NVIDIA GB200 Superchip vs NVIDIA GeForce RTX 3070 Ti — Comparaison GPU (Jun 2026)
NVIDIA GB200 Superchip (384GB HBM3e, 4,500 TFLOPS FP16, Blackwell) vs NVIDIA GeForce RTX 3070 Ti (8GB GDDR6X, 21.7 TFLOPS FP16, Ampere)
|
NVIDIA GB200 Superchip
384GB HBM3e · Blackwell
|
NVIDIA GeForce RTX 3070 Ti
8GB GDDR6X · Ampere
|
||
|---|---|---|---|
| Spécifications | |||
| Fabricant | NVIDIA | NVIDIA | |
| Architecture | Blackwell | Ampère | |
| VRAM | 384 GB HBM3e | 8 GB GDDR6X | |
| Bande passante | 16,000 GB/s | 608 GB/s | |
| FP16 (Tensor) | 4500.0 TFLOPS | 21.7 TFLOPS | |
| FP32 | 150.0 TFLOPS | 10.8 TFLOPS | |
| TDP | 2700W | 290W | |
| Année de sortie | 2024 | 2021 | |
| Segment | Centre de données | GPUs grand public | |
| Idéal pour | Largest-scale AI training multi-trillion parameter models | Budget inference gaming | |
| Tarification Cloud | |||
| Le moins cher à la demande | — | — | |
| Le moins cher en Spot | — | — | |
| Fournisseurs | 0 | 0 | |
NVIDIA GB200 Superchip
NVIDIA GeForce RTX 3070 Ti