trn2.48xlarge vs c9gd.48xlarge
| trn2.48xlarge | c9gd.48xlarge |
| Machine Learning ASIC | Compute optimized |
| Machine Learning (ML) training trn – AWS Trainium 2 – Generation 48xlarge – Size | High-performance web servers, scientific modeling, batch processing, distributed analytics, HPC, machine/deep learning inference, ad serving, scalable multiplayer gaming, and video encoding. c – Compute optimized 9 – Generation g – AWS Graviton processor d – Instance store volume 48xlarge – Size |
| 192 | 192 |
| 2048 GiB | 384 GiB |
| x86_64 | arm64 |
| Intel Xeon Scalable (Sapphire Rapids) | AWS Graviton5 Processor |
| 3.5 GHz | 2.8 GHz |
| 2 | 1 |
| no | no |
| 16 | |
| no | no |
| 15363 GiB | |
| no | no |
| trn2.48xlarge | c9gd.12xlarge, c9gd.16xlarge, c9gd.24xlarge, c9gd.2xlarge, c9gd.48xlarge, c9gd.4xlarge, c9gd.8xlarge, c9gd.large, c9gd.medium, c9gd.metal-48xl, c9gd.xlarge |
| c8gd.48xlarge, c9gd.48xlarge | |
| c9g.48xlarge | |
| ml.trn2.48xlarge | |
| no | no |
| no | no |
| no | no |
| EBS only | 3 x 3800 NVMe SSD |
| default | default |
| supported | supported |
| required | required |
| 81920 MBps | 72000 MBps |
| 80000 Mbps | 72000 Mbps |
| 10000 Mbps | 9000 Mbps |
| 260000 IOPS | 288000 IOPS |
| 260000 IOPS | 288000 IOPS |
| 80000 Mbps | 72000 Mbps |
| 10000 Mbps | 9000 Mbps |
| 3200 Gigabit | 100 Gigabit |
| yes | yes |
| 32 | 24 |
| 16 | 1 |
| 50 | 64 |
| 50 | 64 |
| yes | yes |
| required | required |
| yes | yes |
| 16 | 1 |
| yes | yes |
| no | yes |
| Infinityx |
Regional Prices
Geography | Region | trn2.48xlarge | c9gd.48xlarge |
|---|---|---|---|
| Hidden | Hidden (hidden-1) | n/a | 11.8560 |
| US | US East (Ohio) (us-east-2) | n/a | 10.2528 |
| US | US East (Virginia) (us-east-1) | n/a | 10.2528 |
| US | US West (Oregon) (us-west-2) | n/a | 10.2528 |