The IO device is called PS/2, not P/S 2. It stands for Personal System/2, the IBM product line from 1987. Also, there's a chance there's a big performance gain from Catalyst 14.2, instead of 14.1. Another site did benchmarks comparing 13.6 and 14.2, and the difference was dramatic. It most likely was the somewhere in the 13.6 to 14.1 range, though. I thought the days of one motherboard having significantly faster memory timings than the other died with the IMCs. After being shocked at the horrible performance of the Kaveri, and the almost complete failure of it, between the driver update, and advantage the Asus has, it's clear AMD pulled another "Hawaii", and put their technology in the worst possible light by not packaging it with the proper associated support technologies. Considering the long development cycles for these devices, it's hard to understand how AMD couldn't figure out a proper cooler for the Hawaii, or have drivers degrade performance so much for Kaveri, that within a month or so they were able to increase performance so dramatically. And now, it's clear the memory performance optimizations were far from complete. And Kaveri was a delayed product. It boggles the mind that they consistently fail to find such obvious shortcomings, when everyone else finds them pretty quickly. Clearly, their testing procedures need serious revision in scope.