AMD Frame Pacing Explored: Catalyst 13.8 Brings Consistency to Crossfire
by Ryan Smith on August 1, 2013 2:00 PM ESTThe Test
For the purposes of our testing we’ll be looking at the 6 games we’ve adopted for use with FCAT due to their proven reliability. These are Total War: Shogun 2, HItman: Absolution, Sleeping Dogs, Battlefield 3, Bioshock Infinite, and Crysis 3. All of our results unless otherwise noted are using Catalyst 13.8b1 for the AMD cards, and NVIDIA’s 326.19 beta drivers for the GeForce cards.
Our metric of choice for measuring frame times and frame pacing is a metric we’re calling Delta Percentages. With delta percentages we’re collecting the deltas (differences) between frame times, averaging that out, and then dividing delta average by the average frame time of the entire run. The end result of this process is that we can measure whether sequential frames are rendering in roughly the same amount of time, while controlling for performance differences by looking at the data relative to the average frame time (rather than as absolute time). This gives us the average frame-to-frame time difference as a percentage.
In general, a properly behaving single-GPU card should have a delta average of under 3%, with the specific value depending in part on how variable the workload is throughout any given game benchmark. 3% may sound small, but since we’re talking about an average it means it’s weighed against the entire run, as the higher the percentage the more unevenly frames are arriving. For a multi-GPU setup we’d ideally like to see the delta percentages be equal to our single-GPU setups, but this is for the most part unreasonable. There is no hard number for what is or isn’t right here, but based on play testing we’d say 15%-20% is a reasonable threshold for acceptable variance, with anything under 10% being very good for a multi-GPU setup.
Finally, in our testing we did encounter an issue with Catalyst 13.8 that required we make some slight adjustments to FCAT to compensate for this bug, so we need to make note of this. For reasons we can’t sufficiently explain at this time but has been confirmed by AMD, in some cases in Crossfire mode AMD’s latest drivers are periodically drawing small slices of old frame buffers at the top of the screen. The gameplay impact is minimal-to-nonexistent, but this problem throws off FCAT badly.
To quickly demonstrate the problem, below we have two consecutive frames from one of our Battlefield 3 runs. The correct FCAT color order here is dark blue, green, light blue, and olive. The frames corresponding to dark blue and green occur on frame one, and light blue and olive on frame two. Yet looking at frame two, we see a small 6 pixel high stripe of dark blue at the very top of the image. At this point the dark blue frame should have already been discarded, as the cards have moved on to the green and later light blue frames. Instead we’re getting a very small slice of a frame that is essentially 2 frames old.
The gameplay impact from this is trivial to none; the issue never exceeds a 6 pixel slice, only occurs at the top of the frame (which is generally skybox territory), and is periodic to the point where it occurs at most a few times per minute. And based on our experience this primarily occurs when a buffer swap should be occurring during or right after the start of a new refresh cycle, which is why it’s so periodic.
However the larger issue is that FCAT detects this as a frame drop, believing that over a dozen frames have been dropped. This isn’t actually possible of course – the context queue isn’t large enough to hold that many frames – and analysis shows that it’s actually part of the old frame as we’ve explained earlier. As such we’ve had to modify FCAT to ignore this issue so that it doesn’t find these slices and count them as dropped frames. The issue is real enough (this isn’t a capture error) and AMD will be fixing it, but it’s not evidence of a dropped frame as the stock implementation of FCAT would assume.
Ultimately our best guess here is that AMD is somehow mistiming their buffer swaps, as the 2 frame old aspect of this correlates nicely to the fact that the dark blue and light blue frames would both be generated by the same GPU in a two-GPU setup.
| CPU: | Intel Core i7-3960X @ 4.3GHz |
| Motherboard: | EVGA X79 SLI |
| Power Supply: | Antec True Power Quattro 1200 |
| Hard Disk: | Samsung 470 (256GB) |
| Memory: | G.Skill Ripjaws DDR3-1867 4 x 4GB (8-10-9-26) |
| Case: | Thermaltake Spedo Advance |
| Monitor: | Samsung 305T |
| Video Cards: |
AMD Radeon HD 6990 AMD Radeon HD 7970GE AMD Radeon HD 7990 NVIDIA GeForce GTX 590 NVIDIA GeForce GTX 680 NVIDIA GeForce GTX 690 |
| Video Drivers: |
NVIDIA ForceWare 326.19 AMD Catalyst 13.5 Beta 2 AMD Catalyst 13.6 Beta 2 AMD Catalyst 13.8 Beta 1 |
| OS: | Windows 8 Pro |
Facebook
Twitter
RSS
102 Comments
View All Comments
mwildtech - Friday, August 2, 2013 - link
Are you still signed into AOL...? ;) I also haven't had many issues with either, at least from a single GPU perspective.kyuu - Friday, August 2, 2013 - link
What a surprise, the AMD-bashing trolls are out in force with long rants that nobody will read.Give it a rest guys.
Anyways, great write-up Ryan. Good to see AMD is getting the issue taken care of.
chizow - Saturday, August 3, 2013 - link
Except in this case, "AMD bashing trolls" helped fix your CF drivers. A simple "thank you" would have sufficed.TheJian - Tuesday, August 6, 2013 - link
ROFL...I sincerely thank you for the laugh ;)I liked many products over the years but have been saved by vocal complainers pointing out things to make me run, or at least wait until fixes come. I waited for RROD to get fixed with Jasper. Years of complainers finally got a fix (it took so long I started doubting I'd ever own one). My friend who jumped on x360 early shipped his back multiple times in the first year. I believe it spent more time at MS than in his house...LOL. He was a vocal complainer in their forums etc but I never called him a MS bashing troll for it. I laughed and thanked him for being one of the people who saved me years of that frustration :) He only thought that was funny after some beers...LOL
Thankfully he has a great sense of humor. He's ready with forum accounts everywhere he thinks the complainers will be for xbox1 this time (complainers have value people). But he expects to be a reader this time rather than the complainer ;) I think he'll go PS4 in the end despite the MS love he has vs. Sony. His wallet has no trouble voting against his fanboy thoughts.
I'm torn over the consoles though. I'd love to see AMD start making some cash, but at the same time I'm pretty unhappy they blew a wad of R&D money on something I want completely dead instead of cpus/gpus/arm socs. Had that R&D went to PC's I don't think I'd be making these statements dissing AMD. At the least they could have kept the layoffs from happening (losing 30% of your smartest people will shaft us on PC's for a few years at least and longer if consoles don't take off by the millions), and had good drivers all last year. That also might have given them a better reputation thus not needing to give out free games that are clearly wiping out profits (Q report shows this). AMD has a great gpu. It's a pity they didn't have enough funding for R&D to pair it with a great driver from day1 and funding to avoid the Fcat disaster. Even if it affects a small group it causes a lot of people to paint your other products with that image.
Steveymoo - Friday, August 2, 2013 - link
Interestingly enough, I seem to remember my GTX 460s having microstutter and performance issues in SLI. To the point where your experience in twitch games would be better if you just disabled on of the GPUs. However, over the years, and many driver updates, I don't seem to notice it any more. Nvidia really must have quite a talented software team, who communicate well with the hardware division. I would say there might be some kind of company structure issues for an issue such as this to go unnoticed, and un-fixed for such a long time.anubis44 - Friday, August 2, 2013 - link
Ssshhhhh! TheJian will be all over you like a duck on a june bug! Remember, Nvidia's drivers are always perfect! They never make any mistakes......well, except for the chronic problem I had with the GTX670 card I bought for my 3 monitor setup - kept requiring about 20 steps to get all three screens to display due to bad default refresh rate/synch issue in the Nvidia driver. Got so frustrated having to go through 20 steps every time I updated to a newer driver that I sold the card for close to what I paid (~$400) and bought a Gigabyte 7950 for about ~$100 less and flashed the bios to 1050MHz. 3 monitors in eyefinity set up in about 5 minutes in the Catalyst control panel and not a problem since.
DanNeely - Friday, August 2, 2013 - link
Are you using display port monitors or an active DP-DVI adapter for your third monitor? If the latter, has it finally gotten plug and play vs the problems when it first came out? I was never able to get an adapter to work with my 5870; and since my setup wasn't EF compatable anyway (2x 1200x1600 1x 2560x1600) ended up cutting my losses with a 5450 for the 3rd monitor and went nVidia for my next GPU in response.krutou - Friday, August 2, 2013 - link
Nvidia is known to suck at multi-monitor support because AMD was the first to develop the technology. One of AMD's few strengths is Eyefinity support.TheJian - Tuesday, August 6, 2013 - link
From the article (and this is repeated at every site reviewing the drivers):"So what’s being addressed in phase 1? Phase 1 is being dedicated to Direct3D 10+ games running on a single display. What’s not being addressed in the first driver are the Direct3D 9 and OpenGL rendering paths, along with Eyefinity in any scenario."
So Eyefinity has issues and isn't even touched with phase1. At the very least AMD is the opposite of strength with eyefinity for now. Phase2 maybe? ;)
https://en.wikipedia.org/wiki/File:Graham%27s_Hier...
You've stated a point without backing it (4th, green).
Refutation:
I found your mistake and explained why it is one and backed it with a direct quote (from this article no less...ROFL) thus proving my point ;) That's the purple one :) But I'm pretty sure I made it into the grey anyway. Your central point is debunked. But I can live with purple if it makes you feel better.
Being first has no bearing on who is better later. Horses got us from point A to B first, long before cars right? But that didn't stop a car from blowing them away later. I could say the same about the first car engine vs. say a Lamborghini engine today. First doesn't mean best.
TheJian - Tuesday, August 6, 2013 - link
Why, he's pointing out reality and what most sites point out. All multi cards had issues for a while and still do. NV just spent a lot more to come up with the tools/software to fix it as best as possible (and I'd still go single potent vs. even NV multi given a reasonable choice). You're mistaking an accurate product complaint for fanboyism. That is not what my complaints are. There is no reason to attack his comment as I already know it's at least partially true for all CF/SLI and the fix is proven (so is AMD's lack of it up to now, and still having issues with 3 cards).Would you feel better if I ranted on Bumpgate for a few paragraphs? When a company sucks I point it out. I don't care who it is. Caminogate anyone? I ranted then too. Win8, don't get me started, Vista...(fista? Nuff said). I have equal hate for all crappy releases no matter how much love or hate I have for a company (I hate apple's tactics & pricing, but they do generally have a good polished product). If AMD releases a great 20nm product and NV sucks I will RAVE for AMD and shout at the top of my lungs how NV's product sucks. Based on R&D I doubt NV will suck but AMD can still get out a good product, I just need proof at this point due to lack of funds/engineers pointing to a possible problem launch again.
Comically you miss the entire point of any of my posts (which are backed by data from other sites etc), then rant yourself on NV. Congrats though, at least you made it to the 4th rung here (well sort of):
https://en.wikipedia.org/wiki/Ad_hominem
But not without making the 2nd worst type of argument first...ROFL. You're not outing me here, you're outing yourself.