概要
NVIDIA の GPU には NVIDIA architectures という コードが割り当てられているが、よく忘れるのでまとめたもの。
出典: https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/
世代 | NVIDIA architecture name | ボード名 | 対応CUDA バージョン |
---|---|---|---|
Fermi | sm_20 | GeForce 400, 500, 600, GT630 | CUDA3.2 ~ CUDA 8 |
Kepler | sm_30 | GeForce 700, GT-730 | CUDA 5 ~ CUDA 10 |
sm_35 | GeForce K40 | ||
sm_37 | GeForce K80 | ||
Maxwell | sm_50 | Quadro RTX 4000 | CUDA 6 ~ CUDA 11 |
sm_52 | Quadro M6000, GeForce 900, GTX-970, 980, Titan X | ||
sm_53 | Tegra TX1, Tegra X1, Drive CX, Drive PX, Jetson Nano | ||
Pascal | sm_60 | Quadro GP 100, Tesla P100, DGX-I | CUDA 8 - |
sm_61 | GTX 1080, GTX 1070, GTX 1060, GTX 1050, GTX 1030 (GP108), GT 1010 (GP108) Titan Xp, Tesla P40, Tesla P4, Discrete GPU on the NVIDIA Drive PX2 | ||
sm_62 | Integrated GPU on the NVIDIA Drive PX2, Tegra (Jetson) TX2 | CUDA 8 - | |
Volta | sm_70 | DGX-1 with Volta, Tesla V100, GTX 1180 (GV104), Titan V, Quadro GV100 | CUDA 9 - |
sm_72 | Jetson AGX Xavier, Drive AGX Pegasus, Xavier NX | CUDA 9 - | |
Turing | sm_75 | GTX/RTX Turing – GTX 1660 Ti, RTX 2060, RTX 2070, RTX 2080, Titan RTX, Quadro RTX 4000, Quadro RTX 5000, Quadro RTX 6000, Quadro RTX 8000, Quadro T1000/T2000, Tesla T4 | CUDA 10 - |
Ampere | sm_80 | NVIDIA A100 (the name “Tesla” has been dropped – GA100), NVIDIA DGX-A100 | CUDA 11.1 - |
sm_86 | Tesla GA10x cards, RTX Ampere – RTX 3080, GA102 – RTX 3090, RTX A2000, A3000, A4000, A5000, A6000, NVIDIA A40, GA106 – RTX 3060, GA104 – RTX 3070, GA107 – RTX 3050, Quadro A10, Quadro A16, Quadro A40, A2 Tensor Core GPU | CUDA 11.1 - | |
Hopper | sm_90 | NVIDIA H100 (GH100) | CUDA 12 [計画] - |