Hello guru OpenCL. I have a question and would be very grateful if you answer.
I have a process that uses the GPU, but it also uses 12 to 14% of the CPU. I need to run in parallel for more than 4 of these processes, so that the load on the CPU becomes indecent. Although the process any transactions with the CPU does not perform, it works only with API OpenCL. Is there any explanation for this? (I use i7 - 2600K, HD6950)
Thanks in advance.
i guess same problem is discussed here: http://forums.amd.com/devforum/messageview.cfm?catid=390&threadid=153211&enterthread=y
if you have 13% cpu load on an 8 core CPU = 100% on 1 core