剛出爐的新機,wsl和windows的nvidia共享
![](https://img.laitimes.com/img/9ZDMuAjOiMmIsIjOiQnIsIyZuBnL5IjM0MGM2gDZhZTOxAzYhFGZ5QTNhJmZlhzNzUWNmV2Lc52YucWbp5GZzNmLn9Gbi1yZtl2Lc9CX6MHc0RHaiojIsJye.png)
win+shift+s截圖
wget https://gitee.com/mindspore/mindspore/raw/r1.8/scripts/install/ubuntu-gpu-pip.sh
PYTHON_VERSION=3.9 MINDSPORE_VERSION=1.8.1 bash -i ./ubuntu-gpu-pip.sh
這裡要選擇yes
難道是我移動網卡的問題?
把網卡拔了,用本地網。
ping通,那麼不是我的問題
echo 'deb https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu1804/x86_64' > /etc/apt/sources.list.d/cuda.list
E: Malformed entry 1 in list file /etc/apt/sources.list.d/cuda.list (Suite)
E: The list of sources could not be read.
說明該方法不行
gpg --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
成功!
如果沒有nvidia-smi,這個邏輯可以在bash裡面實作?
sudo apt install nvidia-utils-390
出了錯,重來又在下載下傳了,建議建立一個如果已有檔案不再下載下傳的邏輯。
提個pr
安裝後半截驗證
[ERROR] ME(31402,7f3c64a52740,python):2022-09-12-17:05:24.234.635 [mindspore/ccsrc/runtime/hardware/device_context_manager.cc:93] WaitTaskFinishOnDevice] SyncStream failed, exception:The stream_id: 0 is greater than stream array size: 0
出錯了,那我們
sudo apt install nvidia-cuda-toolkit
排錯
WARNING: The scripts f2py, f2py3 and f2py3.9 are installed in ‘/home/kewei/.local/bin’ which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
WARNING: The script cache_admin is installed in ‘/home/kewei/.local/bin’ which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
然後,我突然發現一鍵運作的腳本是不支援wsl的,這下搞崩了,隻有重新下載下傳wsl,卸了重來。
解除安裝
wsl --list
Ubuntu替換成要解除安裝的核心
wsl --unregister Ubuntu