This is a split board - You can return to the Split List for other boards.

Been outta the loop for a bit, what's the status on microstutter and AMD?

#31DarkZV2BetaPosted 8/5/2013 10:35:59 AM(edited)
Snuckie7 posted...
DarkZV2Beta posted...
Snuckie7 posted...
DarkZV2Beta posted...
Snuckie7 posted...
Since when did "standards" constitute being able to notice the difference between say, a 4ms frame variance and a 2ms one?

No, this is just you being pedantic and arguing over differences that nobody will realistically be able to notice.


If by "nobody", you mean "nobody with no standards".

It's actually pretty damn noticeable in a lot of games, even with small variance.


A 2ms wider frame variance is "damn noticeable"? To you and who else?


No, that's just you downplaying the actual variance.
Never mind that this "fix" comes with it's own problems.

The actual variance is just about that though. If you claim such a difference is so noticeable, why don't I hear you complaining about SLI dropping more runt frames than CF? Surely gaps in the animation are far more noticeable than minute differences in frame pacing.


Except that it's not. Seeing upwards of 5ms variance in charts of largely <10 ms frames.
Never mind that it only works for DX10/11, and only in single display <2460x1440 resolutions. Hardly a "thing of the past".
And average performance is lowered slightly by the fix.
And it attempts to lock the framerate, resulting in consistent frame time variance from minor framerate variation.
---
Want that Shield!
Ball and Cup on ps mobile has framerate issues. -stargazer64
#32ein311Posted 8/5/2013 10:32:04 AM
fire2box posted...
the people who complain about microsuttering are the same people who don't play video games and use their 700-2000 dollar pcs to only post on gamefaqs.


:D
---
Man your own jackhammer
Man your battlestations
#33Snuckie7Posted 8/5/2013 10:40:12 AM
DarkZV2Beta posted...
Snuckie7 posted...
The actual variance is just about that though. If you claim such a difference is so noticeable, why don't I hear you complaining about SLI dropping more runt frames than CF? Surely gaps in the animation are far more noticeable than minute differences in frame pacing.


Except that it's not. Seeing upwards of 5ms variance in charts of largely <10 ms frames.
Never mind that it only works for DX10/11, and only in single display <2460x1440 resolutions. Hardly a "thing of the past".


So runt frames aren't that big of a big deal when Nvidia is dropping them now?

I'm seeing that CF has slightly higher frame variance than SLI in nearly every game. At the higher percentiles where the frame variance may actually approach say 9ms in a chart of 20ms variance, SLI doesn't do much better either.

http://www.pcper.com/files/review/2013-07-31/Crysis3_1920x1080_STUT.png

etc.
---
Intel Core i7 3820 | EVGA X79 SLI K2 | MSI 7950 Twin Frozr III | Samsung / 840 120GB / 8GB RAM | 1TB WD Caviar Blue | Corsair / 550D / H70 | Silencer MKIII 600W
#34DarkZV2BetaPosted 8/5/2013 10:45:57 AM
Snuckie7 posted...
DarkZV2Beta posted...
Snuckie7 posted...
The actual variance is just about that though. If you claim such a difference is so noticeable, why don't I hear you complaining about SLI dropping more runt frames than CF? Surely gaps in the animation are far more noticeable than minute differences in frame pacing.


Except that it's not. Seeing upwards of 5ms variance in charts of largely <10 ms frames.
Never mind that it only works for DX10/11, and only in single display <2460x1440 resolutions. Hardly a "thing of the past".


So runt frames aren't that big of a big deal when Nvidia is dropping them now?

I'm seeing that CF has slightly higher frame variance than SLI in nearly every game. At the higher percentiles where the frame variance may actually approach say 9ms in a chart of 20ms variance, SLI doesn't do much better either.

http://www.pcper.com/files/review/2013-07-31/Crysis3_1920x1080_STUT.png

etc.


http://www.guru3d.com/articles_pages/amd_framepacing_review_with_catalyst_13_8_fcat,9.html

ect.

No one's saying it's not an improvement, but the fact of the matter remains: SLI is still better, and stutter is not a "thing of the past", and never will be for any multi GPU setup, ever. Especially one that's still completely broken for DirectX9, multi-display setups, and extreme resolutions.
---
Want that Shield!
Ball and Cup on ps mobile has framerate issues. -stargazer64
#35Snuckie7Posted 8/5/2013 10:51:48 AM
DarkZV2Beta posted...
Snuckie7 posted...
DarkZV2Beta posted...
Snuckie7 posted...
The actual variance is just about that though. If you claim such a difference is so noticeable, why don't I hear you complaining about SLI dropping more runt frames than CF? Surely gaps in the animation are far more noticeable than minute differences in frame pacing.


Except that it's not. Seeing upwards of 5ms variance in charts of largely <10 ms frames.
Never mind that it only works for DX10/11, and only in single display <2460x1440 resolutions. Hardly a "thing of the past".


So runt frames aren't that big of a big deal when Nvidia is dropping them now?

I'm seeing that CF has slightly higher frame variance than SLI in nearly every game. At the higher percentiles where the frame variance may actually approach say 9ms in a chart of 20ms variance, SLI doesn't do much better either.

http://www.pcper.com/files/review/2013-07-31/Crysis3_1920x1080_STUT.png

etc.


http://www.guru3d.com/articles_pages/amd_framepacing_review_with_catalyst_13_8_fcat,9.html

ect.

No one's saying it's not an improvement, but the fact of the matter remains: SLI is still better, and stutter is not a "thing of the past", and never will be for any multi GPU setup, ever. Especially one that's still completely broken for DirectX9, multi-display setups, and extreme resolutions.


Yeah I edited my first post once I realized that you were actually complaining about frame variance being inherent to multi card setups, even if it was quite unnoticeable to 99% of the population.

SLI, once again is not better than CF to a degree where people not looking for a difference will notice, and the fix for >2560x1600 and DX9 is coming later this month.
---
Intel Core i7 3820 | EVGA X79 SLI K2 | MSI 7950 Twin Frozr III | Samsung / 840 120GB / 8GB RAM | 1TB WD Caviar Blue | Corsair / 550D / H70 | Silencer MKIII 600W
#36DarkZV2BetaPosted 8/5/2013 10:59:46 AM
Statements like "no one will notice" are always false, and not any different than console peasants saying more than 30 or 60fps is unnoticeable. Not everyone will notice, but there will be some. Some people still think N64 had no framerate problems.
That kind of mentality just leads to substandard products.

For all intents and purposes, SLI is still better than crossfire. There's no point in picking the inferior product here. Unless you already have a Radeon and are looking for more performance or just really really hate green and don't want anything loosely associated with it in your PC, CFX is still a less desirable choice.
---
Want that Shield!
Ball and Cup on ps mobile has framerate issues. -stargazer64
#37Snuckie7Posted 8/5/2013 11:08:38 AM
I think when we're discussing frametime variance differences that are mostly under 5ms, it's quite safe to say, that it's a very, minute difference that's not applicable to someone not actively looking for the difference, which is not the case with 30fps vs 60fps. Not to mention that SLI drops more runt frames, but then again, since Nvidia's the one doing it it's probably okay.

There's no point in picking the inferior product here.

Gee, I had no idea that the evenness of frame pacing is the only attribute these products had. Many Radeons are superior to similarly priced GeForces in a number of ways, if I haven't made that clear in the past.
---
Intel Core i7 3820 | EVGA X79 SLI K2 | MSI 7950 Twin Frozr III | Samsung / 840 120GB / 8GB RAM | 1TB WD Caviar Blue | Corsair / 550D / H70 | Silencer MKIII 600W
#38DarkZV2BetaPosted 8/5/2013 11:21:56 AM
And, again, those kinds of blanket statements are always false. Every time. Without fail.

Radeon cards are being sold at a pretty steep discount right now, but the MSRP isn't too different, and there are other advantages to GeForce. However, speaking purely of Crossfire vs SLI, yes, SLI is better.
7990 really aught to be $800-900 msrp at this point.
---
Want that Shield!
Ball and Cup on ps mobile has framerate issues. -stargazer64
#39Snuckie7Posted 8/5/2013 11:26:10 AM(edited)
Ah yes, I forgot the "for 99% of users" part.

As for CF vs. SLI, slightly higher frame variance vs. runt frames. Take your pick.

Edit: http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&DEPA=0&Order=BESTMATCH&Description=7990&N=-1&isNodeId=1

I don't know where you've been looking, but 7990's are $800-900. Never mind MSRP, as there are more discounted 7990's than apparently fully priced ones.
---
Intel Core i7 3820 | EVGA X79 SLI K2 | MSI 7950 Twin Frozr III | Samsung / 840 120GB / 8GB RAM | 1TB WD Caviar Blue | Corsair / 550D / H70 | Silencer MKIII 600W
#40DarkZV2BetaPosted 8/5/2013 11:30:15 AM
You shouldn't need to bank on a discount, especially one reliant on a mail-in rebate.
$820 for that MSI one is a nice deal, though.
---
Want that Shield!
Ball and Cup on ps mobile has framerate issues. -stargazer64