![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
What are Kyle and Brent's opinions on Tech Report's GTX 660 Ti vs HD 7950 findings?
- Thread starter burningrave101
- Start date
wonderfield
Supreme [H]ardness
- Joined
- Dec 11, 2011
- Messages
- 7,396
I don't understand what you're asking for. You do not believe that a single-GPU solution will not exhibit zero inter-frame latency in games?
If what's most important to you is how high the frame counter in the corner of your display is occasionally reading, you probably need to re-align your perceptions about what's important for real-time rendering vis-à-vis gaming.
Why?
Even single card vs. single card I found the GTX 680 4GB cards to be smoother than the Radeon HD 7970's. I didn't expect it when I made the switch. I was just hoping to get my multiGPU setup working. Of course now months later AMD's resolved the lockup issues I had on the X79 and X58 platforms.
BababooeyHTJ
Supreme [H]ardness
- Joined
- Jan 21, 2009
- Messages
- 6,951
Again, you're talking about multi-display. I'm pretty sensitive to microstutter and always seem to be one of the few people who actually seem to notice it on the forums and even I haven't seen a real difference with a single card on a single display.
You also had some serious driver issues that I didn't. Could be a compatibility issue too. I didn't find Kepler to be the smoothest experience in the world. I found it to be a step back from Fermi. Does that mean that what I saw was going to be the case for everyone?
Thats exactly why I sold my 4870x2 and kept my GTX280. I haven't found microstutter to be a major issue this gen. Even Cayman was pretty smooth for me.
- Joined
- May 18, 1997
- Messages
- 55,634
There is a reason that we led the industry in moving away from "canned" benchmarks with highest FPS being the crowning factor.
We have known about frame times for a long time and we have not found an efficient and effective way to capture PRECISE data. I would suggest that current methods are flawed and that is why we are not publishing objective numbers. But we have been talking about this subjectively for years. And honestly, I am not sure we ever will publish frame time numbers. But we will continue to make subjective statements about the gaming experience that the hardware provides.
We have known about frame times for a long time and we have not found an efficient and effective way to capture PRECISE data. I would suggest that current methods are flawed and that is why we are not publishing objective numbers. But we have been talking about this subjectively for years. And honestly, I am not sure we ever will publish frame time numbers. But we will continue to make subjective statements about the gaming experience that the hardware provides.
GoldenTiger
Fully [H]
- Joined
- Dec 2, 2004
- Messages
- 29,669
I noticed the uneven frame times easily when I owned 79xx cards (radeon) versus switching to the nvidia 670 and 680 cards single card to single card. The nvidia setups were smoother and more consistent. The radeon setups felt juddery even when the fps count showed it was supposedly similar. High speed videos and frame time #s only show this more clearly than subjective observations like mine.
BababooeyHTJ
Supreme [H]ardness
- Joined
- Jan 21, 2009
- Messages
- 6,951
According to the early techreport GTX680 and 670 reviews 7970 and 7950 was actually smoother.
I didn't notice any difference in smoothness with a single card. I have noticed microstutter more often in crossfire but its not that often and generally not consistent but I don't use eyefinity.
I also find it odd that they never made a peep about the vsync stutter issues that some people still suffer with.
GoldenTiger
Fully [H]
- Joined
- Dec 2, 2004
- Messages
- 29,669
How on Earth did you get that incorrect conclusion out of this article: http://techreport.com/review/24051/geforce-versus-radeon-captured-on-high-speed-video ? It very clearly, both through graphs, written observations, and high-speed video capture, shows the GeForce cards are far smoother even at about the same framerates, overall.
wabbitseason
[H]ard|Gawd
- Joined
- Jun 16, 2010
- Messages
- 1,511
How on earth did you misunderstand the post you were quoting. Earlier reviews. EARLIER REVIEWS.
This isn't some massive problem that's only now coming to light. You haven't been preaching the truth, and you aren't suddenly vindicated. It didn't exist as recently as a few months ago. It may not even exist now.
GoldenTiger
Fully [H]
- Joined
- Dec 2, 2004
- Messages
- 29,669
No one mentioned "earlier reviews" where I was responding. What are you talking about?
Indeed this isn't some massive thing only JUST NOW coming to light, it's been known for quite a number of years amongst the more technical enthusiast communities of computers and technology.
I'll just quote myself for you from the other thread you posted something identical in:
GoldenTiger said:No, I'd have you believe they have reported frametime info for years and AMD has almost always been on the losing end, which is completely true and can be looked at on TechReport's site for yourself. This current article is about the 79xx series Radeons (not older ones), and clearly they are even worse than prior generations. I'm sure if they had concluded the Radeons were superior, you wouldn't be saying the only two possible realities were that they "messed up" or "a bug".Other sites and people have carried out frametime checks over the years as well, this is not some radical new theory but a hardened, reality-proven method of testing cards in conjunction with framerates.
wabbitseason
[H]ard|Gawd
- Joined
- Jun 16, 2010
- Messages
- 1,511
I'll just quote you too. Then you can provide some proof of your allegations besides Techreport's NEW results. You're saying that AMD frame-time latency has historically been higher than Nvidia's. Further, you state that the 7xxx series is "clearly even worse than prior generations" in terms of frame-time latency.
In the meantime, here's five examples that directly contradict you:
http://techreport.com/r.x/geforce-gtx-680/arkham-comparo.gif
http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-2.gif
http://techreport.com/r.x/geforce-gtx-680/skyrim-comparo.gif
http://techreport.com/r.x/geforce-gtx-680/crysis2-comparo.gif
http://techreport.com/r.x/geforce-gtx-680/sam3-comparo.gif
Also,
Last edited by a moderator:
Brent_Justice
Moderator
- Joined
- Apr 17, 2000
- Messages
- 17,755
I did wish we had another way to visually show performance, or experience, other than fps, I think fps are quite misleading to the overall experience, we've been doing things the way we do them for so long because we figured this out. We are in an interesting time of the gaming/graphics world, it is no longer about the ability to generate framerates the fastest as possible, like it use to be, it is now about the experience the graphics cards deliver in a game, visually, and in other metrics that are not the framerate. Interesting times.
Though, no matter how much evolves, it always comes directly back to the fact that we evaluate video cards just like you, the gamer, we take the card, install drivers, fire up the game, and just play the game and find out what the card can and cannot do to provide a good gameplay experience. In the end, real-world evaluation will always beat out pure fps measurement. Lo and behold, subjective feedback matters.
Though, no matter how much evolves, it always comes directly back to the fact that we evaluate video cards just like you, the gamer, we take the card, install drivers, fire up the game, and just play the game and find out what the card can and cannot do to provide a good gameplay experience. In the end, real-world evaluation will always beat out pure fps measurement. Lo and behold, subjective feedback matters.
wabbitseason
[H]ard|Gawd
- Joined
- Jun 16, 2010
- Messages
- 1,511
I am confident in [H]'s independence. You guys are no-nonsense, and that's why we're all here. If, as you say, times have changed, the [H] reviews might benefit from the kind of data that TR seems to be providing. Intangibles are easy to misinterpret. Facts stand on their own.
BababooeyHTJ
Supreme [H]ardness
- Joined
- Jan 21, 2009
- Messages
- 6,951
I like the way that [H] does things. Pure fps results with a graph and most importantly their subjective opinion. I think that more sites should do that. It always strikes me as odd that sites like techpowerup, guru 3d, or anand test a shit load of games and seem to never run into a single issue. Yeah, sure.
I've said a couple of times in this thread that I agree that microstutter tends to be more of an issue with AMD than Nvidia. I've noticed it more. I haven't found it to be a game breaker though.
I will also say this. I find it odd that all of a sudden techreport is showing more inconsistent frametimes with a single card than they had in the past with crossfire. Don't you think that some people on the forums might have noticed that?
I've said a couple of times in this thread that I agree that microstutter tends to be more of an issue with AMD than Nvidia. I've noticed it more. I haven't found it to be a game breaker though.
I will also say this. I find it odd that all of a sudden techreport is showing more inconsistent frametimes with a single card than they had in the past with crossfire. Don't you think that some people on the forums might have noticed that?
Brent_Justice
Moderator
- Joined
- Apr 17, 2000
- Messages
- 17,755
In the past (as we like to call the 3dfx days) fps meant everything, you needed a lot because graphics cards back then were having trouble even with 1024x768 or 1280x1024 and ouch 1600x1200 resolutions. But now, GPUs are so powerful they are giving us plenty of fps, and now its about taking that fps and doing something constructive with it to improve the gameplay experience.
To most people, 60hz/fps is sufficient (on 60hz display), so if you get 80 fps with one card, and 90 fps with another, it doesn't matter. Most people play with vsync enabled, if you lock the fps to the refresh rate, then what matters is what that extra 20-30fps is going to do for you in games, like enable higher in-game graphics settings. That is the benchmark now, the graphics settings, rather than the highest fps, like it use to be.
The big variable in gaming, to me, now seems to be the display. What resolution are you gaming at, what refresh rate is the display, are you locking vsync to the refresh rate, all of these things are big determinants now in what type of gameplay experience you'll have with each card. There's a lot more variety now.
To most people, 60hz/fps is sufficient (on 60hz display), so if you get 80 fps with one card, and 90 fps with another, it doesn't matter. Most people play with vsync enabled, if you lock the fps to the refresh rate, then what matters is what that extra 20-30fps is going to do for you in games, like enable higher in-game graphics settings. That is the benchmark now, the graphics settings, rather than the highest fps, like it use to be.
The big variable in gaming, to me, now seems to be the display. What resolution are you gaming at, what refresh rate is the display, are you locking vsync to the refresh rate, all of these things are big determinants now in what type of gameplay experience you'll have with each card. There's a lot more variety now.
Brackle
Old Timer
- Joined
- Jun 19, 2003
- Messages
- 8,567
Then why didn't Techreport mention them in the earlier reviews. If anything TR's earlier reviews show Nvidia having more stuttering problems then AMD.
This post is SPOT ON: http://forums.anandtech.com/showpost.php?p=34374276&postcount=545
KazeoHin
[H]F Junkie
- Joined
- Sep 7, 2011
- Messages
- 9,003
I build gaming machines for a living, and I own an AMD 6950. I can honestly say the 660ti is a MUCH smoother experience than 7870 or 7950. Runing Unigine, my favorite benchmark, the 660ti has a very small variance on frame latencies. On the Radeon cards, the latencies are all over the place, easily fluctuating 200% between frames. The way I explain the smoothness to my customers is with drumbeats, impossible to explain with text, instead I'll explain with dashes and "o"s. A dash represents a millisecond of time, and an "o" is when the video card pumps out a rendered frame:
Nvidia:
---o---o---o---o---o--o---o---o---o----o---o---o
AMD:
---o-o--o------o--o--o-o-----o--o----o--o------o
I only notice this because I build and test so many new gaming PCs in a week, but the difference is very much there.
Nvidia:
---o---o---o---o---o--o---o---o---o----o---o---o
AMD:
---o-o--o------o--o--o-o-----o--o----o--o------o
I only notice this because I build and test so many new gaming PCs in a week, but the difference is very much there.
Riddleofsteel
[H]ard|Gawd
- Joined
- Feb 4, 2007
- Messages
- 1,983
Ugh, why do I need to see this just as I'm about to open my Gigabyte 7950 that I've been waiting for a free day to install? All I really care about is smoothness with the best eye-candy possible. Should I be considering waiting for an Nvidia card? This thread is a jumble already.
A lot of this is based on techreports latest frametime measurements.
According to their own measurements there was no such problem in August.
http://techreport.com/review/23419/nvidia-geforce-gtx-660-ti-graphics-card-reviewed/4
The jury is still out and a there are a lot of unanswered questions about what exactly is going on.
Either way if you look at their August review there's nothing major demonstrated. I'm not going to speculate further except to say that more information is needed.
Personally I wouldn't return it because if the supposed issue didn't exist in August there has to be a reasonable explanation and if nothing else use the drivers from then if you experience any issues. In general across the web the 7950 clearly outclasses the 660 ti especially at higher resolutions, even [H] agreed with their evaluation.
I don't think [h] would have recommended it if there were serious issues.
http://www.hardocp.com/article/2012/11/12/fall_2012_gpu_driver_comparison_roundup/
According to their own measurements there was no such problem in August.
http://techreport.com/review/23419/nvidia-geforce-gtx-660-ti-graphics-card-reviewed/4
The jury is still out and a there are a lot of unanswered questions about what exactly is going on.
Either way if you look at their August review there's nothing major demonstrated. I'm not going to speculate further except to say that more information is needed.
Personally I wouldn't return it because if the supposed issue didn't exist in August there has to be a reasonable explanation and if nothing else use the drivers from then if you experience any issues. In general across the web the 7950 clearly outclasses the 660 ti especially at higher resolutions, even [H] agreed with their evaluation.
I don't think [h] would have recommended it if there were serious issues.
http://www.hardocp.com/article/2012/11/12/fall_2012_gpu_driver_comparison_roundup/
MacLeod
[H]F Junkie
- Joined
- Jul 28, 2009
- Messages
- 8,280
I wouldnt worry about it too much especially if youve already got the card in hand. Drop it in and fire it up and see how it performs. I dont see anybody out there saying these smoothness issues are horrible and make the game unplayable, just that theyre not as smooth as Nvidias. Add to that, it doesnt appear to affect all games, just a couple of them so give yours a try and I wouldnt be surprised if you dont have any issues.
GotNoRice
[H]F Junkie
- Joined
- Jul 11, 2001
- Messages
- 12,006
AMD cards didn't show any frame latency issues in their august review because those use older drivers that lack all of the recent performance increases. If you have to revert to using older drivers, you lose all of the performance increases also.
Due in a large part to the performance increase that came from the latest beta drivers, the same beta drivers that appear to have introduced the frame latency issue. It's interesting to recommend reverting to older drivers to avoid the frame latency issue, but then point to performance numbers using the newer drivers when talking about which card is faster. Can you really have it both ways?
Let's all not forget that the 7970 gets stomped by the GTX680 using any driver up until the recent betas (the ones that cause the frame latency problem). It was only then that the 7970 became commonly accepted as a faster card for most purposes. I just wonder if people would still have gotten that same impression if they had known up front what they had to sacrifice in terms of gameplay smoothness by using those drivers.
Reverting to older drivers would only reduce the latest gains (varies per game).
Your second statement is still subjective (noting you say appear), only one site has reported on this so far. Until we get more evidence the exact issue and cause are still undetermined. You appear to have already made up your own mind and keep talking about the performance came at the cost of image quality. Perhaps so, but isn't it a bit premature to form conclusions until at least a couple more reputable sites can investigate this. That is usually the scientific process.
My point on the newer link was merely that [H] recommends it, and they likely wouldn't have recommend it if they saw massive stuttering issues.
If by 7970 you are referring to the original 7970, then yes it was beaten by the 680, but the 7970 GHz edition was faster then the 680 in many games (even the majority?) already way back in 12.7.
A quick google search found this review with 12.7's. The 7970 Ghz was only beaten in BF3 (just scanned very quickly) but even if I missed another you get the point that the ghz was already faster back then.
http://www.legitreviews.com/article/1979/1/
The final nail in the 680's coffin (performance wise) was the latest drivers increasing their lead in just about every game and finally gaining a lead in even the very few games that the 680 still had been (slightly) faster in e.g. BF3.
Let's stick to the facts. If you want to refute something please show me some proof (claims like that 7970 one are in fact false, the AMD TR issue is still up in the air).
Last edited:
The [H] card roundup used the latest beta drivers and it said this
"Middle pricing band This pricing band was far less competitive as the Radeon HD 7950 with Boost simply demolished the GTX 660 Ti across the board with regards to raw frame rates and overall game play experience across our suite of testing. "
(emphasis mine)
I own Nvidia cards, but fair is fair. Perhaps this latency issue is trumped up.
BababooeyHTJ
Supreme [H]ardness
- Joined
- Jan 21, 2009
- Messages
- 6,951
Any proof of that? Techreport hasn't narrowed down when they ran into these issues.
wabbitseason
[H]ard|Gawd
- Joined
- Jun 16, 2010
- Messages
- 1,511
Indeed. I was starting to think rationality was no longer the standard on this forum.
http://www.youtube.com/watch?v=TgM2O8Ci4SU&feature=related
From the vid.
http://hardforum.com/showthread.php?t=1733539
From the vid.
But some people are still getting it regardless of the keyboard.
http://hardforum.com/showthread.php?t=1733539
Last edited: