L40S vs P40
The cheapest P40 ($0.034) costs 91% less per hour than the cheapest L40S ($0.38).
Per unit of performance, the L40S is 19% cheaper: $0.40 versus $0.49 per A100-equivalent hour.
The L40S has 48 GB of memory against 24 GB on the P40.
| Measure | NVIDIA L40S 48GB | NVIDIA Tesla P40 24GB |
|---|---|---|
| Lowest per GPU-hour | $0.38 | $0.034 |
| Median per GPU-hour | $1.30 | $0.21 |
| On-demand from | $0.38 | $0.034 |
| Spot from | $0.45 | — |
| Reserved from | $0.49 | $0.038 |
| Lowest per A100-equivalent hour | $0.40 | $0.49 |
| Live offers | 284 | 32 |
| Providers | 85 | 14 |
| Memory | 48 GB GDDR6 | 24 GB GDDR5 |
| Memory bandwidth | 864 GB/s | 346 GB/s |
| FP16 / BF16 | 362 TFLOPS | — |
| FP8 | 733 TFLOPS | — |
| Performance vs A100 80GB | 0.96× | 0.07× |
| Architecture | Ada Lovelace | Pascal |
| Power | 350 W | 250 W |