NVIDIA H20 vs NVIDIA V100
Side-by-side comparison of specifications, performance, and efficiency.
NVIDIA H20
Architecture
Hopper
Year
2024
VRAM
96 GB
TDP
400 W
NVIDIA V100
Architecture
Volta
Year
2017
VRAM
16 GB
TDP
250 W
Specifications
NVIDIA H20 NVIDIA V100
Architecture
Hopper Volta
Launch Year
2024 2017
Form Factor
SXM PCIe
VRAM
96 GB16 GB
Memory Bandwidth
4.0 TB/s900 GB/s
TDP
400 W250 W
Process Node
4nm 12nm
| Spec | NVIDIA H20 | NVIDIA V100 |
|---|---|---|
| Architecture | Hopper | Volta |
| Launch Year | 2024 | 2017 |
| Form Factor | SXM | PCIe |
| VRAM | 96 GB | 16 GB |
| Memory Bandwidth | 4.0 TB/s | 900 GB/s |
| TDP | 400 W | 250 W |
| Process Node | 4nm | 12nm |
Performance (TFLOPS)
NVIDIA H20 NVIDIA V100
FP64
1.0 TFlops7.0 TFlops
FP32
44.0 TFlops14.0 TFlops
FP16
148.0 TFlops112.0 TFlops
BF16
148.0 TFlops—
INT8
296.0 TFlops—
| Precision | NVIDIA H20 | NVIDIA V100 |
|---|---|---|
| FP64 | 1.0 TFlops | 7.0 TFlops |
| FP32 | 44.0 TFlops | 14.0 TFlops |
| FP16 | 148.0 TFlops | 112.0 TFlops |
| BF16 | 148.0 TFlops | — |
| INT8 | 296.0 TFlops | — |
FLOPS by Precision
Efficiency — TDP vs FP32 Performance
Analysis
Most Efficient
NVIDIA
H20
0.11 TFLOPS/W
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.