NVIDIA L40 对比 NVIDIA GeForce RTX 3070 — GPU 比较 (Apr 2026)
NVIDIA L40 (48GB GDDR6, 181 TFLOPS FP16, Ada Lovelace) vs NVIDIA GeForce RTX 3070 (8GB GDDR6, 20.3 TFLOPS FP16, Ampere)
|
NVIDIA L40
48GB GDDR6 · Ada Lovelace
|
NVIDIA GeForce RTX 3070
8GB GDDR6 · Ampere
|
||
|---|---|---|---|
| 规格 | |||
| 制造商 | NVIDIA | NVIDIA | |
| 架构 | Ada Lovelace | Ampere | |
| VRAM | 48 GB GDDR6 | 8 GB GDDR6 | |
| 带宽 | 864 GB/s | 448 GB/s | |
| FP16 (Tensor) | 181.0 TFLOPS | 20.3 TFLOPS | |
| FP32 | 90.5 TFLOPS | 10.2 TFLOPS | |
| TDP | 300W | 220W | |
| 发布年份 | 2023 | 2020 | |
| 细分市场 | Data center | Consumer | |
| 适用场景 | Inference video processing rendering | Budget inference gaming | |
| 云端价格 | |||
| 最便宜的按需 | — | — | |
| 最便宜的竞价 | — | — | |
| 供应商 | 0 | 0 | |
NVIDIA L40
NVIDIA GeForce RTX 3070