Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I'm truly fascinated by the idea of GPGPU and using the GPU for heavy processing. I'm seeing that also APUs (Accelerated Processing Units, CPU+GPU on the same chip) are gaining a consistent popularity.

Are all of the APUs using a GPGPU? Can it be used for processing? And is it seamless or it requires special code (like Cuda) to have the hard work made by the GPU?

I'm not interested in bare graphic performance, but more about how much the GPU can accelerate the "normal" CPU work.

share|improve this question
    
If the GPU could accelerate normal CPU work, then it would be considered part of the CPU part of the APU, not the GPU. –  David Schwartz Jun 11 '12 at 9:29
    
@DavidSchwartz: and how do you compare this to GPGPUs? "Normal" is not intended in the strict sense, but to mean computation not specific to graphics. Like matrices computation, for instance –  clabacchio Jun 11 '12 at 9:33
add comment

1 Answer

How much a GPU can accellerate depends on the code you run. GPUs are extremely good in running simple, massively paralel instructions. Programs which can use that can gain a massive performace boost. Code which is single treaded or complex will fair quite poorly.

share|improve this answer
    
I understand that, but will any code run on the GPU as well or only specifically optimized software? –  clabacchio Jun 11 '12 at 9:59
    
If CPU code runs on it, then it's part of the CPU. The boundary between the CPU and the GPU is defined this way. –  David Schwartz Jun 11 '12 at 10:09
    
@clabacchio In case of AMD Fusion, APU is simply CPU + GPU on one chip, and to use GPU's power, you need GPGPU-aware software (written with e.g. OpenCL). –  aland Aug 6 '12 at 21:55
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.