CUDA LEARNING
CUDA, installation issues, learning, samples, etc
Sunday, October 17, 2010
GTX4XX cards and issues with HDMI sound with UBUNTU
I setup the GTX460 driver on ubuntu and installed all the necessary drivers. Couldnt get the sound out through the HDMI. Any suggestions welcome.
Friday, October 15, 2010
CUDA resources
Here is a list of few resources for getting started with parallel computing and CUDA
http://www.nvidia.com/object/cuda_home_new.html
http://en.wikipedia.org/wiki/CUDA
http://courses.engr.illinois.edu/ece498/al/
Books
CUDA by Example
Programming Massively Parallel Processors: A Hands-on Approach
http://www.nvidia.com/object/cuda_home_new.html
http://en.wikipedia.org/wiki/CUDA
http://courses.engr.illinois.edu/ece498/al/
Books
CUDA by Example
Programming Massively Parallel Processors: A Hands-on Approach
Thursday, October 14, 2010
INSTALLING CUDA ON UBUNTU 10.04
For a while now, I have been interested in parallel computing. With graphics card becoming programmable, the sheer computing power at user disposable is just amazing. I assembled a PC a few days ago and decided to install Ubuntu 10.04 as I didn’t want to spend money on OS. I am a newbie to linux and CUDA and I decided to share my experiences with everyone and simplify the learning curve for most of us.
As my Laptop did not have hardware support for CUDA, I assembled a PC with Nvidia graphics card. The details are as follows.
Case: ThermalTake V9 Black
MotherBoard: EVGA P55 SLI
Power Supply : ANTEC 850 W Quatro true power modular PSU
RAM: Corsair 4gb DDR 3
Graphics card: EVGA GTX460
After assembling the PC and installing Ubuntu, the next task was installing CUDA. Though NVIDIA has a Linux getting started guide, it is generic in nature with OS specific details missing. After reading a couple of blogs, technical resources, and websites, the set of instructions needed to successfully install CUDA on a PC running ubuntu should be as follows.
1.) This process assumes that ubuntu has been freshly installed on the computer.
2.) Download the graphics card driver, CUDA toolkit and SDK samples and save it on the computer.
sudo apt-get purge nvidia-*
3) For each of installation and avoiding conflicts, we make a file in following location - "/etc/modprobe.d/nvidia-graphics-drivers.conf" and insert the following lines
blacklist vga16fb
blacklist nouveau
blacklist lbm-nouveau
blacklist nvidia-173
blacklist nvidia-96
blacklist nvidia-current
blacklist nvidiafb
4) After this step, follow the procedure as documented in the Linux getting started guide available from Nvidia
5) After using the Make command in SDK, I noticed that libraries like XLi and XMu were missing. Install the -dev version of these libraries and issue Make command again. That will solve the issue.
Subscribe to:
Posts (Atom)