Anand Review: ATI Radeon HD 3870 X2

[Krizby87]

Techreport (use a X38 chipset) and Guru3d's reviews also show 8800GT SLI pwning 3870x2 left and right, stop seeing only what you want to see (ex: some benchmarks from some chinese review site), it's clear that 8800GT SLI is faster. The HD 3870X2 is a nice product, provide power for the masses but hardly an enthusiast product with a measly overclocking headroom.

[Scubar]

One thing that is good is that the 8800GT in SLI outperforms the 3870X2 which means that hopefully the 8800GX2 coming out should be abit more powerful than the SLI'd cards as they will have had time to optimize card design and drivers.

[mascaras]

Countdown until Nvidia decides to releases whatever monstrous ace they have up their sleeves?

why every thread about ATi card ends talkiing about Nvidia and Vice-versa????



btw : nice review , now we need a better driver (same old story)

my HD3870X2 arrive in 2 days , i will post some tests later (in review section)


regards

[munim]

To summarize Anand's testing methodology:

Bioshock: "Our Bioshock test involves a quick run through a section of the Medical Pavillion. Since the game has no built-in benchmark, our runthrough is manual and we rely on FRAPS to measure the average frame rate. All of the enemies are already dead in the area so we don't run the risk of running into one of them during our test."

CoD4: "lacks any sort of in-game benchmark so we benchmark the cut scene at the beginning of the first mission"

Crysis: "Crysis ships with a built-in GPU benchmark"

Enemy Territory: "We used our island benchmark for ET: Quake Wars as always"

UT3: "We used the built-in vCTF flyby in Unreal Tournament 3"

Witcher: "ran FRAPS during the game's first major cutsceneat the start of play"

Take it how you will

P.S. I don't know what the island benchmark is in Quake Wars, and I don't care

[Morais]

argh no CF results

[Syn.]

argh no CF results

errmm you mean QuadCF results? Those can only be done once drivers are there that support it. Wait a little while more m8, they are comming.

[mascaras]

argh no CF results

because theres NO catalyst Driver yet with suport for HD3870X2 crossfire !

Vrozne says the new catalyst Oficial will be released 30th Jan. >> http://www.vr-zone.com/articles/Rade...ance/5530.html


However no mention if the new driver have CrossfireX support ....


regards

[Dave_Sz]

Countdown until Nvidia decides to releases whatever monstrous ace they have up their sleeves?

thank u for that useless contribution of NVIDIA fanboyism...

[Polizei]

They say 8800GT SLI is $150 more than an HD3870x2. I see prices around $480 for the HD3870x2, while I can get an 8800GT for $260.

Try more like, $40 more expensive.

[Hockster]

[H] gets in on the review action as well.
http://enthusiast.hardocp.com/articl...50aHVzaWFzdA==

[zanzabar]

[H] gets in on the review action as well.
http://enthusiast.hardocp.com/articl...50aHVzaWFzdA==

theirs reflects the http://en.expreview.com/?p=219 review and if u compair the expreview 8800gt sli were the ultra was 1% faster so the 3870 x2 should be about 5% faster

[safan80]

Quake Wars is a joke with this card

[grimREEFER]

i gotta say, its their first nice card in a while. but its depressing that technology has made so little progress. the performance is basically that of 2900xt crossfire 9 months ago, albeit it is much cheaper now.

[Scimitar]

Quake Wars is a joke with this card

That review used the old drivers must be. Here is Techpowerup's summary of the HD3870X2's performance in QW.

http://www.techpowerup.com/reviews/H...3870_X2/8.html

[Jodiuh]

Aside from [H], who else actually sits their a$$ down for some time to PLAY the game and use Fraps to show min's and avg's?

Even the ATI release notes Anand listed talk about improving "Island GPU test up to 1920x1200." IMO, pick 2-4 cards that compete in same $ range and play some games. Then tell me which card did a better job. Throw some noise/power/driver/$/avail info in there and call it a day.

[munim]

Aside from [H], who else actually sits their a$$ down for some time to PLAY the game and use Fraps to show min's and avg's?

Even the ATI release notes Anand listed talk about improving "Island GPU test up to 1920x1200." IMO, pick 2-4 cards that compete in same $ range and play some games. Then tell me which card did a better job. Throw some noise/power/driver/$/avail info in there and call it a day.

Bit-tech does.

[disruptfam]

nice...

[disruptfam]

if this review trustworthy ???

Im looking @ selling my gtx if these results are actually true...

[DTU_XaVier]

To summarize Anand's testing methodology:

Bioshock: "Our Bioshock test involves a quick run through a section of the Medical Pavillion. Since the game has no built-in benchmark, our runthrough is manual and we rely on FRAPS to measure the average frame rate. All of the enemies are already dead in the area so we don't run the risk of running into one of them during our test."

CoD4: "lacks any sort of in-game benchmark so we benchmark the cut scene at the beginning of the first mission"

Crysis: "Crysis ships with a built-in GPU benchmark"

Enemy Territory: "We used our island benchmark for ET: Quake Wars as always"

UT3: "We used the built-in vCTF flyby in Unreal Tournament 3"

Witcher: "ran FRAPS during the game's first major cutsceneat the start of play"

Take it how you will

P.S. I don't know what the island benchmark is in Quake Wars, and I don't care

As long as the cards are tested under exactly, or as close to exactly, the ame circumstances and through the same sections of a game, I have no problem whatsoever with them doing it like that...

[Unbornchild]

Quake Wars is a joke with this card

Why? Notice the 2560x1600 resolution and highest settings. Your post is a joke. 54 fps is enough. Anything above is nice but not needed. A card "X" that would have at least 50 fps in every single game at that resolution would be a winner over any other card that has 150+ fps in one game but less than 30 in another.

[aintz]

so you still cant play crysis on high settings smoothly. fail

[grimREEFER]

so you still cant play crysis on high settings smoothly. fail

yea, it performed worse than the 8800gtx according to hardocp. at least that card had the excuse of coming out before crysis was released.

[Ace123]

Dude, When did we start thinking that ATI was going to stomp the yard with this card?
As far as I knew, this was the latest greatest technology, because of the POWER IT SAVED while still producing a comparable performance ratio to a card already released using TWICE the power. Now we get crossfire on one card for people who dont have dual slots, and we still complain?

Come on guys.

[JohnZS]

It's a mixed bag really. I've read a few reviews and there is no doubt it is a serious bit of kit and does perform well in most games, however it has 3 major flaws.

1) Power Draw.... it is one hungry power sucking card. (Perhaps an external power source ala Voodoo5 6000 might be a good idea?)

2) Crossfire, yes I know not all will agree with me, but whether we like it or not Crossfire does require a bit of praying to the Driver Gods (as does SLi) for a game to work correctly

3) it's not that much faster than the single GPU solutions from nVidia which have been out for well.....over a year now?

Aside from those 3 blemishes I would say the HD3870X2 is a good card, but I am not going to buy one... I am happy with my 8800GTS 512MB and will hang onto it until the real Next Gen come out (so probably keep it until October or something).

I think we may see more power efficient, clever and serious polygon pushing GPU's coming out from both ATi and nVidia later this year....which would be nice.

Until then it is sort of like the hack job 9800GX2 Vs the pretty clever yet brute force HD3870X2.

John

[Jodiuh]

...I would say the HD3870X2 is a good card, but I am not going to buy one... I am happy with my 8800GTS 512MB and will hang onto it until the real Next Gen come out (so probably keep it until October or something).

Me too. COD4 plays fine. I've zero desire to kick up power usage and heat for more performance in the only 3 games that already play fine for me...also IMO canned benchmarks fail when compared to actually playing the game and giving a subjective opinion like we do here.