You already have the needed background on what we're doing, so we'll not repeat that. However, be aware that Cypress is a cheeky and tricky beast when it comes to its memory interface, due to its re-training capacity, which probably tightens timings for lower memory clocks, and EDC, which may interfere at higher clocks, where transmission errors may occur, forcing re-transmission and thus lowering performance. We'd expect EDC to be a non-issue in our case, since we're not going out of spec with the clocking, but retraining will make the waters a tad murkier...sadly, that's one parameter we can't mess with for the time being. Here's what happened in practice:



It appears Cypress is not that sensitive to the memory clock/bandwidth to VRAM (taking into account the caveats we mentioned upstream), which is slightly surprising to us-and a testament to ATI's engineers that seem to have nicely balanced the chip, at least when it comes to gaming workloads. Moving on, we've computed the fractions affected by our tinkering:

Finally, it's time to look at how an external factor (the CPU clock) impacts performance, and put all these 3 aspects together to get an image of Cypress' performance constraints.