…
Cloud rental price, compute per dollar and specs, side by side.
Open in the interactive comparer →| Spec / price | A100 SXM 40GB | A100 SXM 80GB |
|---|---|---|
| Memory | 40 GB HBM2 | 80 GB HBM2ebetter |
| Memory bandwidth | 1,555 GB/s | 2,039 GB/sbetter |
| FP16 / BF16 (dense) | 0.312 PF | 0.312 PF |
| FP8 (dense) | — | — |
| FP4 (dense) | — | — |
| INT8 (dense) | 0.624 PF | 0.624 PF |
| TF32 (dense) | 0.156 PF | 0.156 PF |
| FP32 (dense) | 0.0195 PF | 0.0195 PF |
| FP64 (dense) | 0.0195 PF | 0.0195 PF |
| Power | 400 W | 400 W |
| Architecture | Ampere (2020) | Ampere (2020) |
| Lowest on-demand | $0.54better | $0.81 |
| Median on-demand | $1.99better | $2.35 |
| Lowest spot | $0.55better | $0.90 |
| FP16/BF16 $ per PFLOP·h | $1.71better | $2.60 |
| Providers | 8 | 17 |
The A100 SXM 40GB and A100 SXM 80GB have the same dense FP16/BF16 compute. The A100 SXM 80GB has 1.3× more memory bandwidth than the A100 SXM 40GB. Real-world speedups depend on the workload: training and prefill scale with compute, while LLM decoding scales mostly with memory bandwidth.
Right now the A100 SXM 40GB is cheaper per unit of compute: $1.71 vs $2.60 per PFLOP-hour at the lowest on-demand prices.
Running one GPU around the clock at the lowest on-demand price costs about $391 per month for the A100 SXM 40GB and $591 for the A100 SXM 80GB.
The A100 SXM 80GB has more memory: 80 GB vs 40 GB.