take on your most demanding technical computing tasks.

Choose from a range of solutions ideal for technical computing

HPE offers server models to match the full range of your technical computing needs, each with a choice of three AMD EPYC 9004 Series CPUs with 3D V-Cache—from 16 to 32 to 96 cores—to optimize the core count for specific workload demands. The full portfolio of HPE ProLiant Gen11 and HPE Cray systems spans single – and dual – processor solutions, supporting memory bandwidth up to 6TB, high-speed PCIe Gen5 I/O, and a choice of AMD EPYC 9x84X core parts that support up to a 1.152GB L3 cache.

• AMD EPYC 9184X CPU 16 core. For workloads that benefit from more cache per core and higher frequency.
• AMD EPYC 9384X CPU 32 core. The sweet spot for workloads that need the perfect balance of frequency, cache per core, and density.
• AMD EPYC 9684X CPU 96 core. Maximum core density opens more memory bandwidth for the most demanding technical computing workloads.

Benefits-at-a-glance

Leadership density and throughput. Up to 96 cores per socket supports large-scale virtualization, high-performance computing (HPC), and dense computing.

Leadership core performance. AMD “Zen 4” architecture is critical for single-threaded applications and for maximizing per-core licensing.

Leadership I/O bandwidth. 128 PCIe4 lanes enable highly efficient and powerful I/O configurations.

Leadership x86 L3 cache; Up to 96MB / Core. Up to 1.15GB total L3 cache can enable super-linear[vi] EDA/CFD/FEA applications, helping to improve performance by reducing cache misses.

Leadership AMD Infinity Guard Features[vii].  Security features supported by mainstream Linux^® distros, VMware^®, GCP, and Azure.

With this range of computing options, HPE and AMD enable you to fine-tune solutions for just the right combination of raw compute power and platform density to maintain reliable performance as your technical computing workloads scale.

To learn more, please check out:

Accelerate performance-demanding technical workloads

hpe.com/partners/AMD

AMD launch video

Meet HPE Blogger Christina Austin Tiner! Christina is a veteran marketing leader in the Information Technology industry, having contributed to the server business for over 20 years. She has served in a variety of roles that include worldwide product management for Compaq and HP, director of product management for the Apollo line of HPC platforms, director of product management for Dell PowerEdge C servers, and group manager for the worldwide product marketing team responsible for the ProLiant enterprise rack server portfolio.

Compute Experts
Hewlett Packard Enterprise

twitter.com/hpe_compute
linkedin.com/showcase/hpe-servers-and-systems/
hpe.com/servers

[i] 96-core EPYC 9684X

[ii] Updating to the latest BIOS is always recommended when updating processors.

[iii] AMD internal testing: ANSYS^® CFX^® (External Air Flow over an Airfoil 10M/50M/100M) 2P 96C EPYC 9684X vs. 2P 56C Xeon 8480+. Results may vary due to factors including system configurations, software versions and BIOS settings. See ANSYS^® CFX^® and AMD 3D V-CACHE™ Technology Performance Brief.

[iv] AMD internal testing: ANSYS^® LS-DYNA^® (4 Test Cases) 2P 96C EPYC 9684X vs. 2P 56C Xeon 8480+. Results may vary due to factors including system configurations, software versions and BIOS settings. See ANSYS^® LS-DYNA^® and AMD 3D V-CACHE™ Technology Performance Brief.

[v] AMD internal testing: OpenFOAM^® 2P 96C EPYC 9684X. Results may vary due to factors including system configurations, software versions and BIOS settings. See OPENFOAM^® and AMD 3D V-CACHE™ Technolgoy Performance Brief.

[vi] AMD defines “linear scaling” as an equal and proportionate application performance uplift relative to single node performance; that is, when scaling out to 2 nodes results in 2x the performance of a single node, scaling out to 4 nodes results in 4x the performance of a single node, and so forth. “Super-linear” scaling is when the performance uplift achieved by adding one or more node(s) is greater than linear.  AMD allows a +/- of 2% margin of error when claiming linear or super linear scaling. GD-205

[vii] AMD Infinity Guard features vary by EPYC™ Processor generations. Infinity Guard security features must be enabled by server OEMs and/or Cloud Service Providers to operate. Check with your OEM or provider to confirm support of these features. Learn more about Infinity Guard at https://www.amd.com/en/technologies/infinity-guard .

[AmitSharmaAPJ]

Nice article Christina.