Ripperjack
Limp Gawd
- Joined
- Mar 26, 2010
- Messages
- 299
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Yes I did, very interesting link. Let's explore it a bit.
![]()
So this is the example you claimed where AMD has "better" frame-times? Even then it's clear that the Nvidia hardware is not experiencing any sort of issue in terms of frame latency. At that high resolution, in the most VRam intensive game yet released, you would expect a card with an extra GB of VRam to start showing an advantage. Even then the two are basically dead even.
Tamlin_WSGF said:Didn't you catch the post earlier in this thread, showing one of the games that had biggest increase in 12.11 beta, which has both better framerate and frametimes with 7970 vs. the 680 according to Nordichardware's microstutter article?
Aw hell, Google has infilitrated another one and started to rewrite his internal dictionary. THE NEW WAVE OF FREE ADVERTISEMENTS IS HERE, RUN!!
This means we are talking about a conspiracy theory that are build around a driver resession in a single game? Are you being serious?![]()
David Baumann @ AMD said:Over the early part of the year you'll see a few driver updates help this across a variety of games.
I guess he never got the memo that this problem only affects one game.
You are presenting that one game as evidence and are building your whole case around driver regession in a single game.
AMD has found that the new memory management for GCN arch they have been working on could fix some of those games and even improve further on the performance.
You seem to desperately want to put words in my mouth about this being about one game, when it clearly is not.
Again, talk is cheap. I eagerly await new drivers just like everyone else. We'll see if they actually come through. One thing is for sure though, they are sure taking their time...
Its one game you have presented as evidence that there has been a driver regression, namely Skyrim.
Again, one or two bumps in the road like this wouldn't be anything to worry about, but when those blips are everywhere, they begin to matter. Subjectively, the Radeon's performance doesn't feel horrible, but if you were play back to back versus the GeForce, you'd end up picking the GTX 660 Ti as the better option.
the 7950 struggles more in AC3 in Windows 8. Then again, the 7950 doesn't perform especially well in this game in either version of Windowsnot compared to the GTX 660 Ti.
As in Win8, the 7950 posts a higher FPS average than the GeForce, but a small number of high-latency frames detracts from the 7950's performance.
We've already seen terrible frame latency issues in both Skyrim and Borderlands2. Of course those two and BF3 were the only three games that nordichardware even tested... and 2/3 of them had serious issues. Who's to say what they would come up with if they tested more games?
TechReport found similar issues on many different games. Skyrim and Borderlands2 of course, but also several others.
One game?
You are referring to games that has microstutter. We have no dispute there. I even want you to stop with the ridiculous thin conspiracy theory, because its mudding the whole discussion about microstutter. On the last pages, we have derailed the thread discussing your theory and even though you've had ample chance, you haven't provided any solid evidence to support such a claim, so its been a waste of time.
Your claim is that the performance improvements were at the cost of microstutter, where you claim AMD deliberately sacrificed smoothness to get higher framerates before holiday sales.
This you base on a driver regression on a single game: Skyrim. Skyrim worked earlier smoothly, but in the 12.11 beta, it had microstutter. This is the single evidence you have presented towards your conspiracy theory.
I thought I'd write some more, but I really don't have the time to waste on your conspiracy theory anymore, so unless you actually have something substancial, I'm done with it.
when will nvidia fix this...![]()
You are referring to games that has microstutter. We have no dispute there. I even want you to stop with the ridiculous thin conspiracy theory, because its mudding the whole discussion about microstutter. On the last pages, we have derailed the thread discussing your theory and even though you've had ample chance, you haven't provided any solid evidence to support such a claim, so its been a waste of time.
Your claim is that the performance improvements were at the cost of microstutter, where you claim AMD deliberately sacrificed smoothness to get higher framerates before holiday sales.
This you base on a driver regression on a single game: Skyrim. Skyrim worked earlier smoothly, but in the 12.11 beta, it had microstutter. This is the single evidence you have presented towards your conspiracy theory.
I thought I'd write some more, but I really don't have the time to waste on your conspiracy theory anymore, so unless you actually have something substancial, I'm done with it.
Don't worry, GotNoRice is wrong. He is just being a crazy nVidiot at this point. It's pretty obvious to most people that the microstutter was an unintended consequence of something they did in these drivers.
Additionally, when we switched from the old VLIW architecture to the GCN core there was a significant updates to all parts of the driver was needed – although not really spoken about the entire memory management on GCN is different to prior GPU's and the initial software management for that was primarily driven by schedule and in the meantime we've been rewriting it again and we have discovered that the new version has also improved frame latency in a number of cases so we are accelerating the QA and implementation of that.
There was the same stutter in Skyrim before the 12.11 beta driver. It wasn't introduced with it.
Agreed but unfortunately the nature of PC games/hardware means AMD or Nvidia cannot guarantee every game will work perfectly. That is why both AMD and Nivida issue new driver releases regularly. It is also the reason why it is possible to create your own game profiles using the built in GPU control panel or 3rd party software such as RadeonPro or Nvidiainspector. Practically every one of us use 3rd party tools to overclock our cards and think nothing of it. I gaurantee there is not a single Nvidia GPU owner out there who has never had to create a custom profile to get the best out of some game or another. I know I have to make custom game profiles to get the best out of a lot of my games for my GTX 680. Witcher 2 is unplayable without an FPS cap for example.
[H] has been bitching for years about multi GPU Crossfire stuttering, not single GPU. For the vast majority this latency issue has never been an issue. My own testing in Skyrim showed a simple tweak to the flip queue size cured all stuttering.
http://forums.anandtech.com/showthread.php?t=2289641&highlight=
Hitman Absolution Frame Times
http://forums.anandtech.com/showpost.php?p=34376473&postcount=550
Agreed.
Not a problem. Just limit fps to 59 or 60
*adjusting_ma_tin_foil_hat_and_googles*
I remember an OpenGL on Linux performance review on Linux from a couple of years ago that tracked AMD "performance" increases/decreases. Performance increases went hand-in-hand with bugs. Fixing bugs again decreased performance. Hopefully this won't occur again.
You put your tin_foil_hat on backwards. You should have said "AMD increased performance knowing it was causing bugs in Linux, so they could take over the world!"!Now thats a juicy conspiracy theory!
![]()
In Linux however, bugs go hand-in-hand with bugs when it comes to drivers. That goes for Nvidia too. The performance review on Linux you saw years ago isn't something I can comment on as long as you don't link up to the review.
The fix spoken of here in this thread, is about rewriting the memory management to suit the GCM architecture better. AMD/Dave Baumann implied that we will see performance improvements as result of this in games that has been optimized.
As you can see from the Nordichardware microstutter test, Battlefield 3 doesn't show any microstuttering and is infact both marginally better in frametimes AND framerate on 7970 vs. 680. That was one of the games that had the largest increase in framerate in 12.11.
That said, I care less about getting less FPS then getting smoother gameplay. This is one of the reasons I actually READ [H]'s reviews. Their reviews are about highest playable settings and they nudge down settings if they feel gameplay isn't smooth enough, even when framerate should suggest they could use a higher settings.
Except for FPS gaming where higher framerate can help response time, I only need higher framerates to get smoother gameplay.
Thats why kudos to Techreport for making a buzz about this and also kudos to [H] for their reviews which focus on gameplay and not only pure framerates.
Don't put a marketing spin on this with conspiracy theories and all, turning it into an AMD vs. Nvidia war like they have on Anandtech forums. That will only mud the issue.
PCPER is now also going over to a new performance metric that focus on smoothness:
http://www.pcper.com/reviews/Graphics-Cards/Frame-Rating-New-Graphics-Performance-Metric
Vsync should also be enabled and 59fps cap too. In that case I get no stutter at all. Fallout 3 is like a totally different game
They released betas that increased the performance of GCN hardware at the expense of smoothness and did so conveniently right before the holiday shopping season where they sell more cards than any other time of year by far; knowing that people make purchases based on benchmarks.
No WHQL drivers since October... pretty clear that the betas would NOT pass WHQL certification, and it's pretty clear why.
Now it's January. The Holiday shopping season is over. Like clockwork they now "acknowledge" the problem.
The final test will be to see if they can actually come out with a set of WHQL drivers that somehow retain the performance increase of the betas while also eliminating the frame latency issue. I personally don't believe this is something we will ever see.
I still have not see whether the stuttering is resolved or changed in any way, but to say that the drivers were not WHQL certified because of the stuttering is just wrong,