The AotS benchmark doesn't report GPU utilization, it reports % time the benchmark was GPU limited or CPU limited. Even when 49% of the time benchmark was CPU bound that might still mean 99% GPU utilisation
Quoted the numbers AMD gave us on Reddit. Source.
"2x Radeon RX 480 - 62.5 fps | Single Batch GPU Util: 51% | Med Batch GPU Util: 71.9 | Heavy Batch GPU Util: 92.3% GTX 1080 – 58.7 fps | Single Batch GPU Util: 98.7%| Med Batch GPU Util: 97.9% | Heavy Batch GPU Util: 98.7%"
"//EDIT: To clarify this, the scaling from 1->2 GPUs in the dual RX 480 test we assembled is 1.83x. The OP was looking only at the lowest draw call rates when asking about the 51%. The single batch GPU utilization is 51% (CPU-bound), medium is 71.9% utilization (less CPU-bound) and heavy batch utilization is 92.3% (not CPU-bound). All together for the entire test, there is 1.83X the performance of a single GPU in what users saw on YouTube. The mGPU subsystem of AOTS is very robust."
Really confusing. but as I understand it, scaling is 1.83 times that of a single gpu, while the GPU utilization is lower because the cards are CPU bound being in crossfire (simpler batches). So if we want a valid comparison, we should probably look at the batch where GPU utilization was above 90%.
Last edited: