GeForce versus Radeon captured on high-speed video - Skyrim at 120 and 240 FPS

Very interesting. The difference is particularly evident on smaller objects near the edges of the frustum when you focus your eye toward the most grating part of the scene (the bottom). Even on the 240 Hz capture, the motion on the NVIDIA card still seems quite smooth.

I'd like to see this kind of testing done with more games, but this is a good start.
 
that's one reason I've always go on the green side... they are not cheating and fps vs fps they are more smooth. It's been like that for years yet so little people talk about it. I remember seeing artile pop up like this every few year and it's always the same exact conclusion.
 
that's one reason I've always go on the green side... they are not cheating and fps vs fps they are more smooth. It's been like that for years yet so little people talk about it. I remember seeing artile pop up like this every few year and it's always the same exact conclusion.



Yep, same here. This is one of many factors for me that always end up sending me to the green team.
 
wow, with all this smoothness talk, I never thought it would be that noticeable, but the radeon is jittery. Not quite Michael J Fox stuttery, but in comparison to the geforce, wow.
 
The AMD card on the left in those videos is so jerky it's as if the in-game character has a limp.
 
All the TR reports are BS. Just check out how the AMD cards got far better performance months before comparing to this review.

So, the conclusion is that:

a) The latest drivers are crap.
b) There is something else.

But in no way tells us what is going on, because AMD is stronger on Skyrim...
 
I'll believe it when somewhere other than TR shows proof such as this.

Until then, it's bought-off FUD in my book, the timing of which seems entirely too convenient. How wondrous that this issue mystically appears just before the holiday season, and immediately after AMD soundly removes Nvidia's crown.
 
All the TR reports are BS. Just check out how the AMD cards got far better performance months before comparing to this review.

So, the conclusion is that:

a) The latest drivers are crap.
b) There is something else.

But in no way tells us what is going on, because AMD is stronger on Skyrim...

You obviously missed they whole point of giving the data a visual demonstration by recording with high speed cameras.
 
I'll believe it when somewhere other than TR shows proof such as this.

Until then, it's bought-off FUD in my book, the timing of which seems entirely too convenient. How wondrous that this issue mystically appears just before the holiday season, and immediately after AMD soundly removes Nvidia's crown.

They've been saying it for years. They just now released high-speed video capture of it to prove what they've been saying, instead of just frametime charts. There is no "timing" at play here.
 
If a respected review site, like [H], carries out these sorts of tests and comes to the same conclusion I'll be first in line for an Nvidia card. Until then, I think this entire techreport frenzy is dubious at best.

People like GoldenTiger would have us believe that frame-time latency is an ancient AMD problem, which is only now being proven. Oh yes, this totally proves the FUD your like have been slinging for years on end. That's bullshit. Why don't you take a look at the slightly older reviews FROM THE SAME SITE and see the AMD frame-time graphs for yourself.

More likely:

1. Techreport has messed up.

2. There's a bug in the most current AMD drivers.
 
If a respected review site, like [H], carries out these sorts of tests and comes to the same conclusion I'll be first in line for an Nvidia card. Until then, I think this entire techreport frenzy is dubious at best.

People like GoldenTiger would have us believe that frame-time latency is an ancient AMD problem, which is only now being proven. Oh yes, this totally proves the FUD your like have been slinging for years on end. That's bullshit. Why don't you take a look at the slightly older reviews FROM THE SAME SITE and see the AMD frame-time graphs for yourself.

More likely:

1. Techreport has messed up.

2. There's a bug in the most current AMD drivers.

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". :rolleyes: 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.
 
I'll believe it when somewhere other than TR shows proof such as this.

Until then, it's bought-off FUD in my book, the timing of which seems entirely too convenient. How wondrous that this issue mystically appears just before the holiday season, and immediately after AMD soundly removes Nvidia's crown.

I know right? For a moment there I thought I was reading Toms Hardware! The authority on tech :rolleyes: ..... sponsored by nvidia.... :rolleyes: /end thread
 
http://techreport.com/r.x/geforce-gtx-680/arkham-50ms.gif

http://techreport.com/r.x/geforce-gtx-680/arkham-comparo.gif

http://techreport.com/r.x/radeon-hd-7950/skyrim-560ti.gif

In case you'd like to pretend my links don't exist, GoldenTiger.

It stands as a fact that earlier Techreport reviews did not present this problem with AMD. To suggest otherwise is blatant fanboy FUD.


Like this, for example:
They've been saying it for years. They just now released high-speed video capture of it to prove what they've been saying,
 
Last edited by a moderator:
In case you'd like to pretend my links don't exist, GoldenTiger.

It stands as a fact that earlier Techreport reviews did not present this problem with AMD. To suggest otherwise is blatant fanboy FUD.

Hurrdurr that's some terribly erratic frame-time latency.

Mop mop, that's terrible too! Oh my, what a long-standing AMD problem!


Oh my word, here's a direct comparison that totally contradicts the utter nonsense you are spewing:



Cherrypicking one game graph out of dozens doesn't exactly prove your "point" when it's not even representative of most of the charts. That's the same practice you like to do during new releases though to try to proclaim one card a winner, though, so have at it. The rest of us will just look at the articles ourselves.

Anyway, it's not my job to educate you about computers. Have fun out there learning.

To respond to your ninja-edit with a quote from me: I said they've been talking about frametimes for years, which they have been.
 
Oh, you're still talking? I'm completely cherry picking, of course!

Here's three more:

http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-1.gif

http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-2.gif

http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-3.gif

Have the good sense to cut your losses when your position has been completely discredited. Why don't you provide some of the obvious proof available online that you alluded to here:

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.
 
Last edited by a moderator:
Oh, you're still talking? I'm completely cherry picking, of course!

Here's three more:

http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-1.gif

http://techreport.com/r.x/radeon-hd-7970-ghz/bf3-2.gif


Have the good sense to cut your losses when your position has been completely discredited.

That's amusing considering those graphs (from ONE more game, not three), actually show the 680 in a favorable light compared to the 7970 ghz there. The Radeon has many, large spikes in frametimes, while the GeForce stays mostly steady and has smaller spikes when it does have them.

Anyway, done feeding you... have fun!
 
Last edited by a moderator:
2. There's a bug in the most current AMD drivers.

Indeed, as of right now there doesn't appear to be any version of AMD drivers for 7xxx hardware that contains the recent performance increases without also having this frame latency problem. The only current known way to avoid the frame latency issue is to revert to older drivers that contain none of the recent performance increases either.

The two issues might be directly related, or they might be completely separate, but until there is a driver that fixes the frame latency issue while retaining the performance increases, the two issues are effectively linked.

What these drivers have done for AMD, even as betas, has been huge. They served to reverse the common perceptions of which top-end card was fastest, and did so conveniently right before the holiday shopping season...

I'm looking forward to WHQL versions of the recent AMD drivers and hope that clears some things up, but I would bet that we will not see any new WHQL drivers from AMD before Christmas ;)
 
That's amusing considering those graphs (from ONE more game, not three), actually show the 680 in a favorable light compared to the 7970 ghz there. The Radeon has many, large spikes in frametimes, while the GeForce stays mostly steady and has smaller spikes when it does have them.

Anyway, done feeding you... have fun!

I'm still cherry picking here GoldenTiger. Feel free to provide the proof of your statement here:
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.

http://techreport.com/r.x/geforce-gtx-680/skyrim-comparo.gif

Oh look, that's Skyrim the game in question today. Interesting.

http://techreport.com/r.x/geforce-gtx-680/crysis2-comparo.gif

Another game? Couldn't be!

http://techreport.com/r.x/geforce-gtx-680/sam3-comparo.gif

Hmm.

Indeed, as of right now there doesn't appear to be any version of AMD drivers for 7xxx hardware that contains the recent performance increases without also having this frame latency problem. The only current known way to avoid the frame latency issue is to revert to older drivers that contain none of the recent performance increases either.

The two issues might be directly related, or they might be completely separate, but until there is a driver that fixes the frame latency issue while retaining the performance increases, the two issues are effectively linked.

You position seems reasonable enough, GotNoRice. Will need to wait for more information to be sure.
 
Last edited by a moderator:
Indeed, as of right now there doesn't appear to be any version of AMD drivers for 7xxx hardware that contains the recent performance increases without also having this frame latency problem. The only current known way to avoid the frame latency issue is to revert to older drivers that contain none of the recent performance increases either.

The two issues might be directly related, or they might be completely separate, but until there is a driver that fixes the frame latency issue while retaining the performance increases, the two issues are effectively linked.

What these drivers have done for AMD, even as betas, has been huge. They served to reverse the common perceptions of which top-end card was fastest, and did so conveniently right before the holiday shopping season...

I'm looking forward to WHQL versions of the recent AMD drivers and hope that clears some things up, but I would bet that we will not see any new WHQL drivers from AMD before Christmas ;)


This certainly seems to be the most likely explanation. If it turns out to be the case the second question then becomes, did this happen by accident or did AMD knowingly trade latency for FPS performance. My guess is we'll get an answer to the first, but then have a whole new fanboy-war on the second.
 
(Edit) On topic: I found the video enlightening. You can notice the stutter on the 7950 card while the 660 Ti card was more consistent.

The images are stored directly on photobucket from my own account and the link goes direct to Anandtech forums. No bandwidth theft this way. They also show that one single setting with my HD 7970 changed things from stuttery to smooth. Setting flip queue size to 1 made a dramatic difference. IMHO AMD need to look at their individual game profiles to ensure best perdormance and smoothness.
 
Last edited by a moderator:
My post was specifically at wabbitseason. If you right click on wabbitseason's image links here in this thread, they are linking directly to techreport.com's servers. (Your links on Anandtech are good and are how wabbitseason should have shared his/her links here.)

I'm not trying to "mod" the thread, I just wanted to give wabbitseason a heads up to change his links so they aren't stealing techreport's bandwidth.

My apologies, I though it was at mine. :confused:
 
Also the links that Wabbitseason is showing seems to actually support what tech report is saying. All the spikes on the AMD chart is the jitter that is seen in the video. Or am I not reading his charts correctly?
 
Also the links that Wabbitseason is showing seems to actually support what tech report is saying. All the spikes on the AMD chart is the jitter that is seen in the video. Or am I not reading his charts correctly?

No, you are completely right. That is why I have been boggling at his posts, myself.
 
Back
Top