RTX 4090 vs V100 32GB
The cheapest V100 32GB ($0.017) costs 90% less per hour than the cheapest RTX 4090 ($0.16).
Per unit of performance, the V100 32GB is 84% cheaper: $0.04 versus $0.25 per A100-equivalent hour.
The V100 32GB has 32 GB of memory against 24 GB on the RTX 4090.
| Measure | NVIDIA GeForce RTX 4090 | NVIDIA Tesla V100 32GB |
|---|---|---|
| Lowest per GPU-hour | $0.16 | $0.017 |
| Median per GPU-hour | $0.58 | $2.32 |
| On-demand from | $0.19 | $0.017 |
| Spot from | $0.16 | $0.51 |
| Reserved from | $0.18 | $0.19 |
| Lowest per A100-equivalent hour | $0.25 | $0.04 |
| Live offers | 427 | 193 |
| Providers | 91 | 43 |
| Memory | 24 GB GDDR6X | 32 GB HBM2 |
| Memory bandwidth | 1,008 GB/s | 900 GB/s |
| FP16 / BF16 | 165.2 TFLOPS | 125 TFLOPS |
| FP8 | 330.3 TFLOPS | — |
| Performance vs A100 80GB | 0.63× | 0.42× |
| Architecture | Ada Lovelace | Volta |
| Power | 450 W | 300 W |