We noticed a significant drop in OpenGL performance compared to the main competitor GPUs, when the GPU hardware itself barely utilizes 40-50% of GPU time when the CPU OpenGL thread is capped by the CPU, indicating that the AMD driver is not well optimized well in that case. I have heard that it might be related to the competitor's threaded-optimization feature lacking in AMD. Does that ring any bell, do you know of any solution, or do you have any further thoughts on the issue?