Page 34 of 47 FirstFirst ...
24
32
33
34
35
36
44
... LastLast
  1. #661
    I am Murloc! Cyanotical's Avatar
    Join Date
    Feb 2011
    Location
    Aurora, Co
    Posts
    5,103
    im not sure you can make a comparison between a 660ti and a 7950, overclocked or otherwise

    you can definitely compare say a 7970 vs 7970 Matrix or 680 vs 680 classified

    but the technology difference between AMD and Nvidia is enough that simply comparing framerates is not enough anymore, 100fps on an nvidia card is not the same as 100fps on an AMD card

    i7-4790K | Z97 Class. | 8GB DDR3-2133 | GTX-690 Quad SLI | RAIDR | 512GB Samsung 830 | AX1200 | FT02
    Dell U2711 | Ducky Shine3 YoS | Steelseries Sensei Wireless | Xonar Essence One | KRK RP8G2s

  2. #662
    TOTALLY NOT
    Banned
    tetrisGOAT's Avatar
    Join Date
    Jun 2008
    Location
    Sweden
    Posts
    12,561
    Quote Originally Posted by Cyanotical View Post
    you can definitely compare say a 7970 vs 7970 Matrix or 680 vs 680 classified
    Yeah, that was what I was trying to say in the first post.

    Quote Originally Posted by Cyanotical View Post
    but the technology difference between AMD and Nvidia is enough that simply comparing framerates is not enough anymore, 100fps on an nvidia card is not the same as 100fps on an AMD card
    AMD has said to address this in the next driver though.

  3. #663
    Quote Originally Posted by tetrisGOAT View Post
    Yeah, that was what I was trying to say in the first post.

    AMD has said to address this in the next driver though.
    Is that the 13.1 that was released a few days ago as beta or the next one?
    Intel i5-3570K @ 4.7GHz | MSI Z77 Mpower | Noctua NH-D14 | Corsair Vengeance LP White 1.35V 8GB 1600MHz
    Gigabyte GTX 670 OC Windforce 3X @ 1372/7604MHz | Corsair Force GT 120GB | Silverstone Fortress FT02 | Corsair VX450

  4. #664
    I am Murloc! Cyanotical's Avatar
    Join Date
    Feb 2011
    Location
    Aurora, Co
    Posts
    5,103
    Quote Originally Posted by tetrisGOAT View Post

    AMD has said to address this in the next driver though.
    maybe, but afaik the overall smoother operation of the 600 series vs the 7000 series is hardware related

    i7-4790K | Z97 Class. | 8GB DDR3-2133 | GTX-690 Quad SLI | RAIDR | 512GB Samsung 830 | AX1200 | FT02
    Dell U2711 | Ducky Shine3 YoS | Steelseries Sensei Wireless | Xonar Essence One | KRK RP8G2s

  5. #665
    TOTALLY NOT
    Banned
    tetrisGOAT's Avatar
    Join Date
    Jun 2008
    Location
    Sweden
    Posts
    12,561
    Unsure. Basically, what they said was, the fluctuating frame differentials are not a direct result of 'cheating' to get the performance gains they got, but unrelated and will be fixed in an upcoming driver.
    Quote Originally Posted by Cyanotical View Post
    maybe, but afaik the overall smoother operation of the 600 series vs the 7000 series is hardware related
    I'm blaming drivers and coders. After all, they don't have the money oomph like other companies in the business, so it's easy to see they haven't prioritised frame differentials in Crossfire configurations for a while (although here I suspect it's more cheating than in single-card configs, it can definitely improve without sacrificing performance).
    Single card configuration got a side effect from something that slipped by them. That's what I want to believe, at least, and take their word for it.
    Last edited by tetrisGOAT; 2013-01-16 at 10:39 PM.

  6. #666
    I am Murloc! Cyanotical's Avatar
    Join Date
    Feb 2011
    Location
    Aurora, Co
    Posts
    5,103
    well crossfire is something they should work on more, it's just too much trouble compared to SLI, quad 7970s beats the crap out of quad 680s in benchmarks, but the practical use is terrible, meanwhile i rarely have SLI issues, even though im running quad


    honestly they need to apply the active bridging from crossfirepro to the consumer cards, it would put nvidia back on the defensive again, as well as bring game changing abilities to both computer building and multiscreen gaming

    i7-4790K | Z97 Class. | 8GB DDR3-2133 | GTX-690 Quad SLI | RAIDR | 512GB Samsung 830 | AX1200 | FT02
    Dell U2711 | Ducky Shine3 YoS | Steelseries Sensei Wireless | Xonar Essence One | KRK RP8G2s

  7. #667
    I am Murloc! Xuvial's Avatar
    Join Date
    Apr 2010
    Location
    New Zealand
    Posts
    5,031
    You guys hear that measuring frame latencies is most likely going to overtake measuring framerates when it comes to benchmarking? Apparently frame latencies can show stuff like stutter/microstutter, a phenomenon which completely flies over stuff like Fraps.

    I got my first solid taste of stutter with 310.70 drivers in BF3, framerate indicator is telling me a constant 90-110fps but it FELT like goddamn 40 fps, so choppy. 310.90 fixed that for me though :P

  8. #668
    Is the VE228TR notably worse than the H version? It's the only one I can find, and the other suggestions are out of my budget.

  9. #669
    Brewmaster Majesticii's Avatar
    Join Date
    Feb 2010
    Location
    Netherlands
    Posts
    1,413
    Quote Originally Posted by Xuvial View Post
    You guys hear that measuring frame latencies is most likely going to overtake measuring framerates when it comes to benchmarking? Apparently frame latencies can show stuff like stutter/microstutter, a phenomenon which completely flies over stuff like Fraps.

    I got my first solid taste of stutter with 310.70 drivers in BF3, framerate indicator is telling me a constant 90-110fps but it FELT like goddamn 40 fps, so choppy. 310.90 fixed that for me though :P
    I sure hope so. Average FPS just barely says anything. Like you said, avg fps can be good enough, but you can still have poor texture streaming or any other form of popins/stutter.

  10. #670
    Are there any good tools for measuring this sort of latency? I'd very seriously consider the investment if it would help provide more data.

  11. #671
    Brewmaster Majesticii's Avatar
    Join Date
    Feb 2010
    Location
    Netherlands
    Posts
    1,413
    I believe the latest version of Afterburner has this. But i don't know how reliable it is.

  12. #672
    TOTALLY NOT
    Banned
    tetrisGOAT's Avatar
    Join Date
    Jun 2008
    Location
    Sweden
    Posts
    12,561
    Yeah, it does. It seems to work, at least. I have a GeForce card so I can try it in a common game if anyone has a Radeon card.
    However, AMD is very close to have fixed the latencies according to what I've heard. Not better than the nVidia cards, but so close the difference is negligble. (according to Sweclockers)

  13. #673
    I think i have made a log with afterburner, but i have no idea how to actually read the log. It said you need rivatuner, but that didnt really help :P

    nvm, its open xD

    So, i tried it with 5 minutes of BF3, at ultra settings, no MSAA @ 1440p. Running on a sapphire 7970 @ 1000/1450

    Seems im getting a pretty steady 60 fps, which is nice
    Frametime seems to be around 15-20 ms , with some shoot outs to 40-50 ms (no idea what that means, but im not noticing it during gameplay, could be when i died or something like that)
    Last edited by Zeara; 2013-01-18 at 08:16 PM.

  14. #674
    Techreport has benchmarked the 7950 with the 13.2 beta drivers in 3 games and it seems that the changes have solved the frame latency issues. AMD said that they are working on a more general performance fix that should apply to most games.
    http://techreport.com/review/24218/a...on-frame-times

  15. #675
    Quote Originally Posted by stalkerzzzz View Post
    Techreport has benchmarked the 7950 with the 13.2 beta drivers in 3 games and it seems that the changes have solved the frame latency issues. AMD said that they are working on a more general performance fix that should apply to most games.
    http://techreport.com/review/24218/a...on-frame-times
    It's solved the latency issues in three specific DX9 games. AMD response said they're working on generic DX10/DX11 fix and will have it out in few weeks, but considering the had to manually tweak DX9 drivers for each game individually, I wouldn't hope for Q1 release of the patch.
    Never going to log into this garbage forum again as long as calling obvious troll obvious troll is the easiest way to get banned.
    Trolling should be.

  16. #676
    Quote Originally Posted by vesseblah View Post
    It's solved the latency issues in three specific DX9 games. AMD response said they're working on generic DX10/DX11 fix and will have it out in few weeks, but considering the had to manually tweak DX9 drivers for each game individually, I wouldn't hope for Q1 release of the patch.
    To be fair, 'exception' drivers aren't exactly uncommon. nVidia do drivers that are heavily focused around a single game's performance increase quite often.

    That said, there's very little means of figuring out the full update will come, but I would certainly hope it be within this quarter.

  17. #677
    Quote Originally Posted by Drunkenvalley View Post
    To be fair, 'exception' drivers aren't exactly uncommon. nVidia do drivers that are heavily focused around a single game's performance increase quite often.

    That said, there's very little means of figuring out the full update will come, but I would certainly hope it be within this quarter.
    Aren't Nvidia usually doing this with newly released or coming games though?
    Intel i5-3570K @ 4.7GHz | MSI Z77 Mpower | Noctua NH-D14 | Corsair Vengeance LP White 1.35V 8GB 1600MHz
    Gigabyte GTX 670 OC Windforce 3X @ 1372/7604MHz | Corsair Force GT 120GB | Silverstone Fortress FT02 | Corsair VX450

  18. #678
    Quote Originally Posted by Drunkenvalley View Post
    To be fair, 'exception' drivers aren't exactly uncommon. nVidia do drivers that are heavily focused around a single game's performance increase quite often.
    Yeah, but this time they aren't making exception drivers. They're making individual tweaks on individual games. Sure it could be just beta-testing their fixes across multiple different types of 3D engines, but the pessimist in me says it smells like they can't do broad fix that covers all games without reflashing the card or until next gen chips.

    Quote Originally Posted by n0cturnal View Post
    Aren't Nvidia usually doing this with newly released or coming games though?
    Both companies do that, but it's usually for SLI/Crossfire performance or on totally new 3D engines not used anywhere else. They don't do it for yet another 3D shooter using the same old Unreal3 engine which has already been tweaked to max.
    Never going to log into this garbage forum again as long as calling obvious troll obvious troll is the easiest way to get banned.
    Trolling should be.

  19. #679
    Quote Originally Posted by vesseblah View Post
    Yeah, but this time they aren't making exception drivers. They're making individual tweaks on individual games.
    I'm not sure how that's different really.

    Quote Originally Posted by vesseblah View Post
    Sure it could be just beta-testing their fixes across multiple different types of 3D engines, but the pessimist in me says it smells like they can't do broad fix that covers all games without reflashing the card or until next gen chips.
    I won't claim to know the amount of work necessary, or what amount of change will prompt a reflash of the card, etc. I do sincerely hope AMD will deliver on time though, as I'm not one for overdue projects.

  20. #680
    Quote Originally Posted by Drunkenvalley View Post
    I'm not sure how that's different really.
    AMD cards underperform in frame latency on ALL games right now, therefore they need to figure out a fix for ALL games. They can't do it on every single game one by one. On normal driver updates both AMD and Nvidia upgrade and tweak their drivers for ONE underperforming game (like recently AMD got 50% or more speed into Skyrim for example, going from 20% below Nvidia to 20% above in that specific title).

    Remember GTX480 got massive speed increase across the board few months after launch with driver update. That's what AMD needs to come up with now, not tweak few DX9 titles. And that's why I'm concerned with it.

    That's how it's different.
    Never going to log into this garbage forum again as long as calling obvious troll obvious troll is the easiest way to get banned.
    Trolling should be.

Posting Permissions

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