I have successfully(?) installed CUDA and the required components and am trying to compile the Szeta.cu file in the MATLAB Plug-in example, and each time I do, the following text is generated:
abdelali target arch: win32
square_me.cu
tmpxft_000013cc_00000000-3_square_me.cudafe1.gpu
tmpxft_000013cc_00000000-8_square_me.cudafe2.gpu
tmpxft_000013cc_00000000-3_square_me.cudafe1.cpp
Is this something that I need to worry about, or can I ignore this? The reason I ask is that the performance boost is not what I expected (only about 2x).
I am running MATLAB R2008a on Vista Business with CUDA 2.2 and MSVS 9.0. I ended up modifying the nvmex.pl and nvmexopts.bat files from [url=“http://forums.nvidia.com/index.php?s=&showtopic=67538&view=findpost&p=549631”]http://forums.nvidia.com/index.php?s=&...st&p=549631[/url] (had to change from x64 to x86 Vista), and have attached the modified copies below.
If anyone can offer any insight, it would be greatly appreciated!
nvmex.pl.txt (106 KB)
nvmexopts.bat.txt (3.03 KB)