End to End Machine Learning
AMD EPYC 9374F achieves 2.25X the AI throughput performance of Intel® Xeon® Platinum 8462Y+ on TPCx-AI
Enterprise Large Language Models Performance
AMD EPYC 9654 achieves up to 1.36X the AI throughput performance of Intel® Xeon® Platinum 8480+ on Llama2-7B-Chat-hf and Llama2-13B-Chat-hf large language models
AI/ML Inference
Delivering up to 3.5X the AI inference performance of the prior generation for Neural Magic Inference Engine7
Natural Language Processing Inference Performance
4th Gen AMD EPYC with ZenDNN 4.0 delivers 1.8X the AI inference performance of the prior generation for Natural Language Processing (BERT-Large)
Containerized Database Performance
AMD EPYC 9754 based server achieves up to 4.4X the total queries per hour of the Ampere® Altra® Max M128-30 server on TPROC-H
Database Performance
Get 2.5X the DSS database performance, and 2.7X the OLTP database performance of the competition with 4th gen AMD EPYC CPU based servers
Telco Cloud Energy Efficiency
AMD EPYC 8534P achieves 125% higher throughput per watt compared to Intel® Xeon® 8471N running Juniper® vSRX firewall
Server-side Java® Energy Efficiency
AMD EPYC 9754 based server delivers 2.5X the performance per watt of the Ampere® Altra® Max M128-30 server for server-side Java®
Energy Efficiency
Delivering leadership energy efficiency with up to 100% higher performance per watt than the competition1
Fast and Easy Virtual Machine Migration to AMD EPYC
Witness 380 VMware virtual machines migrated from 5 legacy servers to 1 AMD EPYC 9004 server in less than hour, while reducing energy cost by 63%, and footprint by 80%
Enterprise Virtualization
Delivering 2.8X the performance of the competition for VMmark®2
Virtual Desktop Infrastructure
Supporting 4,000 VDI instances using half the servers compared to a competitive solution for LoginVSI™3
Virtual Desktop Infrastructure with Nutanix
Supporting more than twice the VDI instances compared to 3rd Gen AMD EPYC with Nutanix for LoginVSI™4
Molecular Dynamics
AMD EPYC 9754 achieves 2.13X the performance of Intel® Xeon® 8480+ on Molecular Dynamics workloads
Technical Computing
AMD EPYC 9384X achieves 55% higher performance on Ansys Technical computing workloads compared to Intel® Xeon® Max 9462 on Cache Mode
Computational Fluid Dynamics Performance
AMD EPYC 9684X based server achieves up to 2.18X the per-node performance of the Intel® Xeon® 8480+ server on OpenFOAM
Ansys® Fluent®
Delivering up to 75% (avg) faster performance versus the competition for CFD application Ansys Fluent5
Weather Modeling
Offering up to 150% more weather simulations compared to the competitive solution for Weather Research & Forecasting (WRF)6
Option Pricing
Enabling up to 110% higher performance than the competition for the Black-Scholes Model8
Rendering
Delivering up to 90% faster performance than the prior generation for V-Ray9
4th Generation AMD EPYC™ Processors
Advancing data center performance and efficiency.
Stay Connected
Sign up to receive the latest data center news or get in touch with an AMD Sales representative.
Notas al pie
- SP5-011B: SPECpower_ssj®2008 comparison based on published 2U, 2P Windows® results as of 11/10/2022. Configurations: 2P AMD EPYC 9654 (27501 overall ssj_ops/W, 2U, http://www.spec.org/power_ssj2008/results/res2022q4/power_ssj2008-20221020-01194.html) vs. 2P Intel Xeon Platinum 8380 (13670 overall ssj_ops/W, 2U, http://www.spec.org/power_ssj2008/results/res2022q4/power_ssj2008-20220926-01184.html). SPEC® and SPECpower_ssj® are registered trademarks of the Standard Performance Evaluation Corporation. See www.spec.org for more information.
- SP5-049B: VMmark® 3.1.1 matched pair comparison based on published results as of 11/10/2022. Configurations: 2-node, 2P 96-core EPYC 9654 powered server running VMware ESXi 8 RTM (40.19 @ 44 tiles/836 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2022-10-18-HPE-ProLiant-DL385Gen11.pdf) versus 2-node, 2P 40-core Xeon Platinum 8380 running VMware ESXi v7 U2 (14.19 @ 14 tiles/266 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2021-04-20-Fujitsu-PRIMERGY-RX2540M6.pdf) for 2.8x the score and 3.1x the tile (VM) capacity. 2-node, 2P EPYC 7763-powered server (23.33 @ 24 tiles/456 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2022-02-08-Fujitsu-RX2450M1.pdf) shown at 1.6x the performance for reference. VMmark is a registered trademark of VMware in the US or other countries.
- SP5-054: Login VSI™ Pro v4.1.40.1 comparison based on AMD internal testing as of 10/19/2022 measuring the maximum “knowledge worker” desktop sessions (VSImax) within VSI Baseline +1000ms response time using VMware ESXi 8 GA and VMware Horizon 8 on a server using 2x AMD EPYC 9654 (933 average VSImax sessions) versus a server with 2x Intel Xeon Platinum 8380 (400 average VSImax sessions). Results may vary.
- SP5-093: Login VSI™ Pro v4.1.40.1 comparison based on AMD internal testing as of 11/11/2022 measuring the average “knowledge worker” desktop sessions (VSImax) within VSI Baseline +1000ms response time using a pre-release version of Nutanix AOS and Citrix 1912 on a server using 2x AMD EPYC 9654 (1019 average VSImax sessions) versus a server with 2x AMD EPYC 7763 (486 average VSImax sessions) OR a server using 2x AMD EPYC 9654 (1019 average VSImax sessions) versus two servers with 2x AMD EPYC 7763 (486 x 2 = 972 average VSImax sessions) . Results may due to factors such as operating system and configuration. All testing and comparisons were conducted exclusively by AMD using pre-production software
- SP5-035A: Fluent® Release 2022 R2 test cases benchmark comparison based on AMD measurements as of 10/19/2022. Configurations: 2x 32-core Intel Xeon Platinum 8362 vs. vs. 2x 32-core EPYC 9374F for ~1.75x the rating performance. Results may vary.
- SP5-032: WRF® CONUS 2.5KM workload benchmark comparison based on AMD measurements as of 10/4/2022. Configurations: 2x 40-core Intel Xeon Platinum 8380 vs. vs. 2x 96-core EPYC 9654 for ~2.5x the time-step function performance. Results may vary.
- SP5-022: Neural Magic measured results on AMD reference systems as of 9/29/2022. Configurations:2P EPYC 9654 “Titanite” vs. 2P EPYC 7763 “DaytonaX” running on Ubuntu 22.04 LTS, Python 3.9.13, pip==22.12/deepsparse==1.0.2.
BERT-Large Streaming Throughput items/sec (seq=384, batch 1, 48 streams, INT8 + sparse) using SQuAD v1.1 dataset;
ResNet50 Batched Throughput items/sec (batch 256, single-stream, INT8 sparse) using ImageNet dataset;
YOLOv5s Streaming Throughput ([image 3, 640, 640], batch 1, multi-stream, per-stream latency <=33ms) using COCO dataset. Testing not independently verified by AMD.
- SP5-031: Black-Scholes European Option Pricing benchmark comparison based on AMD measurements for 100, 200, 400, 800, and 1600M options as of 10/4/2022. Max score is based on 200M options. Configurations: 2x 40-core Intel Xeon Platinum 8380 vs. 2x 64-core EPYC 9554 all systems on Ubuntu 22.04 and compiled with ICC 2022.1.0. Results may vary.
- SP5-038A: V-Ray based on published scores from https://benchmark.chaos.com/v5/vray as of 11/10/2022. Comparison of 2P AMD EPYC 9654 (209,102 max/206419 median vsamples, https://benchmark.chaos.com/v5/vray/#####) is 3.32x the performance of published 2P Intel Xeon Platinum 8380 (62,619 max/median vsamples, https://benchmark.chaos.com/v5/vray/29746). 2P EPYC 7763 shown for reference (109,248/99,443 median vsamples, https://benchmark.chaos.com/v5/vray/12888). Chaos®, V-Ray® and Phoenix FD® are registered trademarks of Chaos Software EOOD in Bulgaria and/or other countries. Note: Include text in red only if using 7763 score.
Notas al pie
- SP5-011B: SPECpower_ssj®2008 comparison based on published 2U, 2P Windows® results as of 11/10/2022. Configurations: 2P AMD EPYC 9654 (27501 overall ssj_ops/W, 2U, http://www.spec.org/power_ssj2008/results/res2022q4/power_ssj2008-20221020-01194.html) vs. 2P Intel Xeon Platinum 8380 (13670 overall ssj_ops/W, 2U, http://www.spec.org/power_ssj2008/results/res2022q4/power_ssj2008-20220926-01184.html). SPEC® and SPECpower_ssj® are registered trademarks of the Standard Performance Evaluation Corporation. See www.spec.org for more information.
- SP5-049B: VMmark® 3.1.1 matched pair comparison based on published results as of 11/10/2022. Configurations: 2-node, 2P 96-core EPYC 9654 powered server running VMware ESXi 8 RTM (40.19 @ 44 tiles/836 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2022-10-18-HPE-ProLiant-DL385Gen11.pdf) versus 2-node, 2P 40-core Xeon Platinum 8380 running VMware ESXi v7 U2 (14.19 @ 14 tiles/266 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2021-04-20-Fujitsu-PRIMERGY-RX2540M6.pdf) for 2.8x the score and 3.1x the tile (VM) capacity. 2-node, 2P EPYC 7763-powered server (23.33 @ 24 tiles/456 VMs, https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vmmark/2022-02-08-Fujitsu-RX2450M1.pdf) shown at 1.6x the performance for reference. VMmark is a registered trademark of VMware in the US or other countries.
- SP5-054: Login VSI™ Pro v4.1.40.1 comparison based on AMD internal testing as of 10/19/2022 measuring the maximum “knowledge worker” desktop sessions (VSImax) within VSI Baseline +1000ms response time using VMware ESXi 8 GA and VMware Horizon 8 on a server using 2x AMD EPYC 9654 (933 average VSImax sessions) versus a server with 2x Intel Xeon Platinum 8380 (400 average VSImax sessions). Results may vary.
- SP5-093: Login VSI™ Pro v4.1.40.1 comparison based on AMD internal testing as of 11/11/2022 measuring the average “knowledge worker” desktop sessions (VSImax) within VSI Baseline +1000ms response time using a pre-release version of Nutanix AOS and Citrix 1912 on a server using 2x AMD EPYC 9654 (1019 average VSImax sessions) versus a server with 2x AMD EPYC 7763 (486 average VSImax sessions) OR a server using 2x AMD EPYC 9654 (1019 average VSImax sessions) versus two servers with 2x AMD EPYC 7763 (486 x 2 = 972 average VSImax sessions) . Results may due to factors such as operating system and configuration. All testing and comparisons were conducted exclusively by AMD using pre-production software
- SP5-035A: Fluent® Release 2022 R2 test cases benchmark comparison based on AMD measurements as of 10/19/2022. Configurations: 2x 32-core Intel Xeon Platinum 8362 vs. vs. 2x 32-core EPYC 9374F for ~1.75x the rating performance. Results may vary.
- SP5-032: WRF® CONUS 2.5KM workload benchmark comparison based on AMD measurements as of 10/4/2022. Configurations: 2x 40-core Intel Xeon Platinum 8380 vs. vs. 2x 96-core EPYC 9654 for ~2.5x the time-step function performance. Results may vary.
- SP5-022: Neural Magic measured results on AMD reference systems as of 9/29/2022. Configurations:2P EPYC 9654 “Titanite” vs. 2P EPYC 7763 “DaytonaX” running on Ubuntu 22.04 LTS, Python 3.9.13, pip==22.12/deepsparse==1.0.2.
BERT-Large Streaming Throughput items/sec (seq=384, batch 1, 48 streams, INT8 + sparse) using SQuAD v1.1 dataset;
ResNet50 Batched Throughput items/sec (batch 256, single-stream, INT8 sparse) using ImageNet dataset;
YOLOv5s Streaming Throughput ([image 3, 640, 640], batch 1, multi-stream, per-stream latency <=33ms) using COCO dataset. Testing not independently verified by AMD. - SP5-031: Black-Scholes European Option Pricing benchmark comparison based on AMD measurements for 100, 200, 400, 800, and 1600M options as of 10/4/2022. Max score is based on 200M options. Configurations: 2x 40-core Intel Xeon Platinum 8380 vs. 2x 64-core EPYC 9554 all systems on Ubuntu 22.04 and compiled with ICC 2022.1.0. Results may vary.
- SP5-038A: V-Ray based on published scores from https://benchmark.chaos.com/v5/vray as of 11/10/2022. Comparison of 2P AMD EPYC 9654 (209,102 max/206419 median vsamples, https://benchmark.chaos.com/v5/vray/#####) is 3.32x the performance of published 2P Intel Xeon Platinum 8380 (62,619 max/median vsamples, https://benchmark.chaos.com/v5/vray/29746). 2P EPYC 7763 shown for reference (109,248/99,443 median vsamples, https://benchmark.chaos.com/v5/vray/12888). Chaos®, V-Ray® and Phoenix FD® are registered trademarks of Chaos Software EOOD in Bulgaria and/or other countries. Note: Include text in red only if using 7763 score.