byzhang / cudpp

Automatically exported from code.google.com/p/cudpp
Other
0 stars 0 forks source link

A problem in using cudppSparseMatrixVectorMultiply function #53

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1.Run this program http://pastebin.com/2YvWs8yh  with input in file
'matrix.txt'. 
2.The input file(matrix.txt) is here http://pastebin.com/WDBr2L23
3.

What is the expected output? What do you see instead?
Expected Output : [0,0,0,0]
Output I get : [nan,nan,nan,nan]

What version of the product are you using? On what operating system?
cudpp version
      cudpp_1.0a

uname -a gives 

      Linux deepthought 2.6.23.1-42.fc8 #1 SMP Tue Oct 30 13:55:12 EDT 2007
i686 i686 i386 GNU/Linux

nvcc --version gives 

      nvcc: NVIDIA (R) Cuda compiler driver
      Copyright (c) 2005-2009 NVIDIA Corporation
      Built on Thu_Apr__9_07:37:20_PDT_2009
      Cuda compilation tools, release 2.2, V0.2.1221

Please provide any additional information below.

Original issue reported on code.google.com by kiran202...@gmail.com on 30 Apr 2010 at 3:18

GoogleCodeExporter commented 9 years ago
Kiran, would consider updating to the latest CUDPP build (1.1.1) rather than 
1.0a?

Original comment by jow...@gmail.com on 30 Apr 2010 at 3:56

GoogleCodeExporter commented 9 years ago
Also please consider upgrading to CUDA 3.0.  We no longer test CUDPP on CUDA 
2.2 or even 2.3.

Also, we need to know what GPU you are using, and what NVIDIA driver version.

Original comment by harr...@gmail.com on 30 Apr 2010 at 10:37

GoogleCodeExporter commented 9 years ago
I tested the provided test application with CUDPP 1.1.1 on CUDA 3.0 on Windows7 
64-bit 
with Visual Studio 2008, and I get this output, which I think is expected:

Using device 0: GeForce GTX 480
0.000000
0.000000
0.000000
0.000000

Since we've had no response from the filer, and since I can't reproduce the 
problem, 
I'm  marking this issue invalid for now.  If the filer responds and can provide 
more 
info about failure in CUDPP 1.1.1, then we can reopen.

Original comment by harr...@gmail.com on 13 May 2010 at 12:13