Difference between revisions of "Nvidia Drivers and CUDA"

From Interaction Station Wiki
Jump to navigation Jump to search
(Created page with "==Installing CUDA 10.0 on Ubuntu 16.04== check nvidia drivers and cuda version: nvidia-smi Download the latest version of CUDA which is compatible with your version of driv...")
 
Line 1: Line 1:
 
==Installing CUDA 10.0 on Ubuntu 16.04==
 
==Installing CUDA 10.0 on Ubuntu 16.04==
 +
 +
'''General Info'''
  
 
check nvidia drivers and cuda version:
 
check nvidia drivers and cuda version:
 +
nvcc --version
 
  nvidia-smi
 
  nvidia-smi
 +
  
 
Download the latest version of CUDA which is compatible with your version of drivers installed. For example, if you have installed a driver of 410 version, CUDA 10 can be installed, for 375 CUDA 8 should be installed.
 
Download the latest version of CUDA which is compatible with your version of drivers installed. For example, if you have installed a driver of 410 version, CUDA 10 can be installed, for 375 CUDA 8 should be installed.
Line 9: Line 13:
 
Avoid CUDA 10.1, it conflicts with tensorflow-gpu
 
Avoid CUDA 10.1, it conflicts with tensorflow-gpu
  
Installing CUDA 10.0
 
  
First uninstall  previous versions of CUDA:
+
'''Installing CUDA 10.0'''
    sudo apt-get --purge remove cuda-10.1
 
    sudo apt-get remove nvidia-cuda-toolkit
 
  
Then
+
'''1. Uninstall  previous versions of CUDA'''
sudo dpkg -i cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
+
sudo apt-get --purge remove cuda-10.1
sudo apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
+
sudo apt-get remove nvidia-cuda-toolkit
sudo apt-get update
 
sudo apt-get install cuda-10-0
 
  
 +
'''2. Then install:'''
 +
sudo dpkg -i cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
 +
sudo apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
 +
sudo apt-get update
 +
sudo apt-get install cuda-10-0
  
Steps here and installation file downloaded from here:
 
    https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal
 
  
 +
Steps and installation file downloaded from here:
 +
https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal
  
To verify Cuda installation:
+
'''3. Specify your cuda path (cuda 10.0 example)'''
nvcc --version
 
nvidia-smi
 
  
 +
sudo nano /home/interactionstation/.bashrc
  
Specify your cuda path (cuda 10.0 example)
+
locate the two lines regarding cuda9 (in case it was previously installed), comment them out and include instead:
export PATH=/usr/local/cuda-10.0/bin:$PATH
+
export PATH=/usr/local/cuda-10.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH
+
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH
  
Restart your computer.
+
'''4. Save the file'''
  
https://medium.com/@muskulpesent/install-cuda-on-ubuntu-16-04-67ec087f8561
 
  
  
sudo nano /home/interactionstation/.bashrc
 
  
locate the two lines regarding cuda9, comment them out and include instead
+
'''5. Restart the computer'''
export PATH=/usr/local/cuda-10.0/bin:$PATH
 
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH
 
  
Save the file
+
More info:
 
+
https://medium.com/@muskulpesent/install-cuda-on-ubuntu-16-04-67ec087f8561
 
+
https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal
 
 
These instructions:
 
    https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal
 
   
 
   
 
  
 
== Installing cuDNN (for ubuntu 16.04, cuda 10-0) ==
 
== Installing cuDNN (for ubuntu 16.04, cuda 10-0) ==

Revision as of 21:03, 27 November 2019

Installing CUDA 10.0 on Ubuntu 16.04

General Info

check nvidia drivers and cuda version:

nvcc --version 
nvidia-smi


Download the latest version of CUDA which is compatible with your version of drivers installed. For example, if you have installed a driver of 410 version, CUDA 10 can be installed, for 375 CUDA 8 should be installed. CUDA 10.0 (10.0.130) >= 410.48 driver

Avoid CUDA 10.1, it conflicts with tensorflow-gpu


Installing CUDA 10.0

1. Uninstall previous versions of CUDA

sudo apt-get --purge remove cuda-10.1
sudo apt-get remove nvidia-cuda-toolkit

2. Then install:

sudo dpkg -i cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda-10-0


Steps and installation file downloaded from here: https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal

3. Specify your cuda path (cuda 10.0 example)

sudo nano /home/interactionstation/.bashrc

locate the two lines regarding cuda9 (in case it was previously installed), comment them out and include instead:

export PATH=/usr/local/cuda-10.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH

4. Save the file



5. Restart the computer

More info: https://medium.com/@muskulpesent/install-cuda-on-ubuntu-16-04-67ec087f8561 https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=deblocal

Installing cuDNN (for ubuntu 16.04, cuda 10-0)

       Download the:
           - runtime
            https://developer.nvidia.com/compute/machine-learning/cudnn/secure/7.6.4.38/Production/10.0_20190923/Ubuntu16_04-x64/libcudnn7-dev_7.6.4.38-1%2Bcuda10.0_amd64.deb
    - developer

https://developer.nvidia.com/compute/machine-learning/cudnn/secure/7.6.4.38/Production/10.0_20190923/Ubuntu16_04-x64/libcudnn7_7.6.4.38-1%2Bcuda10.0_amd64.deb

    - samples

https://developer.nvidia.com/compute/machine-learning/cudnn/secure/7.6.4.38/Production/10.0_20190923/Ubuntu16_04-x64/libcudnn7-doc_7.6.4.38-1%2Bcuda10.0_amd64.deb and then:

Navigate to your <cudnnpath> directory containing cuDNN Debian file. Install the runtime library, for example: sudo dpkg -i runtimefile.deb Install the developer library, for example: sudo dpkg -i devlib.deb Install the code samples and the cuDNN Library User Guide, for example: sudo dpkg -i codesamples.deb