The Elder Scrolls: Oblivion
Version and / or Patch Used: Default install
Timedemo or Level Used: N/A
Developer Homepage: http://www.bethsoft.com/games/games_oblivion.html
Product Homepage: http://www.elderscrolls.com/games/oblivion_overview.htm
One of the latest games to hit the market, TES: Oblivion has been taking the world by storm. With minimum system requirements that are nearly unheard of and recommended system levels toward the top of the spectrum, this title is guaranteed to stress out any system. It makes use of large amounts of memory, fast processors, faster data bus streams as well as support for all of the latest graphics card enhancements. Whether testing for Shader Model v3 optimizations or HDR lighting effects, this benchmark will successfully test any of the new components and feature sets currently offered in new hardware.
At this time, there is no consistent method to get an identical test when benchmarking with TES: Oblivion. The reason for this is twofold; there is no current way to make a demo run that is replayable, and the game itself randomly generates terrain details on the fly. Testing was conducted using a saved game and performing the same actions for each board and resolution being tested. This does not provide a perfect method of testing performance, so these results are only being included to give an idea as to how well a tested product will handle this new power-hungry game. It should not be considered as a true and accurate result since there is room for error in the methodology used.
Well, here is the new kid on the block and all I can say is DAMN! Testing aside, if you haven't tried this game yet, you're missing a real treat. As mentioned above, however, I cannot give a lot of weight to this particular test just yet. The lack of any true demo capabilities and the random generation of terrain objects makes it an unreliable test on its own... BUT, it can still be used in a general way to see what type of performance we can expect when playing this particular title.
During testing, I once again set all detail levels to their maximum and turned on the HDR lighting effects. This causes about an 18% drop in frames, but the eye candy factor jumps way up the scale. Besides, since we're trying to stress out these boards as much as possible, I didn't figure it hurt to go for maximum details while testing.
Speaking of testing, our numbers show a performance increase of between 5-10% in favor of the GeCube X1900XTX board. Remember, I cannot validate these numbers, but I can say that I practiced my run-through many times before actually testing either board, so the runs are at least somewhat consistent.