r8gn.24xlarge vs dl1.24xlarge
| r8gn.24xlarge | dl1.24xlarge |
| Memory optimized | Machine Learning ASIC |
| High-performance databases, data mining/analysis, in-memory databases, distributed large-scale caches, real-time processing apps for unstructured big data, Hadoop/Spark clusters, and other enterprise apps. r – RAM / memory optimized 8 – Generation g – AWS Graviton processor n – Network and EBS optimized 24xlarge – Size | Deep learning / Machine Learning (ML) inference and training dl – Deep Learning 1 – Generation 24xlarge – Size |
| 96 | 96 |
| 768 GiB | 768 GiB |
| arm64 | n/a |
| AWS Graviton4 Processor | Intel Xeon Platinum 8275L |
| 2.8 GHz | |
| 1 | 0 |
| no | no |
| no | no |
| no | no |
| r8gn.12xlarge, r8gn.16xlarge, r8gn.24xlarge, r8gn.2xlarge, r8gn.48xlarge, r8gn.4xlarge, r8gn.8xlarge, r8gn.large, r8gn.medium, r8gn.metal-24xl, r8gn.metal-48xl, r8gn.xlarge | dl1.24xlarge |
| r8a.24xlarge, r8g.24xlarge, r8gb.24xlarge, r8gd.24xlarge, r8i.24xlarge, r8ib.24xlarge, r8id.24xlarge, r8idb.24xlarge, r8idn.24xlarge, r8in.24xlarge | |
| no | no |
| no | no |
| no | no |
| EBS only | 4 x 1000 GB NVMe SSD |
| default | |
| supported | |
| required | |
| 60000 MBps | 19000 MBps |
| 60000 Mbps | |
| 7500 Mbps | |
| 240000 IOPS | |
| 240000 IOPS | |
| 60000 Mbps | |
| 7500 Mbps | |
| 300 Gigabit | 400 Gigabit |
| yes | no |
| 24 | |
| 1 | |
| 50 | |
| 50 | |
| yes | no |
| required | |
| yes | no |
| 1 | |
| yes | no |
| yes | no |
| 1.00x | 1.50x |
Regional Prices
Geography | Region | r8gn.24xlarge | dl1.24xlarge |
|---|---|---|---|
| US | US East (Virginia) (us-east-1) | 8.7446 | 13.1090 |
| US | US West (Oregon) (us-west-2) | 8.7446 | 13.1090 |
Related Comparisons
Other instances worth weighing against these.