GPU comparison
NVIDIA H100 vs NVIDIA RTX 3090.
Specs, performance and live cloud availability, side by side.
The short answer
Choose NVIDIA H100 if you need 80GB+ on a single GPU (56GB more than NVIDIA RTX 3090).
Choose NVIDIA H100 if you need FP16 tensor throughput (1293% faster than NVIDIA RTX 3090).
Compare another pair
vs
Hopper
NVIDIA H100
from$4.29/hr
Available now · 4 regions- Memory
- 80 GB HBM3
- FP16 tensor
- up to 989 TFLOPS
- Bandwidth
- up to 3.4 TB/s
Ampere
NVIDIA RTX 3090
Not on demand yet- Memory
- 24 GB GDDR6X
- FP16 tensor
- 71 TFLOPS
- Bandwidth
- 936 GB/s
At a glance
Spec by spec. The longer bar is the better number.
Full specs
The complete spec sheet.
| Spec | NVIDIA H100 | NVIDIA RTX 3090 |
|---|---|---|
| Architecture | Hopper | Ampere |
| Memory | 80 GB | 24 GB |
| Memory type | HBM3 | GDDR6X |
| Memory bandwidth | up to 3.4 TB/s | 936 GB/s |
| FP32 | up to 67 TFLOPS | 35.6 TFLOPS |
| FP16 tensor | up to 989 TFLOPS | 71 TFLOPS |
| INT8 | up to 2400 TOPS | 284 TOPS |
| TDP | up to 700 W | 350 W |
| GPUs per node | 8 | 2 |
| Hourly price (live) | from $4.29/hr | Not on demand yet |
Questions
H100 or RTX 3090. What buyers ask first.
Other comparisons