A quick start on CUDA: Difference between revisions

Revision as of 23:31, 3 March 2012

All machines in coover 2050 are installed with CUDA.

Before you run a simple program, you need to set up environment variables

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

The cuda sample code can be downloaded from:CUDA by example

1) go to the directory of your sample code.

2) To compile hello_world.cu

  nvcc hello_world.cu

3) Run the executable file: a.out

 ./a.out

4) Get the output.

@@ Line 1: / Line 1: @@
 All machines in coover 2050 are installed with CUDA.
-Before you run a simple program, you need to:
+Before you run a simple program, you need to set up environment variables
- Setting environment variables
    export PATH=/usr/local/cuda/bin:$PATH
    export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
+The cuda sample code can be downloaded from:[http://developer.nvidia.com/cuda-example-introduction-general-purpose-gpu-programming CUDA by example]
 ) go to the directory of your sample code.
@@ Line 13: / Line 14: @@
     nvcc hello_world.cu
-) This will generate an executable file: a.out
+) Run the executable file: a.out
-) ./a.out
-The cuda sample code can be downloaded from:[http://developer.nvidia.com/cuda-example-introduction-general-purpose-gpu-programming CUDA by example]
-hello_world.cu is in chapter03.
+  ./a.out
+) Get the output.