ASUS ARES CrossFire

[W1zzard]

single card furmark is 450w, so two cards is like 800w because they don't scale 100%

 

[kkaddu]

Awesome review as always W1zz.

Any plans to do a couple of benchies with a modest Eyfinity setup like 5760x1080 or 3240x1920.

If Asus really wanted to show what these cards are capable of they would arrange for a suitable test system as well like a 980x and an Eyefinity setup which is what these beats are aiming for, its really quite disappointing on their part. :shadedshu

HAVE TO AGREE.
atleast the equivalent 6 core AMD would make better sense. the CPU here just cannot feed these monsters enough.
I would still put my money on TRI SLI 480s (personal opinion)

 

[caleb]

Seriously no BadCompany 2...

 

[mdsx1950]

single card furmark is 450w, so two cards is like 800w because they don't scale 100%

Thats good news. Thanks.

 

[tkpenalty]

w1zz i'd redo these reviews with an i7 at 4Ghz+ ....

most of the results are bottlenecking

shows how much we need competition when the GPU is getting held back by the CPU.

 

[W1zzard]

w1zz i'd redo these reviews with an i7 at 4Ghz+ ....

most of the results are bottlenecking

shows how much we need competition when the GPU is getting held back by the CPU.

clock your cpu down by 10% and tell me how much fps you lost (equal to a 10% overclock)

 

[SteelSix]

They're on page one. W1zzard's personal rig runs a PowerColor PCS+ HD 5870.

Yep I saw the cards slotted on the mobo. Does he use an open bench setup? That would have been cool to see. He's the only one I've seen with 2 of these.

 

[Mussels]

HAVE TO AGREE.
atleast the equivalent 6 core AMD would make better sense. the CPU here just cannot feed these monsters enough.
I would still put my money on TRI SLI 480s (personal opinion)

more cores wont help 95% of games out there, you need faster cores, not more of them.

If w1zz wanted to show the true power of these monster cards he'd need one hell of an OC'd system - and it'd be useless for the rest of his reviews, so its a waste of time and money.

 

[Fourstaff]

If w1zz wanted to show the true power of these monster cards he'd need one hell of an OC'd system - and it'd be useless for the rest of his reviews, so its a waste of time and money.

Well, this review is "different", so I say W1z might want to do a special run, if he has the time and also if he wants to prove that the rig is not cpu bottlenecked. How often do you see 4 top end chips running in a rig?

 

[Mussels]

Well, this review is "different", so I say W1z might want to do a special run, if he has the time and also if he wants to prove that the rig is not cpu bottlenecked. How often do you see 4 top end chips running in a rig?

is it worth his time/potential degradation of the components in his benchmark system? what if something goes wrong and fails?

yes, its a point of interest for us all... but at the same time, the benefits we get (only useful in quad GPU reviews at present) arent really going to occur often enough for it to be worthwhile.


I suppose if he had the time he could simply drop 500Mhz off the system (lower the multiplier or something) and see how its affected, and assume it would react the opposite when OC'd higher.

 

[W1zzard]

I suppose if he had the time he could simply drop 500Mhz off the system (lower the multiplier or something) and see how its affected, and assume it would react the opposite when OC'd higher.

correct assumption, but i have no time (do i ever?) .. working on the new nvidia card reviews for monday

 

[lism]

Great review man.

 

[dir_d]

i think with this much video power, you may well be CPU limited - even on a 3.8Ghz i7

I think you would need a 4 to 5Ghz 980X to really see the difference between the cards

 

[crazyeyesreaper]

and i dont get what the hubub is about serious 2 $1000 gpus i can see ppl grabbing 2 5970s maybe since there like $700 but $1k per gpu the review was nice W1zz but i cant see why ppl are complaining its nice to know what these gpus CAN do but the fact it 99% of the ppl here will never have 2 ARES cards. and to get an idea on there performance look at 5970 crossfire reviews where they overclocked the gpus etc and look at the new 4gig 5970 reviews someone will do said review the ARES cards are not THAT special there cherry picked and ran at 5870 clocks last i checked sapphire did the same thing on there 4 gig model.

i agree eyefinity tests are intresting but there not worth the hassel to test since very few ppl will game that way aka lots of extra effort to appease 1% of the world

 

[Meizuman]

Would've been fun to see how a 890FX + 1090T had perform with those. Just because there were benches some time ago where PH II X4 would beat i7 in some games.

 

[xtremesv]

Great review W1zz

We now know that it takes $2000 of graphics hardware, drawing up to 900W on its own, to almost get 50fps in Crysis at 2560x1600.

Maximum strength... maximum speed

how pathetic 5850 looks on those benchies (of course analysing irrelevant >80 fps)

Nice review.

 

[WarEagleAU]

All I can say is WOW!

 

[SamHughe]

First of all, awesome review. One of the best reviews I read about this particular card.
The results for ARES+5870 is the most interesting. I bet the performance is even better with ARES+5870 (2GB) card. I'd love to see someone comparing that config to 3x5870 2GBs.

 

[W1zzard]

First of all, awesome review. One of the best reviews I read about this particular card.
The results for ARES+5870 is the most interesting. I bet the performance is even better with ARES+5870 (2GB) card. I'd love to see someone comparing that config to 3x5870 2GBs.

sure, send me 2x 5870 and 1x 5870 2 GB and i can do that

 

[wteSmithy]

more cores wont help 95% of games out there, you need faster cores, not more of them.

If w1zz wanted to show the true power of these monster cards he'd need one hell of an OC'd system - and it'd be useless for the rest of his reviews, so its a waste of time and money.

I suppose we have to let our imagination run with that one and say the fps in the review (digested every word!) is the minimum you'd expect to get for your hard-earned cash. Whereas typical hardcore gaming CPUs such as the 965/975/980s are oc'd 4 - 4.5Ghz and would probably gain another 20% fps quite easily. So we can forget that until someone posts their bling bench setup here.

I was just happy you included my old-ish 4870 512MB ( I have two so I'm not that poor ) as comparison at the end, since 'other' websites are ignorant to lesser cards such as mine

 

[W1zzard]

no way you can see 20% more fps from 10-15% more cpu clock. show me your data before you make wild claims .. reduce your cpu clock by 20% whats the loss in fps?

 

[wteSmithy]

no way you can see 20% more fps from 10-15% more cpu clock. show me your data before you make wild claims .. reduce your cpu clock by 20% whats the loss in fps?

There is a common fact that there is no way to say universally that all PCs perform the same and I didn't intend for it to come across as that. I was saying in comparison from a friends PC which is a watercooled 980x @4.4Ghz, running a Ati 5970 and Corsair Dominator GT RAM.

I spoke to just now and he said: on 3dMark06 from 3.3Ghz to 4.4ghz, his performance was 20% better. Quite obvious since it's CPU limited. On Stalker: CoP, he found that minimum fps increased quite a bit and average fps was just a small increase. Maximum fps actually didn't increase on Crysis but framerate was more stable than at lower clocks.

 

[blibba]

There is a common fact that there is no way to say universally that all PCs perform the same and I didn't intend for it to come across as that. I was saying in comparison from a friends PC which is a watercooled 980x @4.4Ghz, running a Ati 5970 and Corsair Dominator GT RAM.

I spoke to just now and he said: on 3dMark06 from 3.3Ghz to 4.4ghz, his performance was 20% better. Quite obvious since it's CPU limited. On Stalker: CoP, he found that minimum fps increased quite a bit and average fps was just a small increase. Maximum fps actually didn't increase on Crysis but framerate was more stable than at lower clocks.

That's a 33% increase in CPU clock speed for 20% more benchmark (not real world) performance.

I also am looking forward to some GTX480 quad SLIs thrown in, and maybe a higher clocked i7

To the guy who suggested a 6 core phenom earlier - most games won't make good use of 6 cores, the phenoms would struggle to keep up even more than the 3.8ghz i7. (Wow, yes I did just say that the 3.8ghz i7 was struggling to keep up.)

Seeing as you have so much spare time Wizz, can we see some Nvidia vs ATI comparisons with 3 displays