A16 vs RTX 3090
The cheapest A16 ($0.058) costs 14% less per hour than the cheapest RTX 3090 ($0.067).
Per unit of performance, the RTX 3090 is 59% cheaper: $0.22 versus $0.53 per A100-equivalent hour.
The RTX 3090 has 24 GB of memory against 16 GB on the A16.
| Measure | A16 | NVIDIA GeForce RTX 3090 |
|---|---|---|
| Lowest per GPU-hour | $0.058 | $0.067 |
| Median per GPU-hour | $0.53 | $0.30 |
| On-demand from | $0.082 | $0.11 |
| Spot from | — | $0.067 |
| Reserved from | $0.058 | $0.12 |
| Lowest per A100-equivalent hour | $0.53 | $0.22 |
| Live offers | 26 | 307 |
| Providers | 3 | 63 |
| Memory | 16 GB GDDR6 | 24 GB GDDR6X |
| Memory bandwidth | 200 GB/s | 936 GB/s |
| FP16 / BF16 | 35.9 TFLOPS | 71.2 TFLOPS |
| FP8 | — | — |
| Performance vs A100 80GB | 0.11× | 0.3× |
| Architecture | Ampere | Ampere |
| Power | 62 W | 350 W |