Jump to content

Recommended Posts

Posted

Please note that unlike 3DMark, Furmark or many other 3D benchmarks or 3D games, AIDA64 uses GPGPU compute tasks (via OpenCL) to put stress on the GPUs and APUs.  It means, it has to call OpenCL to first copy a block of memory from system memory into the video memory, then start a program that performs the actual computation, and then when it's finished, it has to copy the memory back from the video memory to the system memory, in order to let it verify the results.  And since all those tasks have to go through OpenCL driver, then the video driver (ForceWare in your case), the PCI Express link, and also the GPU itself, there's a lot of latency and bottlenecks involved in the process.  It effectively means the GPU may not be driven 100% constantly, but sometimes a downward spike may appear when the OpenCL driver, the PCI Express link or the video driver lags a bit.  That's normal, and can only be avoided by directly programming the GPU (which is impossible), or using a 3D scene instead of a GPGPU code.

 

However, the advantage of GPGPU stress test over 3D scenes is that using GPGPU code it's much easier to stress all available GPUs simultaneously.

 

 

Regards,

Fiery

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.



×
×
  • Create New...