たれぱんのびぼーろく

わたしの備忘録、生物学とプログラミングが多いかも

クラウドサービスGPUインスタンス一覧

AWS

GPU EC2 Fargate ECS1 Batch2 SageMaker
NVIDIA
A100 p4d/p4de 3 - p4d p4d p4d/p4de
A10G g5 4 - g5 g5 g5
V100 p3/p3dn 5 - p3/p3dn p3/p3dn p3/p3dn
T4G g5g 6 - - - -
T4 g4dn 7 - g4dn g4dn g4dn
M60 g3s/g3 8 - g3s/g3 g3s/g3 -
K80 p2 9 - p2 p2 p2
Radeon Pro
V520 g4ad 10 - - - -
Others
trainium trn1 11 - trn112 trn1

  1. official doc
  2. "Only instance types that support a NVIDIA GPU and use an x86_64 architecture are supported for GPU jobs in AWS Batch." official docs
  3. "P4d instances ... by the latest NVIDIA A100 Tensor Core GPUs" official web
  4. "G5 instances feature up to 8 NVIDIA A10G Tensor Core GPUs" official web
  5. "P3 instances ... with up to 8 NVIDIA® V100 Tensor Core GPUs" official web
  6. "G5g instances are powered by AWS Graviton2 processors and feature NVIDIA T4G Tensor Core GPUs" official web
  7. "G4dn instances, powered by NVIDIA T4 GPUs" official web
  8. "G3 instances provides access to NVIDIA Tesla M60 GPUs" official web
  9. "P2 instances provide up to 16 NVIDIA K80 GPUs" official web
  10. "G4ad instances, powered by AMD Radeon Pro V520 GPUs" official web
  11. "Trn1 instances, powered by AWS Trainium accelerators" official web
  12. "You can register Amazon EC2 Trn1 ... to your clusters" official docs