Archives Discussions

[scutan]

Hi, all

I am using the Brook+ to compute on the Stream of type `double`. But the sin cos and sqrt functions only support type float, so the results of sin cos and sqrt are a little different from the results computed by CPU. As there are many such computations in my kernel function, the final results of GPU are very different from that of CPU.

I also want to implement my sin, cos and sqrt function in kernel. But the kernel function dosen't support local array, struct or union. So I can't implement these functions.

Therefore, what should I do to solve this precision problem? The requirements of precision for my program is strict.

Thanks

[scutan]

I wrote the sin cos sqrt function using Taylor series method. Now the problem is solved.

[Jawed]

See post near the end of this thread for double-precision sqrt:

http://forums.amd.com/devforum/messageview.cfm?catid=328&threadid=112855

[scutan]

Originally posted by: Jawed See post near the end of this thread for double-precision sqrt:

http://forums.amd.com/devforum/messageview.cfm?catid=328&threadid=112855

Thank you very much .