Why does Xplorer in benchmark mode use a lower frequency than a pro-rated achieve frequency?
In the last run (usually the seventh run), the target frequency can be lower than the achieve frequency in previous runs. The strategy implemented for the last run is not intuitive and appears to be counterproductive. However, Xilinx benchmarking data shows that some designs that were under-constrained actually had a faster frequency than when the clock frequency was over-constrained. As a result of this benchmarking data, the last run implements this strategy.