• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

3x crossfire review

Pure ownage of that GTX 295 which benefits from physX.

2 looks like the best value for performance. Could be the CPU and system 3X HD5870 is a bottleneck.
 
2 295's would be like having 4 gpus. Maybe when the x2 5870 comes out they'll do 2 of those vs 2 295s. The 3 way crossfire is powerful no doubt. Dominates.
 
anyone who went from a 4870x2 or gtx295 to the 5870's already knew how close those three video cards are in performance, which makes this 'revelation' that three 5870's produce impressive results totally pointless. there a few instances where two gtx295's out pace two 5870's so a comparison between two gtx295's and three 5870's would have been interesting and informative. this article is little more than a video card jerk-off session.
 
SonDa5 said:
Pure ownage of that GTX 295 which benefits from physX.
Click to expand...

not really since hardware-accelerated physx was off in batman in the review. so only cpu physx was being used - therefore the 295 didn't benefit at all.
 
Last edited:
rennyf77 said:
anyone who went from a 4870x2 or gtx295 to the 5870's already knew how close those three video cards are in performance, which makes this 'revelation' that three 5870's produce impressive results totally pointless. there a few instances where two gtx295's out pace two 5870's so a comparison between two gtx295's and three 5870's would have been interesting and informative. this article is little more than a video card jerk-off session.
Click to expand...

actually, I moved from GTX 295/4890CF to 5870, and the performance is not even that close..

even when FPS is that high, but its not as smooth as the same fps in 5870...
 
i find it odd that the 295 has higher min fps in all the games in the review than a single 5870. shouldn't it be the other way around?

and what's with the min fps for xfire/ tri-xfire in re5. they're ridiculously low @ all resolutions compared to a single 5870 or a 295. must be a driver issue. at least i hope that is the case, lol.
 
Last edited:
shansoft said:
actually, I moved from GTX 295/4890CF to 5870, and the performance is not even that close..

even when FPS is that high, but its not as smooth as the same fps in 5870...
Click to expand...

the article isn't about how smooth the performance was, it was just raw data. using the same drivers for 58xx and 48xx cards, i found the 4890 cf setup genuinely faster than the single 5870 in 3 out of 4 tests.
 
rennyf77 said:
the article isn't about how smooth the performance was, it was just raw data. using the same drivers for 58xx and 48xx cards, i found the 4890 cf setup genuinely faster than the single 5870 in 3 out of 4 tests.
Click to expand...

fps just fps, it doesn't mean anything to the game experience.

but for raw data, yes... indeed it looks fast..but not really that fast, and data shows nothing but data...

the most noticeable one for you would be Crysis Warhead, try on 4890CF, on the part where you begin in the tunnel after you defeat the alien boss. you will see like 30+fps but the screen lag like shit... but 5870 is opposite...
 
did you not try the link? i did test warhead and you're absolutely right about it being smoother, but again that's not what the pc perspective article was about. all they did was run three 5870's against a single gtx295 and say, 'wow did you see the performance difference?' well, no shit three 5870's are faster than a gtx295, but what do the numbers look like when you challenge three 5870's against two gtx295's. now that would make a more interesting article.
 
i know this thread is about the 5870s but.... does anyone know what heatsink they're using for the cpu in the review? it looks pretty sweet :D
 
kumquatsrus said:
not really, hardware-accelerated physx was off in batman in the review. so only cpu physx was being used - therefore the 295 didn't benefit at all.
Click to expand...

I didn't see that. Are you sure?

What about Vantage?
 
rennyf77 said:
did you not try the link? i did test warhead and you're absolutely right about it being smoother, but again that's not what the pc perspective article was about. all they did was run three 5870's against a single gtx295 and say, 'wow did you see the performance difference?' well, no shit three 5870's are faster than a gtx295, but what do the numbers look like when you challenge three 5870's against two gtx295's. now that would make a more interesting article.
Click to expand...

yea, I did look at the link but they didn't mention anything about it.. just pure raw data..
that is why I absolutely hate those article for none sense.
the numbers look good, but actual thing is, the game experience does not.

btw, this is not just about the smoothness I am talking about.
this is about the actual fps. make your character stay there, don't or do anything, and just look at the screen.. it shows 30+fps in 4890CF/GTX 295. but it looks less than 10fps.

not quite sure what cause that, but thats the most noticeable part I remember.
I did remember it happen on some other games too, but not exactly sure which one...its been awhile..
 
kumquatsrus said:
i was just looking at the picture of the in-game settings they used, that's all - it's the last setting shown on the picture right above the first graph benchmarking the 1680x1050 res.

http://pcper.com/article.php?aid=796&type=expert&pid=2

can't comment on vantage though. haven't paid attention to 3dmark #s in a long time, lol.
Click to expand...



I saw that but that may be the view of the settings for the hd5870 which can't use PhysX anyways. The article doesn't really say if physx was used on the GTX295.

The Vantage score is high for GTX295 and based on the score it does appear to have physX enabled. Not the best review with details like that missing.
 
shansoft said:
yea, I did look at the link but they didn't mention anything about it.. just pure raw data..
that is why I absolutely hate those article for none sense.
the numbers look good, but actual thing is, the game experience does not.

btw, this is not just about the smoothness I am talking about.
this is about the actual fps. make your character stay there, don't or do anything, and just look at the screen.. it shows 30+fps in 4890CF/GTX 295. but it looks less than 10fps.

not quite sure what cause that, but thats the most noticeable part I remember.
I did remember it happen on some other games too, but not exactly sure which one...its been awhile..
Click to expand...

i mean the link in my post to my own test of the 5870 vs 4890's in crossfire. and about the frames showing 30 but looking like 10, that's micro stuttering. i previous posts you've said that micro stuttering was eliminated in the 48xx and gtx2xx series of video cards. now you've experienced it for yourself.
 
SonDa5 said:
I saw that but that may be the view of the settings for the hd5870 which can't use PhysX anyways. The article doesn't really say if physx was used on the GTX295.

The Vantage score is high for GTX295 and based on the score it does appear to have physX enabled. Not the best review with details like that missing.
Click to expand...

exactly - so for a fair "apples to apples" comparison, they have it disabled for both the 5870/ 295 cards. i'm sure if they had separate settings they would have stated as such. moreover, a single 295 would never get those kind of average frame rates with hardware physx enabled. they would be much lower due to the heavier physics processing load. if you go to driverheaven, they had a recent batman analysis of hardware physx performance that will corroborate what i have said based on their numbers. as far as the vantage score, physx should only be enabled in software. otherwise the score would be substantially higher with it processed on the 295. again, they do this otherwise it wouldn't be a fair comparison. it doesn't matter though since 3dmark is pretty useless, imo.
 
kumquatsrus said:
exactly - so for a fair "apples to apples" comparison, they have it disabled for both the 5870/ 295 cards. i'm sure if they had separate settings they would have stated as such. moreover, a single 295 would never get those kind of average frame rates with hardware physx enabled. they would be much lower due to the heavier physics processing load. if you go to driverheaven, they had a recent batman analysis of hardware physx performance that will corroborate what i have said based on their numbers. as far as the vantage score, physx should only be enabled in software. otherwise the score would be substantially higher with it processed on the 295. again, they do this otherwise it wouldn't be a fair comparison. it doesn't matter though since 3dmark is pretty useless, imo.
Click to expand...


Well I have no problems appreciating 3dMark benchmarks.

Also I hear ya on the rest of your comments but that review is still lacking because the review didn't really go into details regarding the PhysX being enabled or not. It's a blank spot in that review.

Apples to Apples it's going to be cool to see ATI Stream computing power being enabled later on down the road to help out 3d performance like PhysX/cuda does for NVIDiA cards.
When that happens it will be a fair apples to apples review.
 
Last edited:
rennyf77 said:
i mean the link in my post to my own test of the 5870 vs 4890's in crossfire. and about the frames showing 30 but looking like 10, that's micro stuttering. i previous posts you've said that micro stuttering was eliminated in the 48xx and gtx2xx series of video cards. now you've experienced it for yourself.
Click to expand...

well, shouldn't micro-stuttering appear all the time? I experience that on 9800GX2 and 3870X2 and it happen almost all the time.
but in this case is different. it only happen in certain place and certain game. And for Warhead, it happen all the time in that map. I was thinking about engine problem.
 
no dude. what you saw is micro stuttering. just because it happens on occasion doesn't mean anything. you just experienced constant micro stuttering with the 3870x2 and 9800gx2.
 
SonDa5 said:
Well I have no problems appreciating 3dMark benchmarks.

Also I hear ya on the rest of your comments but that review is still lacking because the review didn't really go into details regarding the PhysX being enabled or not. It's a blank spot in that review.

Apples to Apples it's going to be cool to see ATI Stream computing power being enabled later on down the road to help out 3d performance like PhysX does for NVIDiA cards.
When that happens it will be a fair apples to apples review.
Click to expand...

that's cool, i don't have a problem with that. just saying for me, i don't find much value in 3dmark benchmarks anymore.

i found toms review of the 5770, your new favorite card, and it shows what hardware physx should look like on and off:

http://www.tomshardware.com/reviews/radeon-hd-5770,2446-14.html

the first chart shows hardware physx off, with a 4870 and 260 performing about the same. while the second chart shows hardware physx on, with performance dramatically dropping on ati cards due to the processing being done on the cpu (with the fps all the same @ 15fps for every resolution, which means it is cpu limited with hardware physx on), while otoh, the nvidia cards are doing the processing. i'm no genius, but i would say based on this, one can reasonably infer that the pcper review has the batman benchmarks with hardware physx off. hence, only the one pic with the game settings showing it off in the review.

as for the last part, i would agree. i hope ati cards will get their own proper gpu physics support as soon as possible.

i would say for sli and xfire, 2 cards are def. the sweet spot. a third card usually offers much smaller gains in most games (in some cases negative gains).

just for you sonda, here is a 5770 tri fire review by guru3d:

http://www.guru3d.com/article/radeon-hd-5770-in-3way-crossfirex-review-test/
 
Last edited:
rennyf77 said:
no dude. what you saw is micro stuttering. just because it happens on occasion doesn't mean anything. you just experienced constant micro stuttering with the 3870x2 and 9800gx2.
Click to expand...

Is there any proof for it to be micro-stuttering rather than engine problem? :)
because it only happen in that map, which is pretty much the entire map in that stage..
 
micro stuttering is a calculable phenomenon, whether you see it is an entirely different matter. some people are more perceptive or sensitive to it than others. but what you saw is the very definition of micro stutter. get over it.
 
kumquatsrus said:
i found toms review of the 5770, your new favorite card, and it shows what hardware physx should look like on and off:
Click to expand...

Right now my favorite single video card is the HD5850. Has tons of over clocking ability and raw power.

I'd love to have 2 of them for Xfire destruction.

Too much money for my budget right now though. Below the Xfire HD5850 configuaration my next weapon of choice is HD5770 in Xfire. 2 HD5770s in Xfire should perform right in between single HD5850 and single HD5870 performance.

Price wise the perfomance seems to scale with gpu power.
 
SonDa5 said:
Right now my favorite single video card is the HD5850. Has tons of over clocking ability and raw power.

I'd love to have 2 of them for Xfire destruction.

Too much money for my budget right now though. Below the Xfire HD5850 configuaration my next weapon of choice is HD5770 in Xfire. 2 HD5770s in Xfire should perform right in between single HD5850 and single HD5870 performance.

Price wise the perfomance seems to scale with gpu power.
Click to expand...

okay 2nd favorite card then, lol.
 
rennyf77 said:
micro stuttering is a calculable phenomenon, whether you see it is an entirely different matter. some people are more perceptive or sensitive to it than others. but what you saw is the very definition of micro stutter. get over it.
Click to expand...

I would like to see a proof on it rather than define it by your words.
I been encountering micro-stuttering and I know how it feels.
but its specifically to the stage and game, that is different problem, its more of a engine issue than card problem
 
rennyf77 said:
did you not try the link? i did test warhead and you're absolutely right about it being smoother, but again that's not what the pc perspective article was about. all they did was run three 5870's against a single gtx295 and say, 'wow did you see the performance difference?' well, no shit three 5870's are faster than a gtx295, but what do the numbers look like when you challenge three 5870's against two gtx295's. now that would make a more interesting article.
Click to expand...

Why? To see if 3 gpu can beat 4 gpu? I suppose you need to do something to help the weaker nvidia side. :)

2x 5870 owned the gtx 295 in all those benches. That's straight up 2 gpu vs 2 gpu. 3x5870 just makes the ATI victory margin even more humiliating.
 
vjcsmoke said:
Why? To see if 3 gpu can beat 4 gpu? I suppose you need to do something to help the weaker nvidia side. :)

2x 5870 owned the gtx 295 in all those benches. That's straight up 2 gpu vs 2 gpu. 3x5870 just makes the ATI victory margin even more humiliating.
Click to expand...

well considering a 5870 is on par with a 295 in most games, this review is unsurprising. 3x 5870 should obviously destroy a single 295. well, except for the min fps in re5 - that is just ridiculous.
 
vjcsmoke said:
2x 5870 owned the gtx 295 in all those benches. That's straight up 2 gpu vs 2 gpu. 3x5870 just makes the ATI victory margin even more humiliating.
Click to expand...

humiliating? how do you hurt a video card's feelings. this is the kind of fanboy rhetoric that we need to avoid. the 5870 is great because it provides premium level performance; the kind of performance that cost $150 to $200 more in the last generation, for under $400. did you really need this article to know that a gtx295 (essentially two gtx275's) are slower than two 5780's (in my testing a single 5870 was almost as fast as two 4890 cards that were originally aimed at the gtx285) we need not these excessive and arbitrary tests to understand that.
 
Can't wait for the 5870x2 to come to cause massive butt hurt to everyone. The arguments on this forums are the same every major video card release, "ONE CARD VS ONE CARD HURR DURR," or "TWO GPUS VS TWO GPUS HERP DERP." I don't see anyone bitching about only comparing quad core to quad cores or dual cores to dual cores in the CPU section. Faster is faster period, no arguing it. Two 5870s is faster than a GTX295. The GTX295 is normally faster than a single 5870, but $100 faster? Hell no, plus I for one would rather we open to crossfire with a second 5870 than quad SLI because the scaling of 4 GPUs tends to suck.

I personally would like to see some super high res reviews, like nvidia+triplehead VS ATI+eyefinity.
 
You must log in or register to reply here.
Back
Top