AI solutions.

Advancing AI from cloud to edge to endpoints.

Making the benefits of AI pervasive.

AI is defining the next era of computing, and this is just the beginning. We see the benefits of AI every day—enabling medical research, curbing credit card fraud, reducing congestion in cities, or simply just making life easier.

The full potential of AI will be realised when the technology is pervasive and spans from the cloud to the edge to endpoints. AMD is helping drive this with a focus on three key areas.

AI architectures.

AMD products are built on scalable, power-efficient, and adaptable architectures designed for workloads ranging from large-scale AI model training to real-time inferencing.

AMD CDNA™ architecture is built to accelerate compute-intensive AI and HPC workloads, offering an advanced platform for tightly connected GPU systems that can share data quickly and efficiently. 

AMD XDNA™ is a spatial dataflow NPU architecture consisting of a tiled array of powerful, custom-designed AI Engines that enable high compute density, ideal for DNN and signal processing workloads.

AMD "Zen” architecture underlies AMD Ryzen™ processors and AMD EPYC™ server processors, offering the ultimate performance, scalability, and efficiency.

The AMD RDNA™ architecture features AI accelerators delivering incredible performance, efficiency, and features to gamers across desktops, laptops, gaming consoles, mobile devices, and the cloud. 

Spotlight –

A lot of AI doesn’t need real-time results.

Modern CPUs can run small to mid-sized AI inference workloads with sub-second latency. As AI inference workloads grow or response times shrink, you may need to add discrete accelerator.

As AI workloads rise, GPUs become increasingly cost-effective.

GPUs alone can support mixed enterprise workloads and AI. As model size, complexity, and volumes increase, GPU clusters can deliver more performance per euro.

Different models have unique processing needs.

Machine learning, graph processing, and statistical methods run exceptionally well on CPUs. Small to mid-sized large language models (LLMs) perform well on the latest CPUs. Larger models can realize significant benefit from AI accelerators.

AMD EPYC^TM  CPUs excel with enterprise-class AI.

5th-generation AMD EPYC^TM CPUs deliver major performance improvements for AI workloads:

• Up to 3.8x the throughput for end-to-end AI compared to competitor CPUs¹

• Up to 90%  faster throughput on Llama 3.1 8B at BF16 compared to competitor CPUs²

• Up to 86%faster Facebook AI Similarity Search (FAISS) compared to previous-generation EPYC^TM CPU³

5th-generation AMD EPYC™ CPUs – The best CPU for enterprise AI.

Shop now

^{1. TPCxAI @SF30 Multi-Instance 32C Instance Size throughput results based on AMD internal testing as of 05/09/2024 running multiple VM instances. The aggregate end-to-end AI throughput test is derived from the TPCx-AI benchmark and as such is not comparable to published TPCx-AI results, as the end-to-end AI throughput test results do not comply with the TPCx-AI Specification. 2P AMD EPYC 9965 (384 Total Cores), 12 32C instances, NPS1, 1.5TB 24x64GB DDR5-6400 (at 6000 MT/s), 1DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS-00A07 NVMe®, Ubuntu® 22.04.4 LTS, 6.8.0-40-generic (tuned-adm profile throughput-performance, ulimit -l 198096812, ulimit -n 1024, ulimit -s 8192), BIOS RVOT1000C (SMT=off, Determinism=Power, Turbo Boost=Enabled) 2P AMD EPYC 9755 (256 Total Cores), 8 32C instances, NPS1, 1.5TB 24x64GB DDR5-6400 (at 6000 MT/s), 1DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS00A07 NVMe®, Ubuntu 22.04.4 LTS, 6.8.0-40-generic (tuned-adm profile throughput-performance, ulimit -l 198096812, ulimit -n 1024, ulimit -s 8192), BIOS RVOT0090F (SMT=off, Determinism=Power, Turbo Boost=Enabled) 2P AMD EPYC 9654 (192 Total cores) 6 32C instances, NPS1, 1.5TB 24x64GB DDR5-4800, 1DPC, 2 x 1.92 TB Samsung MZQL21T9HCJR-00A07 NVMe, Ubuntu 22.04.3 LTS, BIOS 1006C (SMT=off, Determinism=Power) Versus 2P Xeon Platinum 8592+ (128 Total Cores), 4 32C instances, AMX On, 1TB 16x64GB DDR5-5600, 1DPC, 1.0 Gbps NetXtreme BCM5719 Gigabit Ethernet PCIe, 3.84 TB KIOXIA KCMYXRUG3T84 NVMe, Ubuntu 22.04.4 LTS, 6.5.0-35 generic (tuned-adm profile throughput-performance, ulimit -l 132065548, ulimit -n 1024, ulimit -s 8192), BIOS ESE122V (SMT=off, Determinism=Power, Turbo Boost = Enabled) Results: CPU Median Relative Generational Turin 192C, 12 Inst 6067.531 3.775 2.278 Turin 128C, 8 Inst 4091.85 2.546 1.536 Genoa 96C, 6 Inst 2663.14 1.657 1 EMR 64C, 4 Inst 1607.417 1 NA. Results may vary due to factors including system configurations, software versions and BIOS settings. TPC, TPC Benchmark and TPC-C are trademarks of the Transaction Processing Performance Council. (9xx5-012)}

^{2. Llama3.1-8B throughput results based on AMD internal testing as of 05/09/2024. Llama3-8B configurations: IPEX.LLM 2.4.0, NPS=2, BF16, batch size 4, Use Case Input/Output token configurations: [Summary = 1024/128, Chatbot = 128/128, Translate = 1024/1024, Essay = 128/1024, Caption = 16/16]. 2P AMD EPYC 9965 (384 Total Cores), 6 64C instances 1.5TB 24x64GB DDR5-6400 (at 6000 MT/s), 1 DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS-00A07 NVMe®, Ubuntu® 22.04.3 LTS, 6.8.0-40-generic (tuned-adm profile throughput-performance, ulimit -l 198096812, ulimit -n 1024, ulimit -s 8192), BIOS RVOT1000C, (SMT=off, Determinism=Power, Turbo Boost=Enabled), NPS=22P AMD EPYC 9755 (256 Total Cores), 4 64C instances, 1.5TB 24x64GB DDR5-6400 (at 6000 MT/s), 1DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS-00A07 NVMe®, Ubuntu 22.04.3 LTS, 6.8.0-40-generic (tuned-adm profile throughputperformance, ulimit - l 198096812, ulimit -n 1024, ulimit -s 8192), BIOS RVOT1000C (SMT=off, Determinism=Power, Turbo Boost=Enabled), NPS=22P AMD EPYC 9654 (192 Total Cores) 4 48C instances, 1.5TB 24x64GB DDR5-4800, 1DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS-00A07 NVMe®, Ubuntu® 22.04.4 LTS, 5.15.85-051585-generic (tuned-adm profile throughput-performance, ulimit -l 1198117616, ulimit -n 500000, ulimit -s 8192), BIOS RVI1008C (SMT=off, Determinism=Power, Turbo Boost=Enabled), NPS=2Versus 2P Xeon Platinum 8592+ (128 Total Cores), 2 64C instances , AMX On, 1TB 16x64GB DDR5-5600, 1DPC, 1.0 Gbps NetXtreme BCM5719 Gigabit Ethernet PCIe, 3.84 TB KIOXIA KCMYXRUG3T84 NVMe®, Ubuntu 22.04.4 LTS 6.5.0-35-generic (tuned-adm profile throughput-performance, ulimit -l 132065548, ulimit -n 1024, ulimit -s 8192), BIOS ESE122V (SMT=off, Determinism=Power, Turbo Boost = Enabled). Results: CPU 2P EMR 64c 2P Turin 192c 2P Turin 128c 2P Genoa 96c Average Aggregate Median Total Throughput 99.474 193.267 182.595 138.978 Competitive 1 1.943 1.836 1.397 Generational NA 1.391 1.314 1. Results may vary due to factors including system configurations, software versions and BIOS settings. (9xx5-009)}

 ^{3. FAISS (Requests/Hour) throughput results based on AMD internal testing as of 05/09/2024. FAISS Configurations: sift1m Data Set, 16 Core Instances, FP32, MKL 2024.2.1 2P AMD EPYC 9965 (384 Total Cores), 24 16C instances, 1.5TB 24x64GB DDR5-6400 (at 6000 MT/s), 1DPC, 1.0 Gbps NetXtreme BCM5720 Gigabit Ethernet PCIe, 3.5 TB Samsung MZWLO3T8HCLS-00A07 NVMe®, Ubuntu® 22.04.4 LTS, 6.8.0-40-generic (tuned-adm profile throughput-performance, ulimit -l 198096812, ulimit -n 1024, ulimit -s 8192), BIOS RVOT1000C (SMT=off, Determinism=Power, Turbo Boost=Enabled), NPS=42P AMD EPYC 9654 (192 Total cores) 12 16C instances, 1.5TB 24x64GB DDR5-4800, 1DPC, 2 x 1.92 TB Samsung MZQL21T9HCJR-00A07 NVMe, Ubuntu 22.04.3 LTS, BIOS 1006C (SMT=off, Determinism=Power), NPS=4Versus 2P Xeon Platinum 8592+ (128 Total Cores), 8 16C instances, AMX On, 1TB 16x64GB DDR5-5600, 1DPC, 1.0 Gbps NetXtreme BCM5719 Gigabit Ethernet PCIe, 3.84 TB KIOXIA KCMYXRUG3T84 NVMe, , Ubuntu 22.04.4 LTS, 6.5.0-35 generic (tuned-adm profile throughput-performance, ulimit -l 132065548, ulimit -n 1024, ulimit -s 8192), BIOS ESE122V (SMT=off, Determinism=Power, Turbo Boost = Enabled) Results: CPU Median Relative Throughput Generational 2P Turin 192C 64.2 3.776 1.861 2P Genoa 96C 34.5 2.029 1 2P EMR 64C 17 1 NA. Results may vary due to factors including system configurations, software versions and BIOS settings. (9xx5-011)}