Feedback

Frame Rating Part 3: First Results from the New GPU Performance Tools

Author:
Manufacturer: PC Perspective

In case you missed it...

UPDATE: We have now published full details on our Frame Rating capture and analysis system as well as an entire host of benchmark results.  Please check it out!!

In one of the last pages of our recent NVIDIA GeForce GTX TITAN graphics card review we included an update to our Frame Rating graphics performance metric that details the testing method in more detail and showed results for the first time.  Because it was buried so far into the article, I thought it was worth posting this information here as a separate article to solict feedback from readers and help guide the discussion forward without getting lost in the TITAN shuffle.  If you already read that page of our TITAN review, nothing new is included below. 

I am still planning a full article based on these results sooner rather than later; for now, please leave me your thoughts, comments, ideas and criticisms in the comments below!


Why are you not testing CrossFire??

If you haven't been following our sequence of stories that investigates a completely new testing methodology we are calling "frame rating", then you are really missing out.  (Part 1 is here, part 2 is here.)  The basic premise of Frame Rating is that the performance metrics that the industry is gathering using FRAPS are inaccurate in many cases and do not properly reflect the real-world gaming experience the user has.

Because of that, we are working on another method that uses high-end dual-link DVI capture equipment to directly record the raw output from the graphics card with an overlay technology that allows us to measure frame rates as they are presented on the screen, not as they are presented to the FRAPS software sub-system.  With these tools we can measure average frame rates, frame times and stutter, all in a way that reflects exactly what the viewer sees from the game.

We aren't ready to show our full sets of results yet (soon!) but the problems lie in that AMD's CrossFire technology shows severe performance degradations when viewed under the Frame Rating microscope that do not show up nearly as dramatically under FRAPS.  As such, I decided that it was simply irresponsible of me to present data to readers that I would then immediately refute on the final pages of this review (Editor: referencing the GTX TITAN article linked above.) - it would be a waste of time for the reader and people that skip only to the performance graphs wouldn't know our theory on why the results displayed were invalid.

Many other sites will use FRAPS, will use CrossFire, and there is nothing wrong with that at all.  They are simply presenting data that they believe to be true based on the tools at their disposal.  More data is always better. 

Here are these results and our discussion.  I decided to use the most popular game out today, Battlefield 3 and please keep in mind this is NOT the worst case scenario for AMD CrossFire in any way.  I tested the Radeon HD 7970 GHz Edition in single and CrossFire configurations as well as the GeForce GTX 680 and SLI.  To gather results I used two processes:

  1. Run FRAPS while running through a repeatable section and record frame rates and frame times for 60 seconds
  2. Run our Frame Rating capture system with a special overlay that allows us to measure frame rates and frame times with post processing.

Here is an example of what the overlay looks like in Battlefield 3.

View Full Size

Frame Rating capture on GeForce GTX 680s in SLI - Click to Enlarge

The column on the left is actually the visuals of an overlay that is applied to each and every frame of the game early in the rendering process.  A solid color is added to the PRESENT call (more details to come later) for each individual frame.  As you know, when you are playing a game, multiple frames will make it on any single 60 Hz cycle of your monitor and because of that you get a succession of colors on the left hand side.

By measuring the pixel height of those colored columns, and knowing the order in which they should appear beforehand, we can gather the same data that FRAPS does but our results are seen AFTER any driver optimizations and DX changes the game might make.

View Full Size

Frame Rating capture on Radeon HD 7970 CrossFire - Click to Enlarge

Here you see a very similar screenshot running on CrossFire.  Notice the thin silver band between the maroon and purple?  That is a complete frame according to FRAPS and most reviews.  Not to us - we think that frame rendered is almost useless. 

Continue reading our 3rd part in a series of Frame Rating and to see our first performance results!!

 

View Full Size

Data gathered from FRAPS

Here is a typical frame rate over time graph as generated by FRAPS.  Looks good right?  CrossFire and SLI are competitive with the advantage to the HD 7970s.

View Full Size

Data gathered from Frame Rating Capture

This is the same graph with data gathered from our method that omits RUNT frames that only represent pixels under a certain threshold (to be discussed later).  Removing the tiny slivers gives us a "perceived frame rate" that differs quite a bit - CrossFire doesn't look faster than a single card.

 

View Full Size

Data gathered from FRAPS - Click to Enlarge

This is the raw frame times as captured by FRAPS - again we are looking for a narrow band of frametimes to represent a smooth experience.  Both single cards do pretty well, but SLI sees a bit more variance and CrossFire sees a bigger one.  Quite a bit bigger.

View Full Size

Data gathered from Frame Rating Capture - Click to Enlarge

Here is the same data gathered by our new capture system - the CrossFire configuration looks MUCH worse with many frames hitting near 0ms of screen time.  That would be great if they were ALL like that but unfortunately they also scale up to 20ms and higher quite often.  Also notice NVIDIA's is actually MORE uniform indicating that there is some kind of smoothing going on after the frame leaves the game engine's hands. 

 

View Full Size

Data gathered from FRAPS - Click to Enlarge

Let's zoom in a bit - here is 100 frames of the FRAPS frametimes from above.  Notice the see-saw effect that the CrossFire output has...

View Full Size

Data gathered from Frame Rating Capture - Click to Enlarge

And see how much worse it is here in our Frame Rating Capture configuration.  The pattern is actually exaggerated on the CrossFire solution while the SLI configuration is smoother. 

 

Here are a couple more screenshots from our captures.

View Full Size

Frame Rating capture on Radeon HD 7970 CrossFire - Click to Enlarge

 

View Full Size

Frame Rating capture on GeForce GTX 680s in SLI - Click to Enlarge

 

View Full Size

Frame Rating capture on Radeon HD 7970 CrossFire - Click to Enlarge

 

View Full Size

Frame Rating capture on GeForce GTX 680s in SLI - Click to Enlarge

 

This is just a preview of what we have planned for our new Frame Rating Capture performance testing method.  We have gone through many games with this and the results can vary from looking better than FRAPS to looking much, much worse. 

I am eager to get your feedback - please feel free to leave comments below and the follow on to the conclusion of our GeForce GTX TITAN review!

 

February 22, 2013 | 12:14 PM - Posted by Stefen Heiff (not verified)

I like the way you use hardware-based methods to get framerates and frame times. It's seems to be the best setup to evaluate performance instead of software that will use math and averaging to output a number.

March 28, 2013 | 04:33 PM - Posted by Anonymous (not verified)

Oh good, the pathetic nature of AMD's video cards using the new method is exposed finally - just a litle bit.

LOL - I wonder how many goodies AMD promised to keep this under the radar as long as possible while they SCRAMBLE to find a fix... wow those TOP END meet and discus DINING MEALS must have been just awesome and exquisite...
Def jelly.

Anyway, I'm glad AMD is exposed for the garbage that it is - this just shows to me why there are 100 other driver problems with AMD cards, while nVidia is near perfect 100% of the time.

I don't expect the radeon ragers to ever admit it, though.
AMD will die and they will collapse into a puddle of delicious forever tears and blame nVidia and their evil plots... and the "stupid public" "conned by advertising"... but NEVER the results of the above test and those like it we have had for half a decade now.

I do wonder how AMD gets away with it for years at a time before they are exposed... how shameful it must be when they finally realize they have been crap again for half a decade.

February 22, 2013 | 01:33 PM - Posted by gamerk2 (not verified)

Great review. Keep it comming guys.

February 22, 2013 | 01:51 PM - Posted by Aurhinius (not verified)

Great stuff just want to see AMD's response to their crossfire performance!

February 22, 2013 | 02:04 PM - Posted by I don't have a name. (not verified)

Excellent stuff. Thank you for the real insight. :)

February 22, 2013 | 02:39 PM - Posted by serpinati of the wussu (not verified)

dude, that is crazy! You think the amd cards in crossfire just synchronize poorly to produce this? Or maybe there's a conspiracy going on (cue x-files theme).

March 28, 2013 | 04:45 PM - Posted by Anonymous (not verified)

Yes the conspiracy has been to cover up, forgive, not mention, minimize, and give a big fat break to AMD every time, all the time.

That has produced this pathetic result, and a hundred other unresolved AMD vid car issues that carry on forever, maybe getting fixed a month or two before obsolescence, if then.

So this is very good news - the other single time AMD got jumped was on "crossfire drivers" @ hardocp - they have Terry Maked0n "Catalyst Maker" top dog in and he typed it: " I didn't even know we had a problem".

Yes, he didn't know - or he lied, the CF drivers shortly thereafter TOOK A GIANT LEAP FORWARD AND SURPASSED nVidia for the 1st time ever.

So somehow, AMD insulates all it's engineers ( the few they have left ), all it's PR people, all the disappointed users on their tyranny managed lock and delete "help" website, and all that secret no monkey see hear or taste total blackout leaves them blowing forevermore.

When something finally cracks through the impenetrable blackness of contact with the real world, they actually fix it ! (Well, maybe only if they get a LOT of help from end users replicating problems and sending in their cards and systems and wasting their lives while not getting paid).

Yeah, man amd so blows, hope they die soon.
(maybe their hardware (or engineers) are only meant for the direct run on hardware code that the gaming consoles use, and when it gets extrapolated off direct native hardware into a PC and for gaming, it will never be done correctly).

They need to die, be bought off, a then managed properly, without the lalalallalalala fingers in their ears method.

February 22, 2013 | 02:57 PM - Posted by Anonymous (not verified)

You should try this method on a console!

February 22, 2013 | 03:11 PM - Posted by Ryan Shrout

Only possible if we can get some kind of developer kit and run the overlay application on it as well.

Much more complex.  :/

February 22, 2013 | 03:15 PM - Posted by mateo (not verified)

Ok I'll repost here.

Does turning Vsync ON, fixes CF presenting issues.
If it does, is the performance of Vsycned CF in line when compared to Vsync OFF results.

Also does the Vsynced sysetm exhibits frame skipping like with RadeonPro "smoothing"?

And I'm not asking specifically about CF alone, but SLI and single GPUs also.

February 22, 2013 | 03:17 PM - Posted by Ryan Shrout

Yes, Vsync would usually cause frames to be "dropped" from presentation.  

We are going to be doing Vsync enabled testing for our full article as well.

February 22, 2013 | 03:22 PM - Posted by mateo (not verified)

Awesome.
I won't have to flip back and forth through 10 reviews.
PC Perspective is going to be The Mother Of All Reviews And Benchmarks.

Now the question is... WHEN

February 24, 2013 | 03:28 AM - Posted by Anonymous (not verified)

But dose it also change the numbers, and give a more realistic view of whats happening?

February 22, 2013 | 03:57 PM - Posted by Nilbog

Do you plan on explaining what exactly the overlay colors mean, etc? Maybe i missed something, going to go back and re-read the other articles just in case.

It would also be cool if you also compared last generation cards, like the 580, etc. Or maybe that's just too many cards.

I also wanted to ask why exactly you aren't testing CrossFireX? I understand that these testing methods don't put the cards in the best light. However if that's how they are performing then it should be noted. It is important to me that these kind of things are brought up, in order to encourage better support, and properly inform consumers.

I have the exact same issues with my 6970 CrossFireX setup.
After you posted the first Frame Rating preview, i have started looking more closely at my frame times as well.
This is clearly an AMD driver issue, and has been an issue for a while.
I personally think this should be publicized more until they actually fix these issues for all relevant cards, not just the 7000 series. Smooth gameplay is a big deal IMO.

February 22, 2013 | 05:01 PM - Posted by Brokenstorm (not verified)

If I understand correctly, the color overlay is simply a sequence of colors to indicate if any frame were dropped.

February 22, 2013 | 04:24 PM - Posted by kENNY (not verified)

Can you please change the colors of all your charts to be visible by color blind readers?

February 22, 2013 | 05:54 PM - Posted by Ryan Shrout

Hi Kenny, I completely apologize for that.  I didn't think about it all but I'll try to find out more on the topic and pick some better colors next time.

I had selected red and green to represent AMD and NVIDIA.  :)

February 22, 2013 | 07:32 PM - Posted by Dan (not verified)

Damn Ryan, this rocks.

February 22, 2013 | 10:16 PM - Posted by Anonymous (not verified)

You guys are so full of Sh*t with these benchmarks. Even TR is saying Xfire is fixed, you guys and anther biased site is claiming these absurd results for reasons we know well (Hits & slandering AMD).. I am positive nvidia have purchased you guys from a while back bcoz you guys have consistently bashed AMD and praised nvidia.. you even praised the 1000 card lol

I just hope AMD gives you guys what you deserve... Hopefully before next AMD gen cards (No Cards for you anymore)

February 22, 2013 | 10:56 PM - Posted by Kyle Riste (not verified)

What is wrong with you have you even read the previous articles and how much money and time and programming went into getting this right and making sure that gfxcard companies are actually doing what they advertise and not some b/s. PcPer is calling these companies on their crap and its a service to us the consumers.

February 23, 2013 | 06:00 PM - Posted by crikeyson (not verified)

I agree the poster you are replying to is over the top in his statements.

I do feel though that PCPer is definitely nvidia biased and affiliated to a certain extent. Not that I would completely discount these results. There is obvious nvidia partiality from this website though and I hear these sentiments echoed on many tech forums.

If he were truly 'looking out for consumers' he would of crucified the Titan for being nothing beyond the genuine GTX 680 with its priced doubled, yet he praised it and went so far as to have an nvidia PR man shilling on his website in a live stream.

Yeah, clearly not biased.......

Look forward to more investigations about this frametime business, but I do realize many users have been using crossfire and SLI for years now and it's only a few tech sites that are screaming about this issue, while the end-users have been quite happy with their experiences.

Maybe if this site was not so connected to nvidia it would be more palatable to take these results at face value. But with it having nvidia marketers on the site trying to sell a huge rip-off card like Titan, along with a Titan review not once really laying waste to it being a massively overpriced VGA in relation to history and performance...

well, it gets much more difficult not to take this all as a shill.

March 1, 2013 | 10:10 AM - Posted by DrShuey (not verified)

The funny thing is, he had AMD on a couple of days before he had nVidia on. He has done a ton of live stuff with AMD including his next event on 3/5.

March 6, 2013 | 07:15 PM - Posted by anon (not verified)

Do your homework before making a subjective statement. Accusing pcper of giving more positive reviews for Nvidia cards based on pure preference and not user experience is a grievous error.

Before there was pcper.com there was amdmb.com - they were the same site and it was an entirely AMD based review site. amdmb.com evolved into pcper.com to widen its user base and stay relavant. pcper is still evolving and atte4mpting to remain at the tip of the spear by including reviews of tablets and phones occasionally. There is nothing more cerain than change and hopefully pcper will stay ahead of the curfe.
Many of the pcper old hats are still AMD die hards and prefer AMD because they prefer the underdog - especially in the AMD vs. Intel side of things. When ATI was purchased by AMD, many saw this as a great opportunity for the German based company to continue to develop enthusiasts class chips. The further split of AMD's fabs into a separate company, Global Foundries, gave some hope to many that the underdog would once again be able to develop an x86 chip with performance exceeding that of Intel.

The slanderous statements are entirely incorrect -

HyperMinimalism

March 28, 2013 | 04:51 PM - Posted by Anonymous (not verified)

AMD fans are so used to being given a thousand breaks at every website but two, when AMD is kissed gloriously then finally the "bad data that was held back for AMD's sake" makes it to a webpage, the fanboys go crazy...

If you're coddled forever, slightly less coddling is considered unforgivable. That's the real story with AMD and all the website save 2 of them.
If you know anything, you know who those 2 websites are. They don't coddle AMD and baby it and give it endless breaks.

February 23, 2013 | 01:01 AM - Posted by ThorAxe

Appearing to be fixed on limited data is not the same as actually being fixed.

Have you not seen the contests PC Per have done? The majority have been sponsored by AMD?

Take your head out of the sand.

March 28, 2013 | 04:54 PM - Posted by Anonymous (not verified)

One would think a proper AMD fan reaction ( if said person was sane ) would be along the lines: " So I've been gettig screwed all this time ?!!?? What took so long for this to be exposed! Why is AMD "fixing it" finally while claiming "it's not really a problem" ? "

But no, the amd fan merely hates the facts, and is happy no matter how pathetic the amd video card is. They must drop hundreds of Verdetrol pills a day. ( google it ! LOL )

February 23, 2013 | 01:28 AM - Posted by tackle70 (not verified)

TR is not saying xfire is fixed... not by a long shot. Check out their Titan review and watch how 7970 CF looks awesome on average framerate and not so much on 99th % frametime

February 23, 2013 | 05:39 AM - Posted by Stephen H (not verified)

Hi Anonymous (if that's your real name) - I'm going to guess you disagree with the data. Can you please explain on what basis you disagree, including references where possible to data supporting your argument?

Alternatively, come back in a few years once you've made it past puberty and can hopefully make meaningful and useful comments.

Mr Shrout, I came to this article after hearing you on This Week in Computer Hardware over the last few weeks. You have provided very good descriptions in that podcast, and here. I like that PC Perspective is putting some time and thought into what the end user really sees, rather than just going with "average fps". I haven't had the problem lately, but a few years back graphical stutter was the bane of the PC gamer's life.

Have you got any thoughts on why you get this effect with Crossfire?

March 1, 2013 | 09:14 AM - Posted by cyberwire

Have you even looked at this site before? There is almost always a AMD banner or ad somewhere for their video cards. Stop fanboying behind the anonymous tag and join in the discussion and learn a thing or two.

February 23, 2013 | 01:18 AM - Posted by Circuitfeak (not verified)

Have you guys tried changing the Crossfire rendering method to see if it changes your graphs? I know SLI used AFR style while AMD uses the checkerboard style rendering. Just wondering if that is having any effect.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.