Hi all,
When we use memcpy to copy data from one graphicBuffer to another one or between ANatvieWindows, we found CPU usage is high. At first, we use lock method to get address of data in graphicBuffer, then we can use memcpy to copy data.
I doubt whether memcpy here is act as copying directly between two blocks of GPU memory, or copy from GPU memory to main memory, then copy from main memory to GPU memory again.
Is there any method or function to do data copy between two blocks of GPU memory withou much GPU usage? Thx.
regards,
guangx
