| 规格 | NVIDIA A30 | NVIDIA T4 |
|---|
| 架构 | NVIDIA Ampere | Turing (TU104-895) |
| 显存 | 24 GB HBM2 | 16GB GDDR6 with ECC (默认开启) |
| 显存带宽 | 933 GB/s | 320GB/s (早期一页版datasheet口径300GB/s) |
| FP32算力 | 10.3 teraFLOPS | — |
| FP64算力 | 5.2 teraFLOPS | — |
| TF32算力 | 82 teraFLOPS | 165 teraFLOPS | — |
| FP16算力 | 165 teraFLOPS | 330 teraFLOPS | — |
| INT8算力 | 330 TOPS | 661 TOPS | — |
| 功耗TDP | 165 W | — |
| 互联 | PCIe Gen4: 64GB/s / Third-gen NVLINK: 200GB/s | PCIe Gen3 x16 (支持x8运行, Interconnect BW 32GB/s) |
| 形态 | Dual-slot, full-height, full-length (FHFL) | 单槽半高半长 Low-Profile 6.6英寸(长169.53mm), 被动散热, 板重301g |
| MIG | 4 GPU instances @ 6GB each | — |
| 显存位宽 | — | 256bit |
| 算力 | — | FP32 8.1 TFLOPS / FP16·FP32混合精度 65 TFLOPS / INT8 130 TOPS / INT4 260 TOPS |
| CUDA核心 | — | 2,560 (40 SM) / Tensor Core 320 / RT Core 40 |
| 功耗 | — | 70W (Total board power, 默认) |
| 发布年 | — | 2018 (Product Brief v01 2018-10-24初版, Dec18版datasheet © 2018 NVIDIA) |
| 时钟 | — | Base 585MHz / Boost 1590MHz (Base为70W功耗墙下调校) |
| 媒体引擎 | — | 1080p30解码 H.264 32路 / HEVC 44路 / VP9 32路 |
| SR-IOV | — | Supported; 16 VF（支持） |
| 工艺 | — | 12nm FFN, 136亿晶体管, 545mm2 |
| Compute APIs | — | CUDA, NVIDIA TensorRT, ONNX |