more dodgy detonator drivers?
Home | Reviews and Features | Special Reports | Forums |

Page 1 of 2 12 LastLast
Results 1 to 15 of 19

Thread: more dodgy detonator drivers?

  1. #1
    Join Date
    Feb 2003
    Location
    Poole, UK
    Posts
    1,999

    Exclamation more dodgy detonator drivers?

    The other day I downloaded eVGA's 44.67 detonator drivers and took 3dmark 03 (build 330) for a spin to see if they made any difference. The drivers I had before then were 44.04.

    My score went up from ~4900 to ~5900 which seems a hell of a jump. Is this another case of drivers which have been "optimized" for 3dmark, like the old 43.xx ones were (I forget the exact version) or are these genuine improvements?

    I guess the way to tell would be to try something other than 3dmark but I don't have UT2003 or anything like that. I just wondered if anyone else had tried these drivers and what there thoughts are.

    Thanks.
    Windows 7 x64 Ultimate Edition, Gigabyte Z77-D3H, Ivy Bridge i7 3770, nVidia GTX480 1.5Gb, 16Gb Corsair 1600MHz, Creative X-Fi, Seagate 500Gb SATA II, DGM 23" Widescreen TFT

    "Men never do evil so completely and cheerfully as when they do it from religious conviction" -Blaise Pascal

    M.C.S.A.

  2. #2
    Join Date
    Apr 2000
    Location
    UK
    Posts
    8,466
    Yes, the gains are for the most part due to application-specific optimisations although there are obviously some genuine improvements hidden in there as well. You can argue against this but there is a hell of a lot of evidence now, backing it up.

    They are most likely still inserting clip planes into the scene (to avoid having render a substantial proprtion of it) and using shader substitutions at the cost of IQ, but have installed a mechanism to stop such cheats being detected. Since putting an incredible amount of pressure on FutureMark they seem to be able to get away with it.

    UT2003 won't help resolve the situation. nV are taking shortcuts in that as well: http://www.amdmb.com/article-display.php?ArticleID=247

    It seems they are using a hack that completely bypasses anisotropic filtering in order to substantiate the claims of "30% performance improvement" in their drivers. There is a [H]ard|OCP article currently published which addresses a similar issue, but they manage not to notice anything weird at all! The problem is nVidia are taking away the freedom of the user to determine what IQ he/she wishes. As far as I know, full trilinear filtering is now not an option in the latest drivers.

    Such things *should* stop when they release more competent hardware - by the end of the year maybe. At the moment it's all damage control and with the latest revelations that NV3x doesn't actually completely suuport the PS/VS 2.0 model, there should be more to come.

    MuFu.

  3. #3
    Join Date
    Feb 2003
    Location
    Poole, UK
    Posts
    1,999

    Angry What a surprise

    Thanks for the info MuFu.

    I bought the 5900 Ultra after reading that it beat the 256meg 9800 in most (not all) tests - this was before nvidia's driver optimisations.

    I still think I have made the right choice in terms of out-and-out performance, if not value for money, but it does annoy me that nvidia are negating the whole concept of benchmarking. I can understand some of the vitriol I have read on other sites from ATI fans. Surely nVidia credit us enough intelligence not to fall for these sorts of tactics........though apparently not.

    I guess from what you are saying that I should pay more attention to my original score of 4900 ish?
    Windows 7 x64 Ultimate Edition, Gigabyte Z77-D3H, Ivy Bridge i7 3770, nVidia GTX480 1.5Gb, 16Gb Corsair 1600MHz, Creative X-Fi, Seagate 500Gb SATA II, DGM 23" Widescreen TFT

    "Men never do evil so completely and cheerfully as when they do it from religious conviction" -Blaise Pascal

    M.C.S.A.

  4. #4
    Join Date
    Apr 2000
    Location
    UK
    Posts
    8,466
    I wouldn't read too much into the overall score at all. The card is still a fine performer, it's just the price and accompanying PR bull**** that suck. Are FutureMark currently allowing 5x00 submissions to the ORB? I remember reading a while ago that they had stopped letting anybody with one of the dodgy driver sets post their score. Might be wrong though.

    I just hate the hypocrisy of it all - they make all these claims regarding the best gaming exprience for the user ("The Way it's Meant to Be Played" etc) and then force IQ hacks which get them a few cheap benchmarking wins but are in fact to the overall detriment of the game-playing public. They are also giving reviewers nightmares - it is *so* difficult to tell these days whether what you are seeing in benchmarks reflect actual gameplay so writing an honest review is extremely difficult. There are a few articles out now where popular benchmarks are avoided and the 9800 Pro seems to win the majority of tests.

    One interesting thing I saw a webmaster point out recently was the fact that whenever people in a similar position go to visit nVidia, one of the first things they are asked is "what benchmarks are you using?"!!!

    MuFu.

  5. #5
    Join Date
    Feb 2003
    Location
    Poole, UK
    Posts
    1,999
    Roll on NV40 then, I guess.....

    not that I'll be able to afford it
    Windows 7 x64 Ultimate Edition, Gigabyte Z77-D3H, Ivy Bridge i7 3770, nVidia GTX480 1.5Gb, 16Gb Corsair 1600MHz, Creative X-Fi, Seagate 500Gb SATA II, DGM 23" Widescreen TFT

    "Men never do evil so completely and cheerfully as when they do it from religious conviction" -Blaise Pascal

    M.C.S.A.

  6. #6
    Join Date
    Jun 2002
    Posts
    853
    At this rate we will be seeing the original Geforce SDR getting 90+ FPS in DIII with all eye candy enabled....
    Nforce 2 Mobo-oem stock
    Athlon Xp3000-Stock
    768 mgs PC2100 @ 2.5/3/3
    Radeon 9700 PRO/365/342
    Western Digital 8mg cache ATA-133/160G/7200 HD
    Win XP Home SP1
    19" KDS "Extreme Flat"

    Laptop
    Athlon 64 3200+
    512mg Ram
    Mobile Radeon 9600
    80G Hard Drive


  7. #7
    Join Date
    Dec 1998
    Location
    Grand Haven, Michigan, USA
    Posts
    11,332
    Another new one that [H]ardOCP found is that the Quality setting under the desplay properties in the latest drivers use bi-liniar filtering instead of tri-liniar. They also got rid of the application setting now. More artificial bumps in performance at the users cost.
    AMD Phenom II x4 945 3Ghz | ASUS M4A77TD | 2X WD 1TB SATA 2 hard drive | 2x2GB Corsair XMS3 | nVidia GeForce 8800 GTS | ATI TV Wonder Theater Pro 550 | Antec P-160 case | Antec 650w Earth Watts | LG Blu-ray Super Drive | LG DVD RW | Windows 7 Pro

  8. #8
    Join Date
    Nov 2000
    Location
    Connecticut
    Posts
    6,459
    Are FutureMark currently allowing 5x00 submissions to the ORB? I remember reading a while ago that they had stopped letting anybody with one of the dodgy driver sets post their score. Might be wrong though.
    Yeah, the latest dodgy nVidia drivers are allowed, and the guys with 3" penises are taking full advantage of it. Any respect I had for anyone on top of the ORB is now gone. They'll look for any advantage they can get to be on top.
    My Guitar Hero 2 Xbox Live Ranking
    Main Rig (click this link for specs)
    Certified in: Nothing at all!
    Certifiable: 100% hands-on, real-world experience

  9. #9
    Join Date
    Jun 1999
    Location
    Texas
    Posts
    519
    I've never been very interested much in 3DMark, though I understand that it's results can be loosely related to real gaming performance, or at least an indicator of a given card's abilities. I have to be blunt though, it's the fanatical users who have made such a big deal out of this synthetic benchmark over the years. If less (much much less) importance was placed on trying to beat everyone else and brag about it online, companies like Nvidia & ATI wouldn't be so inclined to resort to underhanded tactics. And ATI has fudged in the past IIRC, though I do believe they've got the best deal going ATM.

    It just seems so petty, really. What would honestly be impressive to me is if game developers would code to make use of current hardware efficiently, resulting in better performance for those of us who aren't interested in a new vid card every 3 months. We all know it's possible, but they know they can be sloppy now that we've all got multiple GHz CPU's and at least 512 megs of system RAM. Like I said, that would be something worth crowing about.

    One more thing about 3DMark. Howzabout everyone just ignores it from now on? Don't even install it. I've managed to be perfectly happy not running it, so evidently it's not the key to happiness.
    Last edited by slingblade; July 23rd, 2003 at 02:15 AM.

  10. #10
    Join Date
    Apr 2000
    Location
    UK
    Posts
    8,466
    Originally posted by Todd a
    Another new one that [H]ardOCP found is that the Quality setting under the desplay properties in the latest drivers use bi-liniar filtering instead of tri-liniar. They also got rid of the application setting now. More artificial bumps in performance at the users cost.
    Well using the "Quality" setting forces an adaptive filtering technique that seems to be somewhere in between the two (ATi use something similar on the 9600/Pro in "Performance" mode, BTW). What [H] are arguing is that the IQ is exactly the same as ATi's "Quality" mode, which seems to be the case from the screenshots they use but isn't in many others posted by independent users on forums. I think it's a case of them turning a blind eye yet again. People are getting banned left, right and center on their forums for suggesting such things. Mr Kyle Bennett really doesn't like people who don't agree with him and both threads relating to the article are currently locked. Somebody posted screenshots illustrating the issue they have failed to highlight and they were deleted, can you believe that?!

    These hacks affect gameplay as well, not just benchmarking. The application requests trilinear filtering and the driver is just saying "no, sorry - use this instead".

    MuFu.

  11. #11
    Join Date
    Feb 2003
    Location
    Poole, UK
    Posts
    1,999
    Jeez, sounds like a right old mess, all in all.

    I quite agree with the suggestion to ignore 3dmark, i only use it to get an idea of how the drivers are improving / evolving but with nVidia being d*cks it's clearly impossible to do that now. I would rather have an honest score than a falsely high one, I couldn't care less if it beats the 9800 or not.

    Slingblade, I also read that ATI did something similar a while back, albeit not quite as blatant, but they have apparently confessed and cleaned up their act like you say. Takes a little of the shine off their halo though.

    Anyway I definitely won't be posting my "improved" score anywhere - don't wanna be accused of having a 3-incher

    Thanks for all the viewpoints anyway, hopefully nVidia will wake up one day and release how stupid they are being.......stranger things have happened.

    Cheers.
    Windows 7 x64 Ultimate Edition, Gigabyte Z77-D3H, Ivy Bridge i7 3770, nVidia GTX480 1.5Gb, 16Gb Corsair 1600MHz, Creative X-Fi, Seagate 500Gb SATA II, DGM 23" Widescreen TFT

    "Men never do evil so completely and cheerfully as when they do it from religious conviction" -Blaise Pascal

    M.C.S.A.

  12. #12
    Join Date
    Dec 1998
    Location
    Grand Haven, Michigan, USA
    Posts
    11,332
    But the main point is it looks like there is no way to turn on the tri-liniar filtering even if you wanted to. They only have a performance mode now, basically.
    AMD Phenom II x4 945 3Ghz | ASUS M4A77TD | 2X WD 1TB SATA 2 hard drive | 2x2GB Corsair XMS3 | nVidia GeForce 8800 GTS | ATI TV Wonder Theater Pro 550 | Antec P-160 case | Antec 650w Earth Watts | LG Blu-ray Super Drive | LG DVD RW | Windows 7 Pro

  13. #13
    Join Date
    Jan 2000
    Location
    Plymouth, England
    Posts
    1,399
    So, as a GeForce user, what was the last set of legit drivers I should be using, if there was ever such a thing?!

    Bennyboy.
    Last edited by Bennyboy; July 24th, 2003 at 04:35 AM.
    If you've got your money for nothing, who cares if the chicks are free!

  14. #14
    Join Date
    Feb 2003
    Location
    Poole, UK
    Posts
    1,999

    Drivers

    In terms of 3dmark 03 scores, the original dodgy ones were 43.something (i forget exactly). The 'cheat' in these was eliminated by downloading the build 330 patch for 3dmark.

    As far as I am aware there were no new "optimisations" put in until these latest 44.67 drivers. Before that I was using 44.06 so those are probably your best bet if you want an accurate score IMO. Since build 330 I hadn't seen any increase in score until I went from 44.06 to 44.67.
    Windows 7 x64 Ultimate Edition, Gigabyte Z77-D3H, Ivy Bridge i7 3770, nVidia GTX480 1.5Gb, 16Gb Corsair 1600MHz, Creative X-Fi, Seagate 500Gb SATA II, DGM 23" Widescreen TFT

    "Men never do evil so completely and cheerfully as when they do it from religious conviction" -Blaise Pascal

    M.C.S.A.

  15. #15
    Join Date
    Apr 2003
    Posts
    219
    Screw it, I'm sending back that cursed 5600 Ultra.

    Damn you nVidia! Why must you suck so?

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •