Benchmark optimizations or what NVIDIA was keeping silence about
The editor of our Video section, Andrew Vorobyev, told us a very interesting story related to drivers "optimizations".
"Weíve recently posted a number of reviews dedicated to benchmark "optimizations" implemented by ATI and NVIDIA (it turned out the latter implemented more of them). We found out that all NVIDIA drivers up to 44.65 inclusive contain "optimizations" for all games that have benchmarks (itís strange, but none have been made for Unreal II, though itís based on the same engine, as UT2003, but still doesnít contain a benchmark.)
But there are some optimizations for UT2003. Funny, isnít it? Unreal II performs with Anti-detect same as without it. I doubt that Unreal II just doesnít require any optimizations. Unfortunately it just doesnít have an exposed benchmarking tool to "optimize" :). NVIDIA specialists might have actually overlooked the tweakers who found scripts in Unreal II for executing the test.
Anyway, the thing is that in Code Creatures Anti-Detect found out that speed of all GeForce FX cards was "optimized" by cheating, as Code Creatures is not a game, but a pure benchmark. E.g., FX 5900 shows 46 fps (1024x768, 32bpp), but only 33 fps with Anti-Detect. And the newly leaked 44.90 beta doesnít seem to contain any optimizations, as we obtained the same fps values on the same card at the same resolutions, as 44.03 and 44.65 versions had provided with Anti-Detect. Weíll continue checking the situation, and it will be very good if optimizations have really been removed from other tests as well!"
Write a comment below. No registration needed!