ATI CrossFire - Hands on

it is perfectly acceptable from a refresh rate POV, not from a resolution POV in my opinion.

I think that refresh rate is something worth taking in to account when gaming with CrossFire - I'll be taking that in to account when completing future reviews around CrossFire I think. I don't think 60Hz is 'playable' for more than about 5 minutes - I'd get a headache if I ran 60Hz on a CRT for longer than that.

 

60Hz isn't acceptable on a crt, I havnt read the thread/article, but I guess thats all it can do in 1600x1200, which means lots of people who like high end kit aren't going to buy it. A shame, ATi are really dissapointing me these days

 

Did you mean to switch those two? Makes more sense in my mind that way

 

"Oh My God the *single* X850 XT beat the 6800 Ultra in Doom 3!" - is the first which pops into mine.

ATi never beat nVidia in Doom 3 ... never ...

OK, the CrossFire result is believable, it's new technology and anything's possible, and it's great to see ATi finally clawing back (and passing) nVidia's lead. But with the single card results...?? Either the ATi drivers you used or the reference mobo (I assume you used this for the single X850 figures too) is/are seriously fast! I'd like to see benchies of some single nVidia cards on that mobo sometime, if that's even possible of course.

All in all though, a very good and informative preview.

 

i cant say that im too impressed with this article. not at all. things are run at very low resolutions which seems to be the reason that the ATi side is coming up on top in doom3.. running at these low resolutions is really testing the motherboard and cpu more than anything. i dont see too many people running their new expensive setup at 1024 4xAA (once) and 8xAF.. i know you said you werent going all out on perf numbers but this is giving people an incorrect view of what crossfire is. I have to suggest that people go to a site like anandtech. their review is run at high resolutions talks about the 1600x1200 60hz limitation in greater detail and also shows image quality improvements as bit did. the numbers over there gives you an actual idea of whats going on. frankly i was baffeled at the doom3 perf numbers till i went there.

 

In bit-tech's defence, ATI didnt leave them a lot of options; they were given pretty specific instructions on how to perform the reviews, though I agree, more high resolution gaming should have been done. The fact that Bit cant show a single 7800 GTX (which beats current crossfire options hands down...) hurts a bit too. Really, what I think this crossfire (pre)review needs is time.

 

they were given specific instructions on how to conduct this? thats kinda... shifty IMO.. so i take it anand weren't given the same instructions? or they had more time/resources or what?

 

All sites had conditions on which ATI would allow them to review the Crossfire product, including not using the 7800 GTX in their benchmarks if I understand correctly. Bit-tech is trying not to bend in to all the negative hype that has caught Crossfire in the crossfire as of late. This isnt a full on review by any means, its a preliminary evaluation; the difference that you see is in personal writing style and perspective.

 

ah, i see. it just seems to me that its almost gone too far to show crossfire in a fair light. i mean, yeah, def dont try to directly compare crossfire with a 7800gtx sli setup, theyre not in the same league. but this almost looks like it was made to show crossfire owning. (not saying that its biased, i know its not) it just looks like it worked out that way by accident because of the settings used. needless to say i go to bit-tech pretty much exclusivly untill im interested and want more info then i hunt for other stuff. and when i read this i was going around to my classes today thinking about how well ati did with crossfire, till i read Beatbox's post and thought about it for a little bit.

 

Your Doom 3 numbers looked a little strange, so I thought I'd compare them with the numbers other reviewers got. Bear in mind, test setups are obviously slightly different among each reviewer, but deltas should remain largely the same.

Edit: I compiled a little nifty-grifty chart to make comparisons somewhat easier.



Anand's deltas are the closest out of any reviews (about 1 FPS), but the X850XT doesn't surpass the 6800 Ultra in any review (or preview, if you prefer) let alone surpassing it by 18% as is the case with your results.

Edit: Either your numbers are switched (which is very possible indeed) or timedemo1 isn't an accurate indicator of typical Doom 3 performance.

 

very interesting chart. so now i wonder if it was that they got the nvidia and ati numbers switched instead of what i thought; the low resolution messed up the results.

 

No, 85Hz is acceptable from a refresh rate POV but given the hardware, 1280x1024 is not particularly acceptable IMHO.

 

Hey guys,

You've all raised some brilliant points - thanks for the comments. Let me try and give you some more info on the various things that you're talking about ...

1) Resolutions. What I was trying to show with those graphs was the extra gain from having CrossFire enabled vs SLI enabled. Comparing a 6800GT reference board to a X850 XT reference board is a pretty pointless exercise, since all board partners clock their cards differently anyway. By using a fairly low resolution, I was trying to see if I could get those theoretical performance increases out of the technology.

You are totally right that most gamers won't play with this hardware at 1024. However, it is also true that most timedemos aren't representative of real gameplay - hence why we conduct our graphics card reviews the way we do.

The upshot of this is that we said this is a preview. Tim and Rich will be looking at retail products in full depth very, very soon, and that's when we'll really get a handle on the performance.

Some sites have gone all out to measure performance this week. I don't believe that's a good thing. Number one, CrossFire is obviously still broken in parts - for example, the universal supertiling. I think ATI need more time to fix it, and that will result in differing performance.

Two, we believe that real gameplay evalluations are the best way to test graphics setups, and we simply didn't have time to do that. We had literally one day with the CrossFire system to put together that article, and unfortunately, there are areas where those constraints show - the lack of gameplay evaluations is one.

2) The 7800 GTX and ATI's 'conditions' for review. ATI made it clear to us, and a number of other UK publications, that did not believe that 7800GTX was a competitor to X850 Crossfire and so they did not want it compared. This was a condition of getting the kit for review.

I agreed to this primarily a) because it was so blindingly obvious that 7800GTX canes CrossFire I could just put that in the text b) because the major comparison is cost, and I could still put it in there c) I figured the benchmarks would not be comprehensive anyway and d) Everyone was in the same boat.

It does now appear like ATI changed the rules for certain online mags, or those mags chose not to stick to them. Frankly, good for them - for manufacturers to try to dictate how stuff is reviewed is pretty abhorrent imho.

I also believe that the lack of focus on '7800 just canes everything' means that we can take a more balanced look at CrossFire itself, which was pointed out earlier on in the thread.

All in all, I do believe this was a good article. It was fair, and whilst it was short on performance benchmarks, it did answer the crucial questions - is it any good, and should I buy it? (with a 'sort of' and a 'no', respectively). However, as we said, this is not retail product and it would be stupid, imho, to go all out on reference platforms with iffy drivers. Maybe I should add this as a disclaimer at the beginning of the article.

 

I agree. This was never going to be the definitive review (on any site), it was just meant to fill in a few gaps. Other sites have obviously tried to get a full review from incomplete pieces in order to get a scoop.

 

1600x1200 60hz makes me cry
tbh i never really liked the concept of 2 graphics cards, always seemd a bit daft to me
i want to see what r520 and r580 do

 

Well stated, and thank you for deciding not to be a sheep who tries to release the full scoop with every figure imaginable, most of them devised from RC drivers and a demo board.

It makes you wonder about some of the other sites, who have 15 pages of full review...how did they get time to test all this crap?

By the way, ATI's practices seem to be getting shadier and shadier...demanding reviews be done in small time spaces and only compared to certain cards...

 

Its called control conditions; its obvious, when critical thinking is applied, that they are trying to give their products better appearance at face value. This generation will require a lot of scrutiny to get an accurate picture of just what is going on. Why cant people just be straightforward instead of playing these games of cloak and dagger?

 

If your product is inferior in a straightforward light, then straightforward is the last thing you want. A shame, really...this is not the only gripe I have with ATI's PR lately, and I even was looking really forward to Crossfire. It's a shame to buy a good product with bad business practices, but worse when it's not that great a product to start...

 

i personally like very much this review (and not only).
It's fair, answer questions that users should have and easy to read and understand.
I can't understand why users compares the results with other articles.. This is wrong for a lot of cases. It depends on the hardware/software the way you test the system, even if you see double fps for a card in another article you must NOT compare the results (and this has happen a lot of times)

 

talking about ATi and telling reviewers not to compare the Crossfire to the 7800..well I bet they also dont want people to start comparing warranties either. http://neowin.net/comments.php?id=30728&category=main . It does not look like Ati is the kind of company I would deal with IMO, ill stick with my EVGA 7800GTX with lifetime warranty thank you very much.