Differences
This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision Next revision Both sides next revision | ||
gpu [2020/09/14 15:28] fucik [Rules] |
gpu [2021/12/15 17:57] kasner [Pytorch Environment] |
||
---|---|---|---|
Line 7: | Line 7: | ||
| machine | GPU type | GPU driver version | [[https:// | | machine | GPU type | GPU driver version | [[https:// | ||
- | | dll1 | Quadro RTX 5000 | 440.33 | 7.5 | | + | | dll1 | Quadro RTX 5000 | 455.23.05 | 7.5 | 8 | 16 | 366.2 | |
- | | dll3 | GeForce GTX 1080 Ti | 440.33 | 6.1 | 10 | 11.0 | 248.0 | | + | | dll3 | GeForce GTX 1080 Ti | 455.23.05 | 6.1 | 10 | 11 | 248.8 | |
- | | dll4 | GeForce GTX 1080 Ti | 440.33 | 6.1 | 10 | 11.0 | 248.0 | | + | | dll4 | GeForce GTX 1080 Ti | 455.23.05 | 6.1 | 10 | 11 | 248.8 | |
- | | dll5 | GeForce GTX 1080 Ti | 440.33 | 6.1 | 10 | 11.0 | 248.0 | | + | | dll5 | GeForce GTX 1080 Ti | 455.23.05 | 6.1 | 10 | 11 | 248.8 | |
- | | dll6 | | + | | dll6 | |
- | | dll7 | GeForce RTX 2080 Ti | 418.39 | 7.5 | | + | | dll7 | GeForce RTX 2080 Ti | 455.23.05 | 7.5 | 8 | 11 | 248.8 | |
- | | dll8 | Quadro RTX 5000 | 440.33 | 7.5 | | + | | dll8 | Quadro RTX 5000 | 455.23.05 | 7.5 | 8 | 16 | 366.2 | |
- | | kronos | + | | dll9 | |
+ | | dll10 | GeForce RTX 3090 | 455.23.05 | ||
GPU cluster '' | GPU cluster '' | ||
| machine | GPU type | GPU driver version | [[https:// | | machine | GPU type | GPU driver version | [[https:// | ||
- | | tdll1 | Quadro P5000 | 410.48 | 6.1 | 8 | | + | | tdll1 | Quadro P5000 | 455.23.05 | 6.1 | 8 | |
- | | tdll2 | Quadro P5000 | 410.48 | 6.1 | 8 | | + | | tdll2 | Quadro P5000 | 455.23.05 | 6.1 | 8 | |
- | | tdll3 | Quadro P5000 | 410.48 | 6.1 | 8 | | + | | tdll3 | Quadro P5000 | 455.23.05 | 6.1 | 8 | |
- | | tdll4 | Quadro P5000 | 410.48 | 6.1 | 8 | | + | | tdll4 | Quadro P5000 | 455.23.05 | 6.1 | 8 | |
- | | tdll5 | Quadro P5000 | 410.48 | 6.1 | 8 | | + | | tdll5 | Quadro P5000 | 455.23.05 | 6.1 | 8 | |
Desktop machines: | Desktop machines: | ||
Line 39: | Line 40: | ||
* All the rules from [[:Grid]] apply, even more strictly than for CPU because there are too many GPU users and not as many GPUs available. So as a reminder: always use GPUs via '' | * All the rules from [[:Grid]] apply, even more strictly than for CPU because there are too many GPU users and not as many GPUs available. So as a reminder: always use GPUs via '' | ||
* **Note that you need to use '' | * **Note that you need to use '' | ||
- | * Always specify the number of GPU cards (e.g. '' | + | * Always specify the number of GPU cards (e.g. '' |
* If you need more than one GPU card (on a single machine), always require at least as many CPU cores ('' | * If you need more than one GPU card (on a single machine), always require at least as many CPU cores ('' | ||
* For interactive jobs, you can use '' | * For interactive jobs, you can use '' | ||
Line 54: | Line 55: | ||
You need to set library path from your '' | You need to set library path from your '' | ||
- | CUDA_version=10.1 | + | CUDA_version=11.1 |
- | CUDNN_version=7.6 | + | CUDNN_version=8.0 |
CUDA_DIR_OPT=/ | CUDA_DIR_OPT=/ | ||
if [ -d " | if [ -d " | ||
Line 88: | Line 89: | ||
This environment have TensorFlow 1.8.0 and all necessary requirements for NeuralMonkey. | This environment have TensorFlow 1.8.0 and all necessary requirements for NeuralMonkey. | ||
- | ==== Pytorch | + | ==== PyTorch |
- | If you want to use pytorch, there is a ready-made environment in | + | Install PyTorch following the instructions on https://pytorch.org. |
- | / | + | At the time of writing, the recommended |
- | + | ||
- | It does rely on the CUDA and CuDNN setup above. | + | |
+ | pip3 install torch==1.10.0+cu111 torchvision==0.11.1+cu111 torchaudio==0.10.0+cu111 -f https:// | ||
==== Using cluster ==== | ==== Using cluster ==== | ||
Line 112: | Line 112: | ||
watch nvidia-smi | watch nvidia-smi | ||
# For monitoring GPU activity in a separate terminal (thanks to Jindrich Libovicky for this!) | # For monitoring GPU activity in a separate terminal (thanks to Jindrich Libovicky for this!) | ||
+ | # You can also use nvidia-smi -l TIME | ||
nvcc --version | nvcc --version | ||
# this should tell CUDA version | # this should tell CUDA version |