HD 6800 Crossfire Scaling

cusideabelincoln

Diamond Member
Aug 3, 2008
3,275
46
91
Well I think some of you are curious, and I was, about Crossfire and it seems AMD has indeed improved it with the HD 6800 series compared to the HD 5800 series. I looked over various benchmarks and compiled the data into a spreadsheet. You'll find the numbers and links to the websites I used in the following Google doc:

HD 6800 Crossfire Scaling data
Updated (updating) the spreadsheet, found another review and will be adding more as I come across them over the next few weeks.
Updated again, this time the formulas are in google doc so you can download and play with the numbers for yourself.
Updated Sheet 2 now includes results when SLI/Crossfire don't scale or appear CPU limited in the benchmarks. In summary including these results brings down all the scaling numbers, but the HD 6800 > GTX 400 > HD 5000 the same anyway. I don't include these numbers in the original because we're interested in how they scale when they scale (for both SLI and Xfire), and obviously CPU-limited numbers will give us worse scaling as performance hits a ceiling.

That document has a lot of information. It has all the scaling numbers for any review I found with HD 6800 Crossfire results on a game by game breakdown. So now I'll give you the summaries (averages) and you can look at the individual data for yourselves:

HD 6800 series sees an average 82.1% gain in performance by adding a second card.
HD 5000 series sees an average 67% gain in performance by adding a second card.
GTX 400 series sees an average 75.1% gain in performance by adding a second card.

So from the data points posted across the Internet so far, Crossfire has caught up and in many cases surpassed SLI. AMD has boosted Crossfire scaling by 15% with the HD 6800 series over the HD 5000 series.

However not all reviews included Crossfire or SLI results for other cards. I really only found two (EDIT: So there are more than two out there but I'm not going to put them in this post. Just look at the speadsheet) that did include that: Anandtech and Guru3D. So in this post I'll go ahead and give the details on just how much AMD improved Crossfire over the HD5800 series, as well as compare it to the GTX 400 series. The percentages I'm posting will show you how much performance you gain by adding a second card per game.


http://www.anandtech.com/show/3987/...renewing-competition-in-the-midrange-market/8

HD 5850 vs HD 6850 vs GTX 460 scaling at 1920x1200

Crysis Warhead: 67% vs 82% vs 81%
Battleforge:::::: 83% vs 96% vs 91%
Metro 2033::::: 55% vs 66% vs 67%
HAWX::::::::::: 73% vs 92% vs 82%
Bad Company 2: 80% vs 97% vs 71%
Stalker CoP::::: 74% vs 98% vs 75%
Dirt 2:::::::::::: 66% vs 85% vs 83%
Mass Effect 2::: 87% vs 89% vs 95%
Wolfenstein::::: 35% vs 46% vs 58%
Average: 69% (HD5850) vs 83% (HD6850) vs 78% (GTX460)

(I disregarded Anno results because they appeared CPU-limited with X-fire and SLI didn't scale at all)

http://www.guru3d.com/article/radeon-hd-6850-6870-crossfirex-review/7

HD 5850 vs HD 6850 vs GTX 470 scaling at 1920x1200

Bad Company 2:::: 72% vs 97% vs 82%
Modern Warefare2: 38% vs 33% vs 69%
Dirt 2::::::::::::::: 59% vs 84% vs 82%
Far Cry 2::::::::::: 76% vs 95% vs 75%
Anno 1404:::::::::: N/A vs 63% vs 86%
Crysis Warhead:::: 63% vs 84% vs 62%
Metro 2033:::::::: 84% vs 100% vs 95%
Average: 65% (HD5850) vs 79% (HD6850) vs 79% (GTX 470)

As more reviews come out over the next month we'll get an even better picture of the situation.
 
Last edited:

flopper

Senior member
Dec 16, 2005
739
19
76
BC2 with 97% is promicing in regard to BF3.

overall, crossfire now works, better.
 

Arkadrel

Diamond Member
Oct 19, 2010
3,681
2
0
Yea Crossfire Scaleing is awesum on the 6xxx series... imagine what they ll mean when they release the card that replaces the 5970.

6870 was supposed to a improved 5770 replacement, to compete with a 460.
However the 6870 is closer to 470 performance than it is to 460s.

So 2x470 SLI performance, at 2x 6870 prices = awesum.
 

Grooveriding

Diamond Member
Dec 25, 2008
9,147
1,329
126
:thumbsup::thumbsup: AMD needed better scaling.

6870 Crossfire is definitely the better choice than 460 SLI now.
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
This is by far AMD's biggest improvement/feature with the new cards. Crossfire scaling is incredible and they've taken the one of the only two ace in the holes Nvidia had going for them (the other being the fastest single GPU, which will likely also be taken away from soon) and smacked them in the face with it.

If Nvidia truly wants to stay competitive (long term) in the discrete video graphics market, they are going to need more timely GPU releases divergent from the multi-purpose GF100 type cards.
 

Arkadrel

Diamond Member
Oct 19, 2010
3,681
2
0
SLI has always been scaleing better than Crossfire........ NOT anymore

I agree, this is a huge slap in the face to nvidia, whenever you hear people argue for buying Nvidia cards better SLI is one of the things they mention.
 

n0x1ous

Platinum Member
Sep 9, 2010
2,574
252
126
SLI has always been scaleing better than Crossfire........ NOT anymore

I agree, this is a huge slap in the face to nvidia, whenever you hear people argue for buying Nvidia cards better SLI is one of the things they mention.

I wouldn't go that far. The guru3d results show more consistency out of SLI with teh 6870's dropping into the 30% on MW2. But certainly the are much closer to parity overall now.

We will see if this translates to 6900.....
 

psoomah

Senior member
May 13, 2010
416
0
0
Well I think some of you are curious, and I was, about Crossfire and it seems AMD has indeed improved it with the HD 6800 series compared to the HD 5800 series. I looked over various benchmarks and compiled the data into a spreadsheet. You'll find the numbers and links to the websites I used in the following Google doc:

Thanks for the time and effort you put into producing such a complete and very useful chart of xfire/sli in current cards.

Very nicely done.
 

tonyjones

Member
May 25, 2004
97
0
0
Yeah scaling in general is amazing, I remember when SLI didn't get anywhere close to 50%, now its getting closer to 100%, gotta love it!
 

Grooveriding

Diamond Member
Dec 25, 2008
9,147
1,329
126
Some more impressive 6870 Crossfire performance. Unfortunately too few sites compared 6870CF directly to 460SLI in their reviews.

Stalker 6870 CF 70.6fps > GTX460 SLI 62.8fps

6870CF


460SLI



Dirt2 6870 CF 150.7fps > 114.3fps GTX460 SLI

6870CF


460SLI


Metro 2033 6870 CF 37.1fps > 34.9fps GTX460 SLI

6870CF


460SLI



Here from Guru3D you see 6870CF is faster than GTX470 SLI in Metro 2033




Crysis: Warhead




Crysis Warhead 6870 Crossfire is faster than 5870CF, GTX470 SLI and close to 480SLI.



:thumbsup: Impressive! If this scaling is on 6990 and the 6970/6950 it will be very nice.
 
Last edited:

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com

Stalker
------------
Crossfire scaling - 71%
SLI scaling - 96%

Dirt 2
------------
Crossfire scaling - 79%
SLI scaling - 85%

Metro2033
------------
Crossfire scaling - 85%
SLI scaling - 91%



Uhhh woops. In your first 3 examples SLI scales better every time. Way to not prove the point of this thread.... But in all honesty, crossfire scaling is no longer a linchpin in AMD's lineup. It's every bit as good as SLI (sometimes a little better, sometimes not quite as good, often too close to matter).
 
Last edited:

betasub

Platinum Member
Mar 22, 2006
2,677
0
0
So there's evidence provided data for a balanced argument. Didn't take long for it to get snipped & selectively quoted to suit some-else's agenda.
 

Grooveriding

Diamond Member
Dec 25, 2008
9,147
1,329
126
So there's evidence provided data for a balanced argument. Didn't take long for it to get snipped & selectively quoted to suit some-else's agenda.

Yes it's pretty sad. About all you can expect around here, fans of one camp or the other not wanting to see evidence and just spin it to their preference.

I guess the fact that 6870 CF is faster than 460 SLI every time is not as important as skewing it show a metric of no value to the user, rather than the one that is of most importance.
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
The FC2 results of over 100% made me laugh.

Anybody think that with a newer arch for Cayman the the scaling can improve further?
 

blastingcap

Diamond Member
Sep 16, 2010
6,654
5
76
I wouldn't go that far. The guru3d results show more consistency out of SLI with teh 6870's dropping into the 30% on MW2. But certainly the are much closer to parity overall now.

We will see if this translates to 6900.....

SLI is MANDATORY for Surround, thus I expect NV to give SLI ample resources to make sure it works right. CF is not mandatory for anything in AMD's camp so there is more of a risk that CF for a new game is broken.

That said, the CF scaling is impressive and may balance out the risk of slightly crappier CF support.

If Sideport made its way into Cayman then I'd expect even better scaling with Cayman.

Abe, thanks for putting this together. :thumbsup:
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
So there's evidence provided data for a balanced argument. Didn't take long for it to get snipped & selectively quoted to suit some-else's agenda.

Yes it's pretty sad. About all you can expect around here, fans of one camp or the other not wanting to see evidence and just spin it to their preference.

I guess the fact that 6870 CF is faster than 460 SLI every time is not as important as skewing it show a metric of no value to the user, rather than the one that is of most importance.

If either of you are referring to me previous comment, really? REALLY? Did not see the wink followed up with
crossfire scaling is no longer a linchpin in AMD's lineup. It's every bit as good as SLI (sometimes a little better, sometimes not quite as good, often too close to matter)

When I say "every bit as good as SLI" I don't know what else that could be said to balance that statement out.

And congrats to AMD.... I guess??? I mean honestly, *shouldn't* two $240 video cards win every time vs. two $190 (plus $20 MIR) video cards http://www.newegg.com/Product/Produc...82E16814500169 ???

$480 vs. ($380 - $40 MIR) That's a $140 price premium --- if the more expensive duo isn't winning every single time then something is wrong with their implementation, so what is the big deal here? Hurray the world turned and something else that is supposed to happen did. I was actually giving AMD big props for noticeably improving crossfire, bringing it in line with SLI, and you guys get all upset and pout over me pointing out (and jokingly at that) comical contradictions to what this thread is about. Seriously.... wow....
 
Last edited:

jackstar7

Lifer
Jun 26, 2009
11,679
1,944
126
Yes it's pretty sad. About all you can expect around here, fans of one camp or the other not wanting to see evidence and just spin it to their preference.

I guess the fact that 6870 CF is faster than 460 SLI every time is not as important as skewing it show a metric of no value to the user, rather than the one that is of most importance.

You're overstating its insignificance.

You should also consider there's a $/performance issue. Two 460s will likely be cheaper (based on pricing trends right now) than two 6870s, so while the 6870s outperform, they also should being more expensive.

Potential buyers should be considering their return on investment when going in for two cards over a single card. To me, the best part about the potential for SLI/Xfire is the buy one now and get decent results and get a second later when its affordable (and potentially cheap) and see strong improvement.

Really anyone with a board supporting Xfire/SLI is a winner when scaling is that good for both sides.

It is possible not to care which side is "better" and just see that both sides got better overall (value) in this current wave. I'm really glad to see AMD competing on this front because it's certainly the piece of the puzzle that was missing from their mid-range. Now these cards can be much more appealing for those who are comfortable with the multi-card setups.

I think the lesson of this improvement in Xfire is that we all win.
 

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
This is a bit OT, and I let it slide the first time you misused the word, but every time you misuse "lynchpin," a kitten dies somewhere. Stop killing kittens!

http://www.thefreedictionary.com/lynchpin

I think what you meant was "Achilles Heel" or something like that.

Thank you, I was actually debating how the word is correctly spelled. Firefox's spelling told me I was right so I just went with it. Noted and cataloged for future use. Glad you understood at what I was getting at.
 

betasub

Platinum Member
Mar 22, 2006
2,677
0
0
Edit: ninja'd on this one

Linchpin or lynchpin (from wikitionary.org)

Noun linchpin (plural linchpins)

1. a pin inserted through holes at the end of an axle, so as to secure a wheel
2. (figuratively) a central cohesive source of stability and security; a person or thing that is critical to a system or organisation.
 
Last edited:

tviceman

Diamond Member
Mar 25, 2008
6,734
514
126
www.facebook.com
edit: Ninja'd on this one

linchpin or lynchpin (from wikitionary.org)

noun linchpin (plural linchpins)

1. A pin inserted through holes at the end of an axle, so as to secure a wheel
2. (figuratively) a central cohesive source of stability and security; a person or thing that is critical to a system or organisation.

I SUCK! I know! Get out of my head!
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |