Update content src/site/notes/AI/nvidia/nvidia相关资源.md

src/site/notes/AI/nvidia/nvidia相关资源.md

@@ -2,13 +2,34 @@
 {"dg-publish":true,"permalink":"/AI/nvidia/nvidia相关资源/","noteIcon":"3"}
 ---
 
-#nvidia #docker
+#nvidia #docker #cuda
 https://blog.csdn.net/seasermy/article/details/105298646
 
 [nvidia申请账号密钥密码](https://ngc.nvidia.com/setup/api-key)
 
 [nvidia相关镜像tag](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch/tags)
 
+nvidia cuda toolkit各个版本下载：
+https://developer.nvidia.com/cuda-toolkit-archive
+```sh
+# 环境中存在多个cuda toolkit版本的时候，指定特定的cuda版本, 注意环境中的CUDA_HOME环境变量可能优先级更高，需要先unset
+export CUDA_PATH={path_of_cuda_11.8}
+export GCC_HOME={path_of_gcc_10.2.0}
+export MPFR_HOME={path_of_mpfr_4.1.0}
+export LD_LIBRARY_PATH=${GCC_HOME}/lib64:${MPFR_HOME}/lib:${CUDA_PATH}/lib64:$LD_LIBRARY_PATH
+export PATH=${GCC_HOME}/bin:${CUDA_PATH}/bin:$PATH
+export CC=${GCC_HOME}/bin/gcc
+export CXX=${GCC_HOME}/bin/c++
+```
+
+
+
+查看cuda版本
+https://www.cnblogs.com/wuliytTaotao/p/11453265.html
+```sh
+nvcc --verison
+```
+
 
 
 ```sh
@@ -26,4 +47,24 @@ docker run -it  --gpus all --net host --name xxl-dahua1 \
 --shm-size=64G \
 -v /data:/data \
 nvcr.io/nvidia/pytorch:23.04-py3 /bin/bash
+```
+
+查看当前nvidia服务器是否有nvlink，以及nvlink的速率
+
+```sh
+$ nvidia-smi -q | grep -i nvlink
+        NVLink 0: Enabled
+        NVLink 1: Enabled
+        Nvlink Speed: 25.78 GB/s
+        Nvlink Counter: 0
+        Nvlink Counter: 0
+
+nvidia nvlink --status
+```
+
+
+指定卡训练
+```sh
+export CUDA_VISIBLE_DEVICES=7
+
 ```