5870 CF worse than 5970 in furmark??

[mda96bjh]

MIght just be me - I'm not too worried about benchmarks but just wanting to confirm this is not a problem with my pc...


done 2 identical builds for me an my bro except for graphics cards...

He has i7 930 @ 4ghz Win7 64 and a 5970 - scores 9000 in furmark multigpu

I have the same but 2 x vapor-x 5870 OC & only score 6000


?? Is this a bug in the program or am i likely to have set something up worng?? My system is reading corssfire enabled & seems to be working OK - both on Cat 10.7 officials



ps not liking catalyst drivers at the moment - may have to see & got for SLI 480's at this rate! I have a 1250w PSU so should be able to cope with the power draw.

 

[Mr Krugga]

Try 3DMark Vantage and see what happens there. I wouldn't bother with benchmarking with Furmark tbh.

 

[gingergamer]

Yeah. Use Furmark to put your cards under load if you are lets say overclocking and need to see temps, but as far as benchmarking goes stick to 3D Mark.

 

[Rroff]

If you want to compare cards... forget FurMarks or 3D Marks... throw it through a few benchmarks like resident evil 5.

 

[aliza]

nice post

enjoy here

 

[coupe69]

nice post

enjoy here

nice spam.

 

[Rroff]

RTM'd

 

[coupe69]

RTM'd

I've RTM, I think your sig needs updating

 

[Rroff]

Just can't bring myself to do it

 

[drunkenmaster]

furmark is INCREDIBLY inconsistant with SLI and xfire, look around various reviews that base their results on furmark and its without question not putting a full load on both cards.

I normally refer people to [H]ocp's 480gtx/480gtx/release review and the review a few days later of 470gtx sli.

The 480gtx uses X watts under load, the 480gtx uses x+100w(can't remember X off the top of my head), in furmark, basically one card, full load, second card, not even close, they utterly retarded used that as a reason to suggest 300W is not even close to possible for this card, despite calling it a power hog, and laughing at its temps/heat throughout).

The review not a few days later shows a 470gtx using X watts, and 470gtx sli using x +250W for a galaxy 470gtx custom cooled job, and X+200-220W for a stock 470gtx sli setup.

The difference was they used metro for power measurement(well game measurement, a full load on furmark that was working would be higher still), and the first review they only checked it in furmark.

Basically Furmark is utterly crap, for anything, it fails in many situations to put a full load on your gpu, and does nothing but put a completely unrealistic load on your gpu when it can run full load, so it literally does nothing.

You can be unstable in furmark and dangerous or throttling temps at say 850Mhz on a 5870, yet be able to run 1000Mhz in any game ever made without ever crashing. Its hopeless at finding a stable overclock, hopeless at being accurate, completely inconsistant, and frankly crap.

I have no idea how Furmark became widely used as it simply doesn't work properly and can't give accurate results, literally has no purpose at all.

 

[Mr Krugga]

I have to agree with the above, only used Furmark to see the worst case scenario temps for my overclock (and fix fan profiles for them). Unigine Heaven put a lot of load on my GPU just running there plus benchmarking finds any inconsistency in the results. Other than that I only use GPU demanding games for at least few hours to find out the stable clock. GTA IV served well here for a while, it starts to freeze for about 30 sec if I overclock too far

 

[mda96bjh]

Thanks for comments - I thought the score was a bit unreliable & my Heaven benchmark was way over my bro's 5970

Just wanted to make sure graphics all working OK..... Saved myself a load of troubleshooting!