wrong result from CNC's code?

Hi everybody,

In case u didn’t know, CNC’s code solve the sparse linear equations using CUDA.

I tried the CNC code and when i constructed the simplest example to solve. The code does not output the correct vector. I wonder why this is happening. I didn’t change anything to the Visual studio project and just add a new test data file to test. It’s as follows:
a 0 0 1
a 0 1 1
a 1 0 1
a 1 1 -1
b 0 2
b 1 0
And the result is : array[0] = 2, array[1] = 0, which is not correct. This is a simple equation"x+y = 2; x-y = 0".

If anybody is using CNC’s code and had similar problems, please reply. Really appreciated.

Nobody is using this code?

Just because an application uses CUDA doesn’t mean that there will be people here who can troubleshoot specific problems with that application. You will have better luck on a CNC specific forum/mailing list. Most who read this forum are CUDA programmers working on their own applications.