AMD Instinct MI300X Accelerators Available on Oracle Cloud Infrastructure for Demanding AI Applications
AMD announced that Oracle Cloud Infrastructure (OCI) has chosen AMD Instinct™ MI300X accelerators with ROCm™ open software to power its newest OCI Compute Supercluster instance, BM.GPU.MI300X.8. The OCI Supercluster with AMD MI300X supports up to 16,384 GPUs in a single cluster, designed for demanding AI workloads including large language model (LLM) inference and training.
These OCI bare metal instances have been adopted by companies like Fireworks AI. The AMD Instinct MI300X underwent extensive testing, validated by OCI, showcasing its AI inferencing and training capabilities for latency-optimal use cases and the ability to fit the largest LLM models in a single node. Fireworks AI is leveraging the performance benefits of OCI using AMD Instinct MI300X for their fast platform designed to build and deploy generative AI.
AMD ha annunciato che Oracle Cloud Infrastructure (OCI) ha scelto gli acceleratori AMD Instinct™ MI300X con il software aperto ROCm™ per alimentare la sua più recente istanza del Supercluster OCI, BM.GPU.MI300X.8. Il Supercluster OCI con AMD MI300X supporta fino a 16.384 GPU in un unico cluster, progettato per carichi di lavoro AI esigenti, inclusa l'inferenza e l'addestramento di modelli di linguaggio di grandi dimensioni (LLM).
Queste istanze bare metal OCI sono state adottate da aziende come Fireworks AI. L'AMD Instinct MI300X ha subito test approfonditi, validati da OCI, dimostrando le sue capacità di inferenza e addestramento AI per casi d'uso ottimizzati per la latenza e la capacità di ospitare i più grandi modelli LLM in un singolo nodo. Fireworks AI sta sfruttando i vantaggi prestazionali di OCI utilizzando AMD Instinct MI300X per la loro piattaforma veloce progettata per costruire e distribuire AI generativa.
AMD anunció que Oracle Cloud Infrastructure (OCI) ha elegido los aceleradores AMD Instinct™ MI300X con el software abierto ROCm™ para alimentar su más reciente instancia de OCI Compute Supercluster, BM.GPU.MI300X.8. El Supercluster OCI con AMD MI300X admite hasta 16,384 GPU en un solo clúster, diseñado para cargas de trabajo de IA exigentes, incluida la inferencia y el entrenamiento de modelos de lenguaje de gran tamaño (LLM).
Estas instancias bare metal de OCI han sido adoptadas por empresas como Fireworks AI. El AMD Instinct MI300X pasó por pruebas exhaustivas, validadas por OCI, mostrando sus capacidades de inferencia y entrenamiento de IA para casos de uso óptimos en latencia y la capacidad de alojar los modelos LLM más grandes en un solo nodo. Fireworks AI está aprovechando los beneficios de rendimiento de OCI utilizando AMD Instinct MI300X para su plataforma rápida diseñada para construir y desplegar IA generativa.
AMD는 Oracle Cloud Infrastructure (OCI)가 AMD Instinct™ MI300X 가속기와 ROCm™ 오픈 소프트웨어를 선택하여 최신 OCI 컴퓨트 슈퍼클러스터 인스턴스인 BM.GPU.MI300X.8을 구축했다고 발표했습니다. AMD MI300X가 장착된 OCI 슈퍼클러스터는 단일 클러스터에서 최대 16,384개의 GPU를 지원하며, 대규모 언어 모델 (LLM)의 추론 및 훈련을 포함한 고성능 AI 작업을 위해 설계되었습니다.
이 OCI 베어메탈 인스턴스는 Fireworks AI와 같은 기업에 의해 채택되었습니다. AMD Instinct MI300X는 OCI에 의해 검증된 광범위한 테스트를 거쳤으며, 지연 최적화 사용 사례에 대한 AI 추론 및 훈련 기능과 함께 단일 노드에서 가장 큰 LLM 모델을 수용할 수 있는 능력을 보여주었습니다. Fireworks AI는 AMD Instinct MI300X를 사용하여 생성적 AI를 구축하고 배포하기 위해 설계된 빠른 플랫폼에서 OCI의 성능 이점을 활용하고 있습니다.
AMD a annoncé qu'Oracle Cloud Infrastructure (OCI) a choisi les accélérateurs AMD Instinct™ MI300X avec le logiciel ouvert ROCm™ pour alimenter sa plus récente instance OCI Compute Supercluster, BM.GPU.MI300X.8. Le Supercluster OCI avec AMD MI300X prend en charge jusqu'à 16,384 GPU dans un seul cluster, conçu pour des charges de travail IA exigeantes, y compris l'inférence et l'entraînement de modèles de langage de grande taille (LLM).
Ces instances bare metal OCI ont été adoptées par des entreprises comme Fireworks AI. L'AMD Instinct MI300X a subi des tests approfondis, validés par OCI, démontrant ses capacités d'inférence et d'entraînement IA pour des cas d'utilisation optimisés pour la latence et sa capacité à intégrer les plus grands modèles LLM dans un seul nœud. Fireworks AI exploite les avantages en termes de performances d'OCI en utilisant AMD Instinct MI300X pour sa plateforme rapide conçue pour créer et déployer de l'IA générative.
AMD kündigte an, dass Oracle Cloud Infrastructure (OCI) AMD Instinct™ MI300X-Beschleuniger mit der ROCm™-Open-Software ausgewählt hat, um die neueste OCI-Compute-Supercluster-Instanz BM.GPU.MI300X.8 zu betreiben. Der OCI-Supercluster mit AMD MI300X unterstützt bis zu 16.384 GPUs in einem einzigen Cluster, der für anspruchsvolle KI-Workloads, einschließlich der Inferenz und des Trainings großer Sprachmodelle (LLM), konzipiert ist.
Diese OCI-Bare-Metal-Instanzen wurden von Unternehmen wie Fireworks AI übernommen. Der AMD Instinct MI300X hat umfangreiche Tests bestanden, die von OCI validiert wurden, und zeigt seine Fähigkeiten zur KI-Inferenz und -Schulung für latenzoptimale Anwendungsfälle sowie die Fähigkeit, die größten LLM-Modelle in einem einzelnen Knoten unterzubringen. Fireworks AI nutzt die Leistungssteigerungen von OCI, indem es AMD Instinct MI300X für ihre schnelle Plattform, die zur Erstellung und Bereitstellung generativer KI entwickelt wurde, verwendet.
- OCI chose AMD Instinct MI300X accelerators for its newest Compute Supercluster instance
- OCI Supercluster supports up to 16,384 AMD Instinct MI300X GPUs in a single cluster
- AMD Instinct MI300X demonstrated strong performance in OCI's extensive testing for AI inferencing and training
- Fireworks AI adopted OCI bare metal instances powered by AMD Instinct MI300X
- None.
Insights
The introduction of AMD Instinct MI300X accelerators on Oracle Cloud Infrastructure (OCI) marks a significant advancement in AI computing capabilities. This partnership enables OCI to offer high-performance bare metal instances, particularly beneficial for large language model (LLM) inference and training.
Key points:
- OCI Supercluster can support up to 16,384 AMD Instinct MI300X GPUs in a single cluster
- The solution is designed for AI models with hundreds of billions of parameters
- Fireworks AI has already adopted these instances, indicating early market traction
- The MI300X accelerators offer high throughput with leading memory capacity and bandwidth
This development positions AMD competitively in the AI accelerator market, potentially challenging NVIDIA's dominance. For AMD investors, this could signal increased revenue streams from the data center and AI sectors, which are experiencing rapid growth.
The collaboration between AMD and Oracle Cloud Infrastructure represents a strategic move in the competitive AI hardware market. This partnership could significantly impact AMD's market position:
- Expands AMD's presence in the cloud AI infrastructure segment
- Potentially increases AMD's market share in the high-performance computing (HPC) and AI accelerator markets
- Enhances AMD's credibility in supporting large-scale AI workloads
The adoption by Fireworks AI suggests growing trust in AMD's AI solutions. If this trend continues, it could lead to increased demand for AMD's data center products, positively affecting revenue and potentially stock performance. However, investors should monitor how this translates into actual market share gains against established competitors like NVIDIA in the coming quarters.
— Customers including Fireworks AI are powering their AI inference and training workloads with new OCI Compute instances --
— OCI Supercluster leads among cloud providers with support for up to 16,384 AMD Instinct MI300X GPUs in a single ultrafast network fabric --
SANTA CLARA, Calif., Sept. 26, 2024 (GLOBE NEWSWIRE) -- AMD (NASDAQ: AMD) today announced that Oracle Cloud Infrastructure (OCI) has chosen AMD Instinct™ MI300X accelerators with ROCm™ open software to power its newest OCI Compute Supercluster instance called BM.GPU.MI300X.8. For AI models that can comprise hundreds of billions of parameters, the OCI Supercluster with AMD MI300X supports up to 16,384 GPUs in a single cluster by harnessing the same ultrafast network fabric technology used by other accelerators on OCI. Designed to run demanding AI workloads including large language model (LLM) inference and training that requires high throughput with leading memory capacity and bandwidth, these OCI bare metal instances have already been adopted by companies including Fireworks AI.
“AMD Instinct MI300X and ROCm open software continue to gain momentum as trusted solutions for powering the most critical OCI AI workloads,” said Andrew Dieckmann, corporate vice president and general manager, Data Center GPU Business, AMD. “As these solutions expand further into growing AI-intensive markets, the combination will benefit OCI customers with high performance, efficiency, and greater system design flexibility.”
“The inference capabilities of AMD Instinct MI300X accelerators add to OCI’s extensive selection of high-performance bare metal instances to remove the overhead of virtualized compute commonly used for AI infrastructure,” said Donald Lu, senior vice president, software development, Oracle Cloud Infrastructure. “We are excited to offer more choice for customers seeking to accelerate AI workloads at a competitive price point.”
Bringing Trusted Performance and Open Choice for AI Training and Inference
The AMD Instinct MI300X underwent extensive testing which was validated by OCI that underscored its AI inferencing and training capabilities for serving latency-optimal use cases, even with larger batch sizes, and the ability to fit the largest LLM models in a single node. These Instinct MI300X performance results have garnered the attention of AI model developers.
Fireworks AI offers a fast platform designed to build and deploy generative AI. With over 100+ models, Fireworks AI is leveraging the benefits of performance found in OCI using AMD Instinct MI300X.
“Fireworks AI helps enterprises build and deploy compound AI systems across a wide range of industries and use cases," said Lin Qiao, CEO of Fireworks AI. "The amount of memory capacity available on the AMD Instinct MI300X and ROCm open software allows us to scale services to our customers as models continue to grow."
Supporting Resources
About AMD
For more than 50 years AMD has driven innovation in high-performance computing, graphics, and visualization technologies. Billions of people, leading Fortune 500 businesses, and cutting-edge scientific research institutions around the world rely on AMD technology daily to improve how they live, work, and play. AMD employees are focused on building leadership high-performance and adaptive products that push the boundaries of what is possible. For more information about how AMD is enabling today and inspiring tomorrow, visit the AMD (NASDAQ: AMD) website, blog, LinkedIn, and Twitter pages.
AMD, the AMD Arrow logo, Instinct, ROCm, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other names are for informational purposes only and may be trademarks of their respective owners.
Trademarks
Oracle, Java, MySQL and NetSuite are registered trademarks of Oracle Corporation. NetSuite was the first cloud company—ushering in the new era of cloud computing.
Contact:
David Szabados
AMD Communications
+1 408-472-2439
david.szabados@amd.com
Mitch Haws
AMD Investor Relations
+1 512-944-0790
mitch.haws@amd.com
FAQ
What is the new OCI Compute Supercluster instance powered by AMD Instinct MI300X called?
How many AMD Instinct MI300X GPUs can the OCI Supercluster support in a single cluster?
What company is using OCI bare metal instances with AMD Instinct MI300X for generative AI?
What types of AI workloads is the AMD Instinct MI300X designed to handle on OCI?
