|
At present, we’re saying the overall availability of Amazon Elastic Compute Cloud (Amazon EC2) P6-B200 cases powered by NVIDIA B200 to deal with buyer wants for prime efficiency and scalability in synthetic intelligence (AI), machine studying (ML), and excessive efficiency computing (HPC) purposes.
Amazon EC2 P6-B200 cases speed up a broad vary of GPU-enabled workloads however are particularly well-suited for large-scale distributed AI coaching and inferencing for basis fashions (FMs) with reinforcement studying (RL) and distillation, multimodal coaching and inference, and HPC purposes reminiscent of local weather modeling, drug discovery, seismic evaluation, and insurance coverage danger modeling.
When mixed with Elastic Material Adapter (EFAv4) networking, hyperscale clustering by EC2 UltraClusters, and superior virtualization and safety capabilities by AWS Nitro System, you’ll be able to practice and serve FMs with elevated pace, scale, and safety. These cases additionally ship as much as two occasions the efficiency for AI coaching (time to coach) and inference (tokens/sec) in comparison with EC2 P5en cases.
You’ll be able to speed up time-to-market for coaching FMs and ship quicker inference throughput, which lowers inference value and helps improve adoption of generative AI purposes in addition to elevated processing efficiency for HPC purposes.
EC2 P6-B200 cases specs
New EC2 P6-B200 cases present eight NVIDIA B200 GPUs with 1440 GB of excessive bandwidth GPU reminiscence, fifth Era Intel Xeon Scalable processors (Emerald Rapids), 2 TiB of system reminiscence, and 30 TB of native NVMe storage.
Listed below are the specs for EC2 P6-B200 cases:
| Occasion measurement | GPUs (NVIDIA B200) | GPU reminiscence (GB) |
vCPUs | GPU Peer to see (GB/s) | Occasion storage (TB) | Community bandwidth (Gbps) | EBS bandwidth (Gbps) |
| P6-b200.48xlarge | 8 | 1440 HBM3e | 192 | 1800 | 8 x 3.84 NVMe SSD | 8 x 400 | 100 |
These cases function as much as 125 p.c enchancment in GPU TFLOPs, 27 p.c improve in GPU reminiscence measurement, and 60 p.c improve in GPU reminiscence bandwidth in comparison with P5en cases.
P6-B200 cases in motion
You should use P6-B200 cases within the US West (Oregon) AWS Area by EC2 Capability Blocks for ML. To order your EC2 Capability Blocks, select Capability Reservations on the Amazon EC2 console.
Choose Buy Capability Blocks for ML after which select your complete capability and specify how lengthy you want the EC2 Capability Block for p6-b200.48xlarge cases. The entire variety of days which you could reserve EC2 Capability Blocks is 1-14 days, 21 days, 28 days, or multiples of seven as much as 182 days. You’ll be able to select your earliest begin date for as much as 8 weeks upfront.
Now, your EC2 Capability Block will likely be scheduled efficiently. The entire worth of an EC2 Capability Block is charged up entrance, and the worth doesn’t change after buy. The cost will likely be billed to your account inside 12 hours after you buy the EC2 Capability Blocks. To be taught extra, go to Capability Blocks for ML within the Amazon EC2 Person Information.
When launching P6-B200 cases, you should utilize AWS Deep Studying AMIs (DLAMI) to assist EC2 P6-B200 cases. DLAMI supplies ML practitioners and researchers with the infrastructure and instruments to shortly construct scalable, safe, distributed ML purposes in preconfigured environments.
To run cases, you should utilize AWS Administration Console, AWS Command Line Interface (AWS CLI) or AWS SDKs.
You’ll be able to combine EC2 P6-B200 cases seamlessly with numerous AWS managed providers reminiscent of Amazon Elastic Kubernetes Companies (Amazon EKS), Amazon Easy Storage Service (Amazon S3), and Amazon FSx for Lustre. Help for Amazon SageMaker HyperPod can be coming quickly.
Now out there
Amazon EC2 P6-B200 cases can be found at present within the US West (Oregon) Area and may be bought as EC2 Capability blocks for ML.
Give Amazon EC2 P6-B200 cases a attempt within the Amazon EC2 console. To be taught extra, confer with the Amazon EC2 P6 occasion web page and ship suggestions to AWS re:Publish for EC2 or by your typical AWS Help contacts.
— Channy
How is the Information Weblog doing? Take this 1 minute survey!
(This survey is hosted by an exterior firm. AWS handles your data as described within the AWS Privateness Discover. AWS will personal the info gathered through this survey and won’t share the knowledge collected with survey respondents.)
