CUDA "Compatible Device" Error on Ubuntu 11.10 / 12.04

I have been trying to set up an Ubuntu environment on my laptop for some time to program CUDA. I am currently downloading Windows 8 and Ubuntu 12.04 twice and want to install CUDA 5 on Ubuntu.

A GeForce GT 640M graphics card is installed on the laptop (see detailed specifications below). This is an Optimus card.

Initially, I was a dual boot Ubuntu 11.10 and tried the tutorials on both 11.10 and 12.04.

I have tried many textbooks of all shapes and sizes, including this tutorial . The installation process shows the installation of the device driver and the installation of the Toolkit, and the samples are unsuccessful, but when I go to test the simple Vector Add CUDA program in NSight, I get the error "No compatible CUDA device".

Ubuntu Details Also Shows Unknown For Graphics

Suggestions?

Notebook Specifications: Acer V3-771G Intel Core i7 2670QM nVidia GeForce GT 640M 2GB - Optimus 16 GB RAM DDR3-1600 120 GB SSD + 500 GB HDD + 32 GB Cache SSD

+4
source share
4 answers

Since this is an optimizer, there are a few extra steps to use the nvidia GPU. Although this is not necessary, I suggest you use the bumblebee wrapper program because this is the easiest solution.

After you have installed the bumblebee shell, you can run your programs with optirun programname or run the shell with the nvidia card activated: optirun bash --login

An added bonus is that the bumblebee daemon will turn off the GPU when it does not work and save you some battery.

If you do not care about battery life and just want CUDA to always turn on without commands, you can load the nvidia kernel module and then manually create the necessary device nodes:

 mknod /dev/nvidia0 c 195 0 mknod /dev/nvidiactl c 195 255 

(This advanced method allows you to run cuda programs from the console without starting Xorg, for example, when you SSH connect to a machine without starting the X server.)

See also https://askubuntu.com/questions/131506/how-can-i-get-nvidia-cuda-or-opencl-working-on-a-laptop-with-nvidia-discrete-car for a more detailed discussion .

+2
source

Try the sudo apt-get install mesa-utils command .

See if the graphics are recognized, and then try to install cuda

If not recognized with the first try command:

 sudo add-apt-repository ppa:ubuntu-x-swat/x-updates sudo apt-get update sudo apt-get install nvidia-current 
0
source

First install the following libraries and Tools:

 sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev 

Next, we will close some modules (drivers), in the terminal enter:

 sudo gedit /etc/modprobe.d/blacklist.conf 

Add the following to the end of the file (one for each line):

 blacklist amd76x_edac blacklist vga16fb blacklist nouveau blacklist rivafb blacklist nvidiafb blacklist rivatv 

Save the file and close the editor. Now we want to get rid of any nvidia risiduals in the terminal:

 sudo apt-get remove --purge nvidia* 

Then you need to restart your computer ( sudo reboot ).

0) Press Ctrl + Alt + F1 on the login screen (you do not need to log in, we still need to restart), then log in.

1) sudo service lightdm stop

2) cd Downloads

3) chmod +x devdriver*.run (driver file name)

4) sudo ./devdriver*.run

You may need to run the installer driver once, reboot (it will remove the nouveau drivers) and repeat the steps again. Follow the instructions of the installer, and everything will be fine when it asks you; yes, you need 32-bit libraries, and you want it to modify the xorg.conf file.

As soon as the installation program completes, restart ( sudo reboot ). You have finished:]

To install the SDK and Toolkit, use steps 3 and 4 with the downloaded files. ( .run )

0
source

In theory, drivers included in CUDA 5.5 should support Optimus (as well as debugging a single GPU for laptops without Optus). I have not tried it yet, because I am expecting the Optimus laptop to calculate 3.5 to support kernel recursion and HyperQ. Theoretically, the HP Envy 15t-j000 has a GK208 version of the GT 780m, but I'd rather have an Ultrabook form factor, like the upcoming Acer S3-392 with GT 735m. The NVIDIA guys at GTC assured me that Optimus should work with CUDA 5.5 RC. I found this "CUDA Getting Started Guide for Linux" released this month, which contains some flags for installing Optimus drivers correctly:

http://www.google.com/url?q=http://developer.download.nvidia.com/compute/cuda/5_5/rc/docs/CUDA_Getting_Started_Linux.pdf

In addition, more information about the GK208 and Compute 3.5 chips in laptops:

https://devtalk.nvidia.com/default/topic/546357/sounds-like-gk208-laptops-cards-will-support-most-sm_35-features/

Someone lucky with CUDA 5.5 and Optimus laptops under Linux?

0
source

Source: https://habr.com/ru/post/1444602/


All Articles