A couple of days ago, Nvidia release a driver update:

New in Version 285.62

  • This is the recommended driver for Battlefield 3. It contains several performance and compatibility enhancements for the final release of the game. Check here to see if your PC is ready for Battlefield 3.

  • This is also the recommended and enhanced driver for Batman: Arkham City and RAGE. Check here to see if your PC is ready for Batman: Arkham City

  • Contains a fix for the driver timeouts reported with the R285 beta drivers.

What is about this release that makes it the recommended version for these games? I can think of three possibilities:

  1. There is nothing about the driver that makes it good specifically for those games. It's just a general performance increase for any software, and the mention of the games, along with the timing of the release, is just a PR exercise.
  2. While those games were in development, the developers noticed certain driver-dependant functionality that wasn't as fast as they expected it to be. They go to Nvidia, say 'Hey, what's up with feature X taking Y milliseconds instead of Y/2 milliseconds? " Nvidia say "Hey, you know what, we can improve that! " and so the next driver update will make the game run better.
  3. The driver actually has special cases in it for the game. E. G if (application == 'BF3'){DoABunchOfTweaksSpecificallyForBF3();}

I suspect the truth is somewhere between 1. And 2. But does anyone have an informed answer?

  • I don't have an informed answer but I wouldn't rule out 3 by a long, long shot. If people don't buy games, they won't buy graphics cards, so I would fully expect cooperation between game makers and GPU driver writers. – LawrenceC Oct 26 '11 at 23:09
  • All of the above. Mostly #3 in a slightly different context. Game specific Crossfire/SLI profiles, shader type/usage, etc. – Lamar B Oct 26 '11 at 23:20

More and more, the real answer is 3. Video drivers are getting more and more application-specific tweaking. It's not just disabling costly features but tweaking of all kinds.

  • and testing the driver for all the bugs reported by users in the specific application. – Psycogeek Oct 26 '11 at 23:30
  • 2
    Oh wow, I wouldn't have expected that. Do you have any sources? – Cam Jackson Oct 27 '11 at 0:08
  • no I dont, do you think they just ignore all the bug reports , and just guess at what might need to be adjusted most :-) Check out AMD and NVIDA forums and locations for bug reports. ATI made a (stupid) promise to put out a new driver monthly, that means either keep adding it junk or fix something – Psycogeek Oct 27 '11 at 5:05
  • 1
    @Psycogeek My comment was directed more at David, as to whether he has any sources that they put in application-specific tweaks. Obviously they get bug reports for specific applications, but my question is whether they go "Oh, that bug has always been there, but <BF3> is the first application to do <some obscure sequence of operations> and exhibit the bug." So the fix is really a general improvement/bug-fix, not really application-specific. Or are they actually writing in special case code that changes behaviour for some applications and not others, to make them run better? – Cam Jackson Oct 27 '11 at 23:49
  • 1
    Most bugs are bugs that were always in the driver that were triggered by unusual application use and thus reported and seen with a specific application. Bugs in application-specific code are possible but rare -- obviously application-specific code is well-tested with the application it is specific to. – David Schwartz Oct 28 '11 at 0:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.