Running SDK on Linux system

I trying to understand the CUDA API, so i figured the SDK example would be the perfect place to start. I’m having trouble running the sample programs. I have installed CUDAtoolkit3.0, I also understand that the files in the SDK are precompiled. So my question is, how do I run the programs?


I’m not sure what you mean by “the files in the SDK are precompiled”. The directory structure might be a bit confusing, though.
To compile sample, you [font=“Courier New”]cd[/font] to [font=“Courier New”]C/src/sample[/font], type [font=“Courier New”]make[/font] and find the executable in a directory somwhere under [font=“Courier New”]C/bin[/font]

I don’t think they are precompiled either. :rolleyes:

Yes, and if you have gcc 4.4, probably you’ll have to use 4.3 to avoid some errors. Anyway, I don’t think the nVidia SDK is the best place to start, perhaps the nVidia GPU Programming Guide is a better place.