Zardon

Read The Review Here
___________________

A few weeks ago Nvidia had one of its most successful launches ever with the 8800GT. Reviews across the web, including ours, were incredibly positive with the excellent performance at a midrange price. At the time of our original review the XFX 8800GT XXX edition was not ready and so we featured only the standard edition. It was a competitive part but not as outstanding as we have come to expect from XFX so when the XXX edition arrived we were quick to put it through some benchmarks and today we will be showing just how well it fared.

As always in such a short space of time, the marketplace has changed and the direct competition to the 8800GT is no longer the 2900XT/Pro, it is the Radeon HD 3870. There is still a gap in the ATI product line-up with no single card flagship enthusiast product; instead ATI's stance is to crossfire the 3870's and with a recommended price of $219/£160 each that is not too outrageous.

Our review will be covering comparisons between the 8800GT XXX, 3870, 2900 XT and stock 8800 GT using the latest games and high definition movies. We will also cover 3870 CrossFire, overclocking, image quality and whether it is possible to connect a 3870 and 2900 XT in a working CrossFire system.

josh2thad

awesome review. We all knew the 8800GT was faster...but dang i wasn't thinking that much faster. i wanted to crossfire 2x 3850's but its hard to do that when you can now see that the performance of the 8800gt alone would be about the same as 2 of those cards.

__________________

ChaosMinionX

Great review guys, hopefully with some driver releases the 3870 gets better as well.

Would have been nice if you had 2 8800GT's to compare CF/SLI of the new ones, a couple reviews already show that in multi card setups that the 3870's are scoring better than the GT's.

But none the less was worth the wait to publish the review, and was a good read

__________________

Dark Lord

any chance well be seeing a quad crossfire 3870 review some time soon?

__________________

swimtech

Outstandingly thorough review Veridian3, thank you!

Yeah, it seems the 8800GT is even faster in comparison to the 3870 when more eye candy is applied than is apparent from other reviews. That's a Crysis indeed...

__________________
It's not so much getting your way that matters or not - what matters is how you go about getting it.

geraldmango

erm, there were pages of crossfire testing, and even the fact that DH found bugs which AMD didn't and now have even opened offical support tickets on their site, id say it was worth the wait.

I think they didnt include SLI because there was no need, two 3870s in Crossfire had a hard time beating a single overclocked GT in most cases (when they worked).

Veridian3

considering a single XXX is somewhat competitive with the 3870 Crossfire (when it works) i think its pretty safe to assume that the SLI would win.

In the current state 3870 crossfire against SLI in a review would be pointless seeing as many games dont work at all. (See the Crossfire/Ugly page).

ChaosMinionX

A single GT is not going to beat dual 3870s.... just not possible, and in several known reviews of a comparison between the 2 3870s performed better than 2 GT's in crossfire/SLI.

And it was an overclocked GTX not a GT, that was compared to the CF setup.

EDIT:
What driver release was used? I have no problems with HL2 crashing, or any of the source engine with crossfire enabled on 2 3870's. I was using 7.11 with Hotfix installed overtop.

I would link the crossfire/sli review, but I wouldnt want to derail this thread

EDIT2: I am also not stating that a single HD3870 is currently faster than a Nvidia 8800GT, however that crossfire 3870's are faster than crossfire GT's in the majority of games.

__________________

Zardon

chaos, not my review however I will comment on a few things and let stuart reply to the rest. (some of your posting I dont even understand to be honest).

Firstly, i am in no way interested in other reviews on the subject, there isnt a good crossfire review up on the net right now, and the fact we were reporting issues no one else did (confirmed by AMD), which AMD verified due to our thorough testing speaks volumes. Yes AMD found problems due entirely to DH testing. This alone makes you wonder what exactly reviewers do when they "test" these cards to get them published in time for NDA, obviously to get hits.

I made the call to hold publication of the review because we ran into so many nagging problems, even at the expense of big hits on launch day. We ending up working with AMD HQ and the driver development team and finding out just what was going on. Full credit to AMD for opening their doors to us as well.

Your Half life system might be perfect for you, however again with our testing AMD successfully reproduced the problem and opened a report AMD Customer Care. The fact you don't have a problem means really very little in the grand scheme of things and this is why it is a very hard job delivering a great review, you need to go one step further than merely just "playing" a few games to get accurate results.

This is exactly why our GFX reviews get such high hits on the net, people wait for them to see the problems/findings we are reporting, because quite frankly very often no other site is capable of finding them. It just shows how thorough Stuart is, and we have no agenda to "kiss ass" with companies either which is abundantly clear with this review. All packages on test today have issues and we made sure everyone knows it.

Will it cause problems for us in the future? I don't care, because everything we published is exactly how we found it. Not how any company would like us to find it.

riles9262

Great review guys, one of the best thus far imo. Very detailed.

__________________
A beer a day keeps the doctor away.

slowhandluke

great review.

ChaosMinionX

Understood, I was just saying it has been confirmed a few places that crossfire was working better than SLI on the 3870s, than GT's... It was a fantastic review and very indepth, I take most reviews from other sites (besides DH and endusers personal experiences) with a grain of salt, but the reports on Crossfire/SLI comparison between the 2 new cards was that the 3870s were faster in a dual GPU setup when compared to GT in dual card on the majority of games and especially with higher filtering added. And I was just wondering about a few things regarding some of the crossfire bugs

I know there has been a thread detailing these differences, GT excels in some games, and then 3870's beat GT in other games. I was just making a case for 3870 crossfire, and it being a great buy for people looking for dual graphics, and that ATI is slowly coming back at Nvidia, while the single card comparison GT is almost always going to beat the 3870 by a few fps, but that a dual card setup the 3870s actually handled some games better than the GT especially with higher AA and AF.

__________________

[hobo]eclipse

Great review, IMHO if the 8800GT had a cooler that expelled the air from the case rather than into it it would be a no brainer, that being said anyone see a drastic difference in the IQ comparison screen shots in crysis? everything looked pretty much identical except for the mountain in the backround to the right in these 2 screen shots. the ATI and the nvidia screen shots. the nvidia shot has the mountain being way lower res it looks like from my standpoint.

__________________

brutusmaximus

Are your views colored by the fact your profile has you having two of these cards? Neither CF or SLI has worked out as a good solution for the end user, I think we have users on both sides of the field extremely unhappy with their configurations. if you look on any forum you have nvidia users bitchfesting about SLI in vista, and on the other hand its clear from this review, that unless you play 3 or 4 games, that crossfire has more flaws than selling points. You are even getting LOWER performance with forced AFR in some games without profiles than you would be running it in single card mode. These aren't just personal opinions of the review on DH either, these are findings which AMD has verified !

Yes you aren't alone, I have read many a rushed review on launch day online, but I always tend to wait for DH to report the facts as they frequently seem to miss launch day intentionally to get a more thorough article up. I also have major respect for Veridian3's reviews. They aren't boring like some other sites but they are fantastically indepth in the areas im interested in. Many other sites don't even bother with HD testing which is ridiculous.

Well I have no idea who has said this because its clear there aren't even CF profiles for a lot of the key games, and the fact that a single XXX is beating a CF rig even when its working is quite concerning.

yeah me too, ive been reading the AMD reports detailed on the CFUGLY page that DH published. great work DH saying it like it is, I also hope AMD haven't pulled support here for that, its the first time ive seen a site finding bugs, then reporting them, then waiting on AMD to verify them before publishing. Top drawer stuff and respect needs to be given.

EPR 224330 Oblivion: CrossFire. 2x 3870. Flickering when the game first loads : 737-30750:

EPR 224331 LOST PLANET. CrossFire 2x RV670 corrupt text with CF enabled. Water in caves is also a bit corrupt. 737-30751:

EPR 224344 World In Conflict: CrossFire. 2x 3870. Driver crashes (atikmdag error) 737-30752:

EPR 224345 Half-Life 2: Episode 2: CrossFire. 2x 3870. Random crashes to desktop, atikmdag error 737-30753:


well unless you are playing COD4 or Crysis thats about as strong a case as you will make. Certainly until ATI sort out their drivers and profiles anyway.