backing up state

slaclab · Jul 22, 2024 · 3426515 · 3426515
1 parent 15aa096
commit 3426515
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 8 deletions.
diff --git a/data_gpu/driver/README.md b/data_gpu/driver/README.md
@@ -1,18 +1,14 @@
 # GPU Enabled Driver
 
-To build this driver you need to have the NVIDA Open GPU Kernel Modules installed. This driver will not compile gainst the CUDA toolkit drivers.
+To build this driver you need to have the NVIDA Open GPU Kernel Modules installed. This driver will not compile against the CUDA toolkit drivers.
 
 https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
 
 See section 5.
 
-A script in this directory 'comp_and_load_drivers.sh' is provided to compile and load the nvidia drivers as well as the driver in this directory. Edit this file and update the NVIDIA_PATH value at the top to the install directory for the nvidia drivers.
+<!--- ######################################################## -->
 
-```bash
-$ sudo apt-get install nvidia-kernel-source-545-open
-$ sudo apt-get install nvidia-cuda-toolkit
-$ sudo ./comp_and_load_drivers.sh
-```
+# System Configurations
 
 Disable the Xserver and nvidia-persistenced to prevent rmmod due to Module XXX is in use by: YYY
 because the Nvidia driver gets loaded by default at startup 
@@ -25,3 +21,34 @@ $ sudo systemctl disable lightdm # For LightDM
 $ sudo systemctl disable sddm    # For SDDM
 $ sudo systemctl disable nvidia-persistenced
 ```
+
+Add the nvida cuda package for nvidia-545.23.08 and install the cuda toolkit:
+
+```bash
+$ wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.0-1_all.deb
+$ sudo dpkg -i cuda-keyring_1.0-1_all.deb
+$ sudo apt update
+$ sudo apt install nvidia-kernel-source-545 # Tested with nvidia-545.23.08
+$ sudo apt install nvidia-cuda-toolkit
+```
+
+Add `iommu=off nouveau.modeset=0 rd.driver.blacklist=nouveau` GRUB_CMDLINE_LINUX:
+
+```bash
+$ sudo nano /etc/default/grub
+      GRUB_CMDLINE_LINUX="iommu=off nouveau.modeset=0 rd.driver.blacklist=nouveau"
+$ sudo update-grub
+$ sudo reboot
+```
+
+<!--- ######################################################## -->
+
+# How to build and load the nvidia and datagpu drviers
+
+After you completed all the "System Configuration" configuration steps above, run the following script to build and load the nvidia and datagpu drviers
+
+```bash
+$ sudo ./comp_and_load_drivers.sh
+```
+
+<!--- ######################################################## -->
diff --git a/data_gpu/driver/comp_and_load_drivers.sh b/data_gpu/driver/comp_and_load_drivers.sh
@@ -74,6 +74,8 @@ fi
 
 # Go to nvidia path and build Nvidia driver
 cd "$RET_DIR" || { echo "Error: Failed to change directory to $RET_DIR"; exit 1; }
+# Clean previous builds
+make clean
+# Build datagpu driver
 make CC=$CC NVIDIA_DRIVERS=$NVIDIA_PATH
 /usr/sbin/insmod $RET_DIR/datagpu.ko || { echo "Error: Failed to insert datagpu.ko."; exit 1; }
-