NVIDIA GB200 Superchip vs NVIDIA RTX A4000 — Comparaison GPU (Apr 2026)
NVIDIA GB200 Superchip (384GB HBM3e, 4,500 TFLOPS FP16, Blackwell) vs NVIDIA RTX A4000 (16GB GDDR6, 19.2 TFLOPS FP16, Ampere)
|
NVIDIA GB200 Superchip
384GB HBM3e · Blackwell
|
NVIDIA RTX A4000
16GB GDDR6 · Ampere
|
||
|---|---|---|---|
| Spécifications | |||
| Fabricant | NVIDIA | NVIDIA | |
| Architecture | Blackwell | Ampere | |
| VRAM | 384 GB HBM3e | 16 GB GDDR6 | |
| Bande passante | 16,000 GB/s | 448 GB/s | |
| FP16 (Tensor) | 4500.0 TFLOPS | 19.2 TFLOPS | |
| FP32 | 150.0 TFLOPS | 16.0 TFLOPS | |
| TDP | 2700W | 140W | |
| Année de sortie | 2024 | 2021 | |
| Segment | Data center | Professional | |
| Idéal pour | Largest-scale AI training multi-trillion parameter models | Entry workstation rendering | |
| Tarification Cloud | |||
| Le moins cher à la demande | — | — | |
| Le moins cher en Spot | — | — | |
| Fournisseurs | 0 | 0 | |
NVIDIA GB200 Superchip
NVIDIA RTX A4000