Oddball TRIFIRE 5970+5850 benches

[gerardfraser]

5970+5850 Benches

Warhead 1920x1080 Enthusisast Avalanche run



Warhead 1920x1080 Enthusisast Avalanche 0x,2x,4x,8x



Crysis 1920x1080 Very High Island 0x,2x,4x,8x



Call Of Pripyat Benchmark 1920x1080 4x DX10



Call of Juarez DX10 Benchmark 1920x1080 4x



StreetFighterIV_Benchmark 1920x1200 4x



Batman - Arkham Asylum 1920x1080



RE5DX10 1920x1200 4x



FarCry 2 1920x1080 Ultra 4x Ranch Med



Heaven 1920x1080 4x



Well I did some compares 5970 VS5970+5850


COP 1920x1200 DX11 MAXED



Dirt 2 DX11 1920x1200 8xMSAA Due to fps changes with view take this with grain of salt



Heaven DX11 1920x1200 8AA 16 AF MAXED



SF4 1920x1200 8x Notice benchmark jumped with higher AA then previous TRIFIRE weird



[Farcry 2 1920x1200 4x Ranch this time ULtra settings[/b]

 

[cadaveca]

DX update needed? Also, 5970 + 5850 isn't "offically" supported in Crossfire, so some oddities are expected.

 

[crazyeyesreaper]

need to overclock the i7 those gpus are being heavily bottlenecked

 

[cadaveca]

It's @ 4.2...check the warhead screenie.

 

[crazyeyesreaper]

ah there it is well as u said he still needs the DX11 update Call of Pripyt actually gets better frames in DX11 with settings maxed then it does DX10 and heaven bench is only usefull when tessellation is on hell if he checked the DX10.1 box he would see a large performance jump in CoP as well in DX10 his 3 gpus are getting the same frames i get with 2 5850s since hes running dx10 vs 10.1 or 11

 

[kylzer]

4.2 is still a bottleneck (benching)

Owait its the serial spammer lol

 

[gerardfraser]

DX 10 was on purpose,DX11 works fine.

 

[crazyeyesreaper]

well i dont see the point in running dx 10 in games that have dx11 support all its doing is holding your hardware back since DX11 CoP and heaven bench are far harder on the hardware and since 5k series can produce more eyecandy with DX11 at the same speeds of DX10 and lower image quality why not run the DX11 settings and see what the cards really have in them perfromance wise

 

[gerardfraser]

Just wanted to run DX10 benches,thats all.Sorry if it disappoints yeah.

 

[a_ump]

yes we're all extremely pissed lol nice, it'd be tight if u would just run HD 5970 benchmarks for comparison to see how well the HD 5850 scaled with the HD 5970, and just trifire in general for the HD5k series.

 

[Melvis]

Dude those are some awesome scores there, your system owned every game with ease, i think its just amazing

 

[HammerON]

yes we're all extremely pissed lol nice, it'd be tight if u would just run HD 5970 benchmarks for comparison to see how well the HD 5850 scaled with the HD 5970, and just trifire in general for the HD5k series.

It would be nice to see those comparisons

And throw in some Vantage score as well~

 

[crazyeyesreaper]

yes indeed and im not saying the benches you did werent great they are they give a good representation of trifire just id like to see the cards run DX11 CoP and DX 11 heavens bench to see how trifire scales with DX11 and how well it handles tessellation etc

 

[gerardfraser]

You asked no vantage though but GPU scores in TRIFIRE would be around 37000

Well I did some compares 5970 VS5970+5850


COP 1920x1200 DX11 MAXED



Dirt 2 DX11 1920x1200 8xMSAA Due to fps changes with view take this with grain of salt



Heaven DX11 1920x1200 8AA 16 AF MAXED



SF4 1920x1200 8x Notice benchmark jumped with higher AA then previous TRIFIRE weird



[Farcry 2 1920x1200 4x Ranch this time ULtra settings[/b]

 

[crazyeyesreaper]

thanks mate as i thought DX 11 with more eyecandy hits the minimum frame rates hard but otherwise it seems the average fps stay consistent meaning DX11 is shaping up nicely and so is trifire even tho a 5870 would be better paired with a 5970 i must say i want your gpus anyway XD

 

[gerardfraser]

I know what and why your saying a 5870 would match better.

In gaming though it makes not difference and I have compared my benches to 3x5870,have not seen many 5970+5870 but I have seen some, and they are equal and better in gaming benches.

Now if your a pro bencher and do 3dmark stuff then 3x5870 and quad fire would be better,but I am a gamer and this combo is the best way to go.
Way cheaper and as fast or faster

 

[crazyeyesreaper]

true enough and if u oc the 5850 heavily enough it can overcome the shorter shader amout anyway

just i like to have hardware specs match its like 5870+5850 its not as good as 5870+5870 but its only a 1% difference from 5850+5850 so its not worth it but with this setup it still adds a huge boost making it worth while

 

[gerardfraser]

I know what your saying man,I bought 2 5970,and some 5870 but the 5850 came in first and guess what.
Trifire is faster then quad due to drivers and no real perceivable difference between 5850 and 5870 in games with the 5970.

So I cancelled the cards and keeping what I have.

 

[crazyeyesreaper]

yea quad gpus are pretty much worthless even an i7 at 4ghz + bottlenecks a 5870 crossfire setup in many cases so more gpus wont equal more performance quad gpus wont really come into play untill heavy tessellation takes off namely on the scale of the heaven bench till then dual or triple is the best setup for the time being

hell the only game i can recall really loving 4 gpus was tom clancys hawx as it just kept scaling with more gpus

 

[gerardfraser]

Also RE5 scales well with quad,but not many more.What do ATI guys always say,wait till the next driver,all will be good,I lol everytime I read that one.

 

[crazyeyesreaper]

well the 9.12 hotfix helped me alot saw a boost of over 100fps in DMC4 with crossfired 5850s also saw a 20fps boost in crysis warhead and Dirt 2 i also saw a noticeable increase so if the new drivers out this month are better i cant really complain

 

[gerardfraser]

For me 9.12 hotfix there were no real gains in gaming,but a Significant Loss in GPU transcoding with the ATI stram 2.0
If anyone does video encode,transcode time is a huge factor.Beta drivers were better.

MP4 Video 480x270
TO
MP4 .M2TS Video 1920x1080


Each Transcode done 3 times with restarts in between.+-4 seconds difference

Core I7 4095
GPU Default Clocks

CPU Transcode
Elapsed time 2.59 minutes


GPU Transcode Driver 8.663.1_Beta5_Hemlock_VistaWin7_Nov11
Elapsed time 1.22 minutes


GPU Transcode Driver Catalyst_9.12_Hotfix_Win7_Vista_8.682.2RC1_Dec15
Elapsed time 1.37 minutes

 

[gerardfraser]

Just like to say hey to some of you smart people here

DX update needed? Also, 5970 + 5850 isn't "offically" supported in Crossfire, so some oddities are expected.

First post after I posted benches

Second post after I posted here,this guy is the smartest,duhHHHHHHHHH

need to overclock the i7 those gpus are being heavily bottlenecked

4TH post,oh smart guy again

ah there it is well as u said he still needs the DX11 update Call of Pripyt actually gets better frames in DX11 with settings maxed then it does DX10 and heaven bench is only usefull when tessellation is on hell if he checked the DX10.1 box he would see a large performance jump in CoP as well in DX10 his 3 gpus are getting the same frames i get with 2 5850s since hes running dx10 vs 10.1 or 11

5 th post

4.2 is still a bottleneck (benching)

Owait its the serial spammer lol

the smartest guy ever .Sorry about your sister,I needed a shit.

 

[crazyeyesreaper]

well thats nice bashing someone on a forum YAY no wonder your usual posts are on other forums and not here

and its been shown that even Gulftown I9 at 4ghz + is still a bottleneck in multi gpu setups

 

[johnnyfiive]

Holy shit at the Crysis results....