Conclusion
Well, did it make sense to switch to ATI Catalyst 8.12 hotfix and does it make sense to install the new catalyst 9.1 now? The numbers grouped in the following summary tables will answer this question best of all:


The performance did improve significantly in two tests out of six we used in our today’s experiments - Crysis Warhead and Enemy Territory: Quake Wars. This gain sometimes reached 30-37%! It is a huge improvement for a new driver version. As a rule, in situations like that performance either doesn’t change at all or changes minimally so that you can hardly notice anything with a naked eye. I have to note that the performance increased stably only in Crysis Warhead, while in Enemy Territory: Quake Wars only dual-chip Radeon HD 4870 X2 benefitted from the new driver versions, which should have got correct multi-GPU support implemented for this game.
Far Cry 2 has also benefitted a bit from the new drivers but only in 1920x1200 and only with Radeon HD 4870 X2. In all other cases, except the first 3DMark Vantage test, there was either no noticeable performance increase, or it was so small that could be regarded as a measuring error. Overall, we see that Catalyst 8.12 hotfix has already unwrapped the main performance potential, so the new recently released catalyst 9.1 has barely added anything to that. Nevertheless, it definitely makes sense to upgrade to the new ATI driver versions, because every now and then the developers may pleasantly surprise us. It will be especially rewarding if you happen to own a Radeon HD 4870 X2 or 4850 X2, or even a discrete CrossFireX system built of several single-GPU Radeon HD cards.



