NVIDIA A100 vs NVIDIA H100
Side-by-side comparison of specifications, performance, and efficiency.
NVIDIA A100
Architecture
Ampere
Year
2020
VRAM
40 GB
TDP
250 W
NVIDIA H100
Architecture
Hopper
Year
2022
VRAM
80 GB
TDP
350 W
Specifications
NVIDIA A100 NVIDIA H100
Architecture
Ampere Hopper
Launch Year
2020 2022
Form Factor
PCIe PCIe
VRAM
40 GB80 GB
Memory Bandwidth
1.6 TB/s2.0 TB/s
TDP
250 W350 W
Process Node
7nm 4nm
| Spec | NVIDIA A100 | NVIDIA H100 |
|---|---|---|
| Architecture | Ampere | Hopper |
| Launch Year | 2020 | 2022 |
| Form Factor | PCIe | PCIe |
| VRAM | 40 GB | 80 GB |
| Memory Bandwidth | 1.6 TB/s | 2.0 TB/s |
| TDP | 250 W | 350 W |
| Process Node | 7nm | 4nm |
Performance (TFLOPS)
NVIDIA A100 NVIDIA H100
FP64
9.7 TFlops26.0 TFlops
FP32
19.5 TFlops51.0 TFlops
FP16
312.0 TFlops756.5 TFlops
BF16
312.0 TFlops756.5 TFlops
FP8
—1513.0 TFlops
INT8
624.0 TFlops1513.0 TFlops
| Precision | NVIDIA A100 | NVIDIA H100 |
|---|---|---|
| FP64 | 9.7 TFlops | 26.0 TFlops |
| FP32 | 19.5 TFlops | 51.0 TFlops |
| FP16 | 312.0 TFlops | 756.5 TFlops |
| BF16 | 312.0 TFlops | 756.5 TFlops |
| FP8 | — | 1513.0 TFlops |
| INT8 | 624.0 TFlops | 1513.0 TFlops |
FLOPS by Precision
Efficiency — TDP vs FP32 Performance
Analysis
Most Efficient
NVIDIA
H100
0.15 TFLOPS/W
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.