Nothing Stacks Up
To EPYC™
Performance leadership that helps you thrive in the most demanding business environments of today.
AMD EPYC™ processors set a new standard for data centers – delivering continued performance leaps and opening a new world of possibilities for enterprises worldwide. Modernize efficiently and advance data center sustainability goals, while addressing today’s business vulnerabilities.
Unmatched efficiency
Revolutionary multi-die design delivers industry-leading innovation for your business
Architecture leadership
AMD Infinity Architecture offers breakthrough performance to deliver next-generation computing.
Advanced security
AMD Infinity Guard helps organizations decrease risks to their most important assets
The World’s Best Data Center CPU
Our latest AMD EPYC™ 8004 series processors bring the “Zen 4c” core into a purpose-built CPU to elevate your data center performance and efficiency – meeting the needs of space and power-constrained infrastructure.
300+ World Records and Still Counting
AMD EPYC™ processors power the highest-performing x86 servers for the modern data center with World Record performance across major industry benchmarks.
world records across
major benchmarks in
HPC workloads
world records
for HPC apps*
partnerships
with major OEMs,
ISVs and IHVs
*AMD EPYC Family of Processors as of May 2022. See amd.com/worldrecords for the full list.
World’s Fastest HPC and Al Data Center Accelerators²
Designed to power discoveries in exascale systems, AMD Instinct™ MI200 series accelerators deliver the processing power needed by even the most complex Generative AI models.
Powering Transformative Science at Exascale
AMD EPYC™ processors and AMD Instinct™ MI200 accelerators are enabling scientists to tackle our most pressing challenges from climate change to vaccine research. Discover how they power discoveries in exascale systems to bring about a whole new era of High Performance Computing.
Unlock The True Sustainability Potential of Your Data Center
With the AMD EPYC™ Server Virtualization and Greenhouse Gas Emissions TCO Estimation Tool, you can easily see the potential value AMD EPYC™ processors may deliver for your data center. You can also compare your current x86 based server solution to a solution powered by AMD EPYC™ processors!
Delivering Leadership Performance Across Key Computing Workloads
Enterprise
Create a more agile data center with AMD EPYC™ processors that support workloads with the right balance of resources and a flexible architecture that can adapt to specialized workload needs
HPC
Easily optimize IT infrastructure for the right balance of per-core and overall server performance with the balanced architecture of AMD EPYC™ processors, without compromising on key processor features
Verticals
High performance cores, large memory, and fast I/O to support workloads across a wide range of industries including automotive, financial services, healthcare, manufacturing, oil and gas, retail, and telco
Discover how you can elevate business productivity
with AMD EPYC™ processors
Footnote
-
- SP5-013A: 96-core EPYC 9654 CPU processors results as of 11/10/2022 using SPECrate®2017_int_base. The AMD EPYC scored 1790 SPECrate®2017_int_base which is higher than all other 2P scores published on the SPEC® website. 2P AMD EPYC 9654 (1790 SPECrate®2017_int_base, 192 total cores, www.spec.org/cpu2017/results/res2022q4/cpu2017-20221024-32607.html). SPEC®, SPECrate® and SPEC CPU® are registered trademarks of the Standard Performance Evaluation Corporation. See www.spec.org for more information
- World's Fastest HPC and Al Data Center Accelerators: World's fastest data center GPU is the AMD Instinct™ MI250X. Calculations conducted by AMD Performance Labs as of Sep 15, 2021, for the AMD Instinct™ MI250X (128GB HBM2e OAM module) accelerator at 1, 700 MHz peak boost engine clock resulted in 95.7 FLOPS peak theoretical double precision (FP64 Matrix), 47.9 FLOPS peak theoretical double precision (FP64), 95.7 FLOPS peak theoretical single precision matrix (FP32 Matrix), 47.9 FLOPS peak theoretical single precision (FP32), 383.0 FLOPS peak theoretical half precision (FP16), and 383.0 TFLOPS peak theoretical Bfloat16 format precision (BF16) floating-point performance.