Got my 3870x2 setup running finally....

[BOBDOLER]

Blah, most frustrating computer I've ever built, one of my stick of ram and my new PSU were both freaking busted, then after I got my setup going I couldn't figure out why Crossfire wasn't working for 3 days! I was trying everything I could think of, tweaking, reinstalling stuff and I just right now found out that.

CROSSFIRE WON'T WORK IN WINDOW MODE YOU MUST BE AT FULL SCREEN :rofl:

You see I had I had just kept running games in window mode so I could see the graphs for my second GPU activity. Just wanted to share that with anyone who might do the same thing I did :eek:

Hopefully someone gets some high AA tests out between the 3800's and 8800GT because I really think from what I've been seeing the 3800's will destroy the 8800GT when you get into higher AA's

computerbase has an article up were 3870 vs the 8800GT the 3870 came back from behind in 3 of 4 of the tests they used 8xAA on to overtake the 8800GT and on the one test it didn't it it should be noted it didn't loose ground unlike the 8800GT which dropped like a rock at 8x

I just had to throw that in there for you Rollo :rofl:

I'll do more testing and post results later, on the default 3Dmark06 settings though I got 18000 on stock GPU clocks and the Q6600 at 3.4Ghz.

Amazingly upping the resolution to 1680 by 1050 only knocked off a few hundred points

[Rollo]

Hopefully someone gets some high AA tests out between the 3800's and 8800GT because I really think from what I've been seeing the 3800's will destroy the 8800GT when you get into higher AA's

computerbase has an article up were 3870 vs the 8800GT the 3870 came back from behind in 3 of 4 of the tests they used 8xAA on to overtake the 8800GT and on the one test it didn't it it should be noted it didn't loose ground unlike the 8800GT which dropped like a rock at 8x

I just had to throw that in there for you Rollo :rofl:



You should have linked that article, I like how it shows the difference between 3870s and 8800GTs:

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/8/
At Jericho 12X10 4X16X, the 8800GT beats the 3870 37fps to 24fps, a 54% performance advantage. Both cards are totally unplayable at 8X AA, at 19fps and 21fps respectively.

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/10/
At COH, the 8800GT was 5fps faster than the 3870 at 16X12 8X AA .

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/11/
At FEAR the 8800GT is 20% faster at 16X12 4X AA, and the same speed at 8X AA.

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/12/
At Gothic 16 X12 0X16X, the 8800GT is 81% faster than the 3870, 37.6 vs 20.8 fps.

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/13/
The 3870 does squeek out a 2.4fps victory at Oblivion, but 51.3 vs 48.9 wouldn't really be discernible in game play.

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/14/
Prey is a tie, with a .1 advantage to the GT at 4X and 16X inh 16X12. (12X10 does give the 3870 a decent advantage)

Anyway- you should have linked that article- good benches!

[BOBDOLER]

You sure know how to spin this stuff, :rofl: seriously that COH example was just pathetic... :rofl:

Did you miss the fact that they also tested COH in DirectX 10? Lets put this in perspective buddy, if you have an 8800GT or a 3870 are you going to freaking run COH in DX9? No your going to put it on DX10 and in DX10 the 3870 is faster by a big 23% on a game that I believe has the Nvidia logo on the startup screen.

here (http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/19/#abschnitt_company_of_heroes)

Way to change the subject to Nvidabot, I clearly was discussing the AA performance since trolls like you spam that the 3800's just blow when it comes to AA. The gothic test isn't even including AA.

And in the Oblivion test it wins by 11% I wouldn't exactly call that a squeak (btw you spelled squeak wrong..)


In prey at 16x12 at 8xAA the 3870 wins by 20%

Your right FEAR is the one game it does win 8x at unless you want to count the DX9 version of COH, idk why you would run that with these cards though... the thing I took away from FEAR's test though was that at 8X the ATI card didn't loose ground so in other words we are seeing the 3870 consistently do better at 8xAA


Now add the fact that the 3870 scales far better in crossfire and I think the 3870 will clean house at high levels of AA in cross/SLI like 8 and 16x





And while your at it since you wanted to go a bit off course and mention ATI's gothic performance how about taking a look at the 88 on DIRT :shocked:

Wow is that bad the

3870 gigantic 251% faster then the 8800GT in DIRT at 16x12 with 4xAA link (http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/9/#abschnitt_colin_mcrae_dirt)


:eek: what makes it even worse is the fact that the 8800GT was actually WINNING by 21% with 1xAA ouch thats an epic hit right there.


Hopefully someone gets out some 8x/16x AA benchmarks with these two in cross/sli so we can see a real test..

I got TF2 running at 16xAA at 40 min to 100 frames a second at 1680x1050 with Super sampling maxed out on top of that on 32 player servers, I don't know if I'd be doing that with 8800GT's in sli tbh.

Oh and my bad their were actually 5, 8xAA tested games and the 3870 won in anno 17 at 8xAA as well by 14% while the 8800GT had a hefty 27% lead before AA was applied

[zanzabar]

rollo way to bring up a bench that shows that the 680i at stock speeds beats the 975 and how limiting the bandwidth on pci-e2 cards affects them. not to mention the most nv biased games next to lost planet (and strange results on the others)

[BOBDOLER]

Oh snap your right zan! They aren't even running a proper CF mobo!


Even in the Nvidia favored lost planet the 3870 comes out on top with AA at 16x12 with no AA the 8800GT has a massive 50% lead, even at 4xAA it still has a hefty 37% lead, but once theirs 8xAA it's ends up loosing by 15% OUCH

These benchmarks show a clear pattern of the 8800GT taking big hits at 8xAA

ON top of that I was just looking through all the screenshots and the only one so far that I've noticed a DRASTIC image quality difference is the Lost Planet screens were surprise ATI's image quality is significantly better.

Look for yourself

Nvidia notice how awfully low rez the building textures look

http://pics.computerbase.de/1/8/3/5/0/38.jpg



ATI look at how much sharper they are here, on top of that the smoke coming out of the building looks better as well.

http://pics.computerbase.de/1/8/3/5/0/39.jpg

[demowhc]

i dont know if its just me but the whole ATI screenie looks sharper, especially the building you pointed out...

But i would like to point out something.. look at the missing 'fur' on the players collar.. it has no 'fur' in ATI screenie where as you can clearly see it in the Nvidia pic!


btw bob i take it you have 2 x 3870 in CF and not a 3870x2??

[Lord AnthraX]

Red building is defiantly more sharper and defined in the ATI picture.

Mountains off the the left also have sharper textures.

[Rollo]

You sure know how to spin this stuff, :rofl: seriously that COH example was just pathetic... :rofl:

Did you miss the fact that they also tested COH in DirectX 10? Lets put this in perspective buddy, if you have an 8800GT or a 3870 are you going to freaking run COH in DX9? No your going to put it on DX10 and in DX10 the 3870 is faster by a big 23% on a game that I believe has the Nvidia logo on the startup screen.
Spin from me? I posted resolutions and framerates. You referenced a performance advantage at a setting neither card is even close to playable at. I think I can bet no one is playing at 20 fps average to realize your "23% ATi advantage". That would be more like "frame counting" than "gaming". :rolleyes:



here (http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/19/#abschnitt_company_of_heroes)


Way to change the subject to Nvidabot, I clearly was discussing the AA performance since trolls like you spam that the 3800's just blow when it comes to AA. The gothic test isn't even including AA.
Aren't personal attacks a violation of TOS on this board BobDoler?
5/6 benchmarks I referenced were with AA, so I was clearly discussing AA as well, and proving you wrong.:)


And in the Oblivion test it wins by 11% I wouldn't exactly call that a squeak (btw you spelled squeak wrong..)
Framerates would have been useful here too- 35.7 vs 39.7 isn't exactly a "knockout" either.


In prey at 16x12 at 8xAA the 3870 wins by 20%
And what I said about Prey is 100% accurate-
Prey is a tie, with a .1 advantage to the GT at 4X and 16X inh 16X12. (12X10 does give the 3870 a decent advantage)


Your right FEAR is the one game it does win 8x at unless you want to count the DX9 version of COH, idk why you would run that with these cards though... the thing I took away from FEAR's test though was that at 8X the ATI card didn't loose ground so in other words we are seeing the 3870 consistently do better at 8xAA
No, not really.


Now add the fact that the 3870 scales far better in crossfire and I think the 3870 will clean house at high levels of AA in cross/SLI like 8 and 16x
Except it's not a "fact" 3870s scale better, it depends on the game, and often even when they're "scaling better" they aren't exceeding 8800GT SLi in performance. You may "think" they will "clean house" but you haven't posted one CF vs SLi benchmark to prove it. I care more about benchmarks than supposition personally.


And while your at it since you wanted to go a bit off course and mention ATI's gothic performance how about taking a look at the 88 on DIRT :shocked:

Wow is that bad the

3870 gigantic 251% faster then the 8800GT in DIRT at 16x12 with 4xAA link (http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/9/#abschnitt_colin_mcrae_dirt)


:eek: what makes it even worse is the fact that the 8800GT was actually WINNING by 21% with 1xAA ouch thats an epic hit right there.
Another totally meaningless reference by the guy who "doesn't spin". Is 15fps average more playable than 4fps BobDoler? Is having an average framerate at half what's generally considered necessary for smooth animation worth even running? Could a person even play that?


Hopefully someone gets out some 8x/16x AA benchmarks with these two in cross/sli so we can see a real test..
We won't know till they do.


I got TF2 running at 16xAA at 40 min to 100 frames a second at 1680x1050 with Super sampling maxed out on top of that on 32 player servers, I don't know if I'd be doing that with 8800GT's in sli tbh.
That's not surprising as you don't have GT SLi. Do you have a link to GT SLi playing that game worse than 3870 CF, or are we just supposed to believe that because you "don't know if you'd be doing that"?



Oh and my bad their were actually 5, 8xAA tested games and the 3870 won in anno 17 at 8xAA as well by 14% while the 8800GT had a hefty 27% lead before AA was applied

TBH- I wouldn't call either of these cards "8X" gaming cards- but I like to play games, not watch them stutter.

[Rollo]

rollo way to bring up a bench that shows that the 680i at stock speeds beats the 975 and how limiting the bandwidth on pci-e2 cards affects them. not to mention the most nv biased games next to lost planet (and strange results on the others)

BobDoler pointed me at the article, I just recapped some of the benches contained therein that cast a bit of doubt on his assertions.

As far as "nV biased games" go- games aren't "biased".

[tiro_uspsss]

Aren't personal attacks a violation of TOS on this board BobDoler?

so is trolling, which imo ur doing by entering a thread clearly intended for those who like a bit of ATI action - so stfu & gtfo

edit: btw i've reported u this time

[Lord AnthraX]

As far as "nV biased games" go- games aren't "biased".

Games aren't biased? Anything to back that up or is that just something you assumed.

I'm pretty sure some games are designed to run on certain cards better then others. I'd call that biased, wouldn't you?

Examples. Call of Juarez, Orange Box, I mean those are just a few that are "ment for ATI", and both show better framerates with ATi cards, not to mention you have more resolutions to choose from in each aspect ratio in Source games with ATi, and the nightvision is MUCH much clearer versus an Nvidia card. I'd call that biased.

[SNiiPE_DoGG]

BobDoler pointed me at the article, I just recapped some of the benches contained therein that cast a bit of doubt on his assertions.

As far as "nV biased games" go- games aren't "biased".

yeah, but you refuse to acknowledge just how pathetic it is that the 3870's in 16x/4x can beat or come close to the gt's which are on 16x/16x, and are supposedly SOOOOO much better, and scale SOOOOO much better :\

[Lord AnthraX]

yeah, but you refuse to acknowledge just how pathetic it is that the 3870's in 16x/4x can beat or come close to the gt's which are on 16x/16x, and are supposedly SOOOOO much better :\

Word.

[demowhc]

in COH @ 12x10 i see 3870 4xaa 32fps... beating the GT, also at 8xaa it says 27fps..also beating the GT..infact the only time the GT beats it in COH is at 1xaa :rolleyes: and i thought the 3870 was meant to suck at aa...

and then the 3870 beats it in every 16x12 test.. according to the link you provided.

http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_3850_rv670/19/#abschnitt_company_of_heroes

[BOBDOLER]

Your right alot of the games aren't running at playable frame rates, but your not seeing the significant part of this, the thing I see looking at these numbers is the 8800GT just doesn't seem to have the power to push out the higher AA, well what about fear? What about it? that games a practically all indoor so you would have to go up to 16x it seems for the 8800GT to start choking.

Clearly all these big drops from the 8800GT once 8XAA is reached is a significant find. Logic would suggest the pattern indicates that the 8800GT just doesn't have enough of something when you get to that high of an AA level. Whether it's shader bandwidth or whatever is beyond the point, the point is this will translate to SLI so while it may not matter here with these unplayable frames it will matter when you have two cards.

Now this is an assumption, but based on the circumstantial evidence if one were to buy a 8800GT/3870 x2 for crossfire/sli until their is a clear test published showing 8x/16x results logic suggests that the 3870's will be the victor and other then Crysis with SLI/Crossfire your actually going to want to and should expect to be able to run most games at, at least 8x AA.

oh and aren't you going to comment on the lost planet screens?

[demowhc]

ill comment on the lost planet screens..what about the missing fur in the ATI screenie! :shakes: i dont know why but that really bothers me..

apart from that i agree with you the ATI screenie is much sharper, even look at the fences etc, the barbed wire is much clearer etc.

[Vapor]

Rollo, BOBDOLER, please stop trolling...

Yes, you're both trolling and targeting each other...no more.

Final warning.

[Rollo]

Games aren't biased? Anything to back that up or is that just something you assumed.

I'm pretty sure some games are designed to run on certain cards better then others. I'd call that biased, wouldn't you?

Examples. Call of Juarez, Orange Box, I mean those are just a few that are "ment for ATI", and both show better framerates with ATi cards, not to mention you have more resolutions to choose from in each aspect ratio in Source games with ATi, and the nightvision is MUCH much clearer versus an Nvidia card. I'd call that biased.

What I meant is games aren't "biased"- they have no opinions.

Games are coded in ways that some work better on one GPU design than another, so we're on the same page.

[oneeyedollar]

ill comment on the lost planet screens..what about the missing fur in the ATI screenie! :shakes: i dont know why but that really bothers me..

apart from that i agree with you the ATI screenie is much sharper, even look at the fences etc, the barbed wire is much clearer etc.

i agree totally. when comparing the pics i was impressed with the detailed texture quality and sharpness in the ati image, but as soon as i saw that the nvidia one has that highly detailed fur and the ati didn't i declared the evidence inadmissable :p:

[oneeyedollar]

here's the 8800GT SLI vs HD 3870 CS review

it's translated from chinese or something i think, so the wordings a little weird, but here it is. i'd like to see a newer review though, as the aa doesn't seem fixed for ati in this review.

http://en.expreview.com/?p=53&page=1

[newls1]

Blah, most frustrating computer I've ever built, one of my stick of ram and my new PSU were both freaking busted, then after I got my setup going I couldn't figure out why Crossfire wasn't working for 3 days! I was trying everything I could think of, tweaking, reinstalling stuff and I just right now found out that.

CROSSFIRE WON'T WORK IN WINDOW MODE YOU MUST BE AT FULL SCREEN :rofl:

You see I had I had just kept running games in window mode so I could see the graphs for my second GPU activity. Just wanted to share that with anyone who might do the same thing I did :eek:

Hopefully someone gets some high AA tests out between the 3800's and 8800GT because I really think from what I've been seeing the 3800's will destroy the 8800GT when you get into higher AA's

computerbase has an article up were 3870 vs the 8800GT the 3870 came back from behind in 3 of 4 of the tests they used 8xAA on to overtake the 8800GT and on the one test it didn't it it should be noted it didn't loose ground unlike the 8800GT which dropped like a rock at 8x

I just had to throw that in there for you Rollo :rofl:

I'll do more testing and post results later, on the default 3Dmark06 settings though I got 18000 on stock GPU clocks and the Q6600 at 3.4Ghz.

Amazingly upping the resolution to 1680 by 1050 only knocked off a few hundred points

Just wondering what motherboard your using. I just ordered 2 3870's and a ASUS Maximus for dual 16x action. Just wondering what im in for.

[BOBDOLER]

Goto the intel section of this site, you'll find many people with the Maximus and a super long thread on it..it seems to be the most popular X38 right now, but I went with the Abit X38 QuadGT..

[Ethelred]

here's the 8800GT SLI vs HD 3870 CS review

it's translated from chinese or something i think, so the wordings a little weird, but here it is. i'd like to see a newer review though, as the aa doesn't seem fixed for ati in this review.

http://en.expreview.com/?p=53&page=1

Extremely interesting what they had to say...does go to show that CS does scale alot better than the 8800gt SLI. They said that the 3870 in the review costs the same as the 8800gt, however, that doesn't carry over very well. with the 3870 at about 240(?) about 50 less than the 8800gt it seems to be a much better option as well as for longevity to simply CS them rather than constantly upgrade.

Seemed very objective at the very least.

[Rollo]

Extremely interesting what they had to say...does go to show that CS does scale alot better than the 8800gt SLI.

The thing is, even if it gets higher scaling, it's not a "better" solution if you take for granted people don't buy 2 video card rig with the hope of not running AA.

If you only look at the game benchmarks, and use AA where possible, the 8800GT SLi wins 7, the 3870 CF wins 6.

So even if the amount of scaling is higher, the amount of performance isn't.

IMO, as long as the performance is equal, you'd want the card with less scaling, because that card started out higher to begin with, and not all games scale.

[Ethelred]

If you only look at the game benchmarks, and use AA where possible, the 8800GT SLi wins 7, the 3870 CF wins 6.

So even if the amount of scaling is higher, the amount of performance isn't.

IMO, as long as the performance is equal, you'd want the card with less scaling, because that card started out higher to begin with, and not all games scale.

We're talking by one maybe two frames...while the 8800gt does perform better...it is only by the slightest margin. And while the 8800gt does perform better stock, the 3870 has IMO far more potential than the GT.

Yeah, we all want the card that doesn't scale as well...if I'm running CS or SLI I would be angry if my card was scaling in much the same way the GT is. And while not all games scale, I think the ones that do are the only ones that matter.

[Rollo]

We're talking by one maybe two frames...while the 8800gt does perform better...it is only by the slightest margin. And while the 8800gt does perform better stock, the 3870 has IMO far more potential than the GT.

Yeah, we all want the card that doesn't scale as well...if I'm running CS or SLI I would be angry if my card was scaling in much the same way the GT is. And while not all games scale, I think the ones that do are the only ones that matter.

Why do you think the R600 arch based 3870 has "far more potential"? It's currently the 5th or 6th best card you can buy? What do you base this on?

Scaling % is irrelevant, end result is all that matters. If one card starts 25% faster, scales another 25%, and the other scales 50%, you still get the same framerate, except in games that don't scale, where you're better off being 25% faster to start with.

While you may think the games that don't scale "don't matter", tell that to the guys with CF who have to disable one card to play some of their favorite games because CF degrades performance. It's a very big deal to a lot of current and ex CF owners, check it out on Rage3D, the ATi fansite if you don't believe me.

[LaMpiR]

I thought of going with 2x3870 but after latest review on hothardware i saw that ATI with it's crossfire can't go near new 8800GTS 512? So is that about drivers or just because they suck?

[Jakko]

Rollo, your anti ati rambles are getting old.
The 3870 is cheaper, gets less hot, and uses less power.
It's overclocking potential is better, so is its scaling.

Single card, 8800 GT performs somewhat better than the 3870, yet do you know of any game where the difference can be noticed?
CF/SLI, the 3870 performs on par with or better than dual 8800 GT, yet also in this case it's hard to really tell the difference.

So if these two cards roughly perform likewise, what's the most important remaining factor?
To me it would be price(ati wins) or upgradability(ati wins again).
I doubt any of your nvidia propaganda is going to make me change my mind.

[AliG]

not sure what you're talking about, the xfire setup beats the 8800gts sli setup in all the 3dmark tests, and for gaming you really only need enough to smoothly play the game, above 60fps your eyes won't notice the difference anyways. It's only in crysis and the opengl test that it really kinda fails

[xMrBunglex]

I thought of going with 2x3870 but after latest review on hothardware i saw that ATI with it's crossfire can't go near new 8800GTS 512? So is that about drivers or just because they suck?

3870 CrossFire is faster than an 8800 Ultra in most games. if the review you read said otherwise, it is wrong.

[Ethelred]

Why do you think the R600 arch based 3870 has "far more potential"?

I base it on the fact that it has proven so far to be very overclock-able, cheaper, more efficient (energy-wise) and a pretty solid card. Perform marginally less than the 8800GT but as was said early it is by the slightest of margins...and I guess percentages don't matter.

Really a lot of this is just coming from people splitting hairs over a few frame rates (Now, I know there out there but don't show a stupid review of the 8800gt beating the 3870 by 20 frames or some such nonsense) and in my book they are pretty much neck-a-neck much like the difference between goose foie gras and duck foie gras...only the anally can tell and who cares about them, right?

Seriously Rollo, just take your little spiel elsewhere or something. Because at this point you have lost a lot of credibility as all you preach is to your NVidia overlord.

[BOBDOLER]

Thats great isn't it the 8800GT x2 squeaks out wins at 4xAA, it's just to bad that computerbases results suggest they are going to fall like a rock when you goto higher AA's. No their may not be any sli/cross benches of higher AA's to show it yet, nevertheless the circumstantial evidence does not look good for the 8800GT.

[LaMpiR]

3870 CrossFire is faster than an 8800 Ultra in most games. if the review you read said otherwise, it is wrong.

I needed this type of answer. Tnx :)

[Rollo]

I needed this type of answer. Tnx :)

Here's another type of answer you need when considering Crossfire:

http://www.rage3d.com/board/showthread.php?t=33908236&page=2

Your drivers use profiles even though those are not exposed. Unless the:added Crossfire profile for X game in your release notes means something magical and inherently different, correct? Profiles as in detecting the exe and applying a multi-GPU rendering technique to it, be it your compatible AFR or the performance AFR, as I don't think you still have SuperTiling available(I may be wrong on this).

The fact is that:a)many newer games don't scale at all because the abovementioned profile isn't in the drivers, and it takes a lot of time for it to be included;b)forcing AFR through CatAI doesn't always work;c)compatible AFR doesn't always work. When you get all of those together, it doesn't necessarily build confidence for a part(3870X2) that'll rely on multi-GPU rendering to give it its oomph, so I was asking if you had found a way to ensure that the vast majority of titles will scale, including new unreleased ones, without forcing hypothetical 3870X2 users to wait for a new Catalyst release. Thanks for your answer anyway.

Confidence in CF is not high on Rage3d because a lot of the people there have used it and been frustrated waiting for profiles on new games, never getting profiles or scaling on many games, and strange things happening on some games.

With an 8800Ultra, you get the same performance in every game, and although it may sometimes be a bit less on some games that scale well with multicard, that is definitely a consideration.

Buying a CF setup is not a plug it in and "Wow all my games run fast" now type of proposition- there are trade offs.

[Ethelred]

Here's another type of answer you need when considering Crossfire:

http://www.rage3d.com/board/showthread.php?t=33908236&page=2



Confidence in CF is not high on Rage3d because a lot of the people there have used it and been frustrated waiting for profiles on new games, never getting profiles or scaling on many games, and strange things happening on some games.

With an 8800Ultra, you get the same performance in every game, and although it may sometimes be a bit less on some games that scale well with multicard, that is definitely a consideration.

Buying a CF setup is not a plug it in and "Wow all my games run fast" now type of proposition- there are trade offs.

Way to quote an 'expert', Rollo. The gentlemen was asking a question, he wasn't speaking of; nor is he a representative of the community. Much like if I said "I have very little confidence in Intel's power to process information at a decent rate, very little confidence." And you if quoted that it doesn't really make it true. There are issues on all sides of the table with in regards to Crossfire and SLI, the system takes time. There are issues on both sides with SLI and Crossfire, both of them have issues with games. That being said a single card solution beats Crossfire and SLI in the case of High-end vs 2 - Mid-Range that is a fact, however, the cost vs performance is something has to be weighed by the user not you.

I would say reading the posts, that the confidence is high and the ones who have the 38x0 are satisfied. That holds true both here and elsewhere. The only one who is bringing it down are those who come into a forum with an objective such as yourself.

Try to be more Objective, Rollo.

[Rollo]

I base it on the fact that it has proven so far to be very overclock-able, cheaper, more efficient (energy-wise) and a pretty solid card. Perform marginally less than the 8800GT but as was said early it is by the slightest of margins...and I guess percentages don't matter.
http://www.firingsquad.com/hardware/amd_rv670_performance_preview/page16.asp
There are two things to note on that page:
1. OCing doesn't seem to bump 3870 performance much at all.
2. The 8800GT is using less energy under load. Here's another review showing it using less energy under load:
http://www.anandtech.com/video/showdoc.aspx?i=3151&p=11
I've also seen reviews where the 8800GT uses a few watts more, but the point is these cards are tied for energy consumption.



Really a lot of this is just coming from people splitting hairs over a few frame rates (Now, I know there out there but don't show a stupid review of the 8800gt beating the 3870 by 20 frames or some such nonsense)
It's not "nonsense" to the guy buying the cards Ethelred- it's the difference in his gaming experience-

http://www.anandtech.com/video/showdoc.aspx?i=3151&p=6
Ten games tested. At 16X12, 3 games are within 5%, 4 games are 14-18% slower, and 3 games are 22-27% slower.
That's not "splitting hairs"- it's a pretty big difference.


and in my book they are pretty much neck-a-neck much like the difference between goose foie gras and duck foie gras...only the anally can tell and who cares about them, right?

In a way you're right, but sometimes that difference can mean minimums that dip under 30fps and impact on gameplay.


Seriously Rollo, just take your little spiel elsewhere or something. Because at this point you have lost a lot of credibility as all you preach is to your NVidia overlord.

Sorry, but it seems to me people reading this board need some balance.

I'm not alone in my views- AMD is spiralling dangerously close to being bought out or just going out of business these days:

http://www.bit-tech.net/news/2007/12/10/amd_now_worth_less_than_it_paid_for_ati/1


AMD reached its lowest share price for more than four years last week and, as a result, its market value dropped to around US$5 billion – that's $400 million less than it paid for ATI in July 2006.

Intel, AMD’s major competitor in the CPU business, has been on a roll for the past 18 months and is now worth around US$162 billion, which makes the chip giant more than 32 times the size of AMD in monetary terms.

Even worse for AMD is that its partner-cum-archrival, Nvidia, has a market cap of around $19 billion, which makes it almost four times as valuable as the struggling platform company.

People just aren't buying into their products, although the 38XXs have been a good seller. They're really the only bright spot in the AMD line up these days, and you can't keep a company as big as AMD/ATi afloat on a couple of $200 cards.

[tiro_uspsss]

Sorry, but it seems to me people reading this board need some balance.

yeah the problem is when *u* start spouting BS, ur too illogical & irrational to let someone 'balance' it.... ur pretty much a one sided individual (theres a word for this, forgotten what it is tho...) ..... quit trolling, or i'll report u again

[demowhc]

While you may think the games that don't scale "don't matter", tell that to the guys with CF who have to disable one card to play some of their favorite games because CF degrades performance. It's a very big deal to a lot of current and ex CF owners, check it out on Rage3D, the ATi fansite if you don't believe me.


you make CF support sound worse than it really is... when i first got my CF set up i was expecting only a handful of games to work, but honestly all of my games scale just fine in CF..

at first i had a few teething problems but now 1 out of around 20 of my games doesnt work with CF...the other 19 all show good performance improvments.

the one game that CF didnt work in, 1 card had over 70 fps in that title anyway..

CF support is ok and its getting better, at the most you might have to wait a day or 2 for a 'hotfix' as in the case of crysis.

and no you dont always have to wait for a profile, cod4 is a new title for example and that scales great, about 70 - 80% performance increase out of the box.

i admit i have only been using CF since the 7.10 cats were released which apparently made big advances on the CF side of things, so i cant really comment on how it was before then.

[Ethelred]

http://www.anandtech.com/video/showdoc.aspx?i=3151&p=6
Ten games tested. At 16X12, 3 games are within 5%, 4 games are 14-18% slower, and 3 games are 22-27% slower.
That's not "splitting hairs"- it's a pretty big difference.


Everyone knows that the 3870 performs lower than the 8800gt...it's no secret, no one is trying to say that it doesn't. So really, what's your point or do you just want to :horse: As well as looking at the trend the 8800gt does worse at higher resolutions with more AA/AF...what does that say?

I really doubt that AMD is in dire straits. Yeah, they aren't doing the best recently however, they seems to be on a good trend with what is out and what is to come. I know many people who are looking forward to switching to AMD procs rather than Intel, even to me personally I may be redoing my set up in the next coming months and move over to an AMD. I think it's over hyping the issue to say they are near bankrupt or some such thing.

[C'DaleRider]

As I am late to the Rollo party, all of you do understand Rollo is paid by nVidia to be a viral marketer/booster/shill of nVidia products and has been banned on numerous boards across the internet, including Anandtech, amongst others.

[SNiiPE_DoGG]

As I am late to the Rollo party, all of you do understand Rollo is paid by nVidia to be a viral marketer/booster/shill of nVidia products and has been banned on numerous boards across the internet, including Anandtech, amongst others.

ive been preaching that story the whole time, fortunately he has now been banned, as noted in his rank thing under his name

[BOBDOLER]

Oh snap

computerbase has their 8800GTS 512 review up and even that seems to fall big when it comes to 8xAA

here (http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_nvidia_geforce_8800_gts_512/)

Poor rollo, bu bu the 8x doesn't matter! It's not going to look very good when the 8800GTS 512 in sli gets beat by the 3870 crossfire in 8x AA

[tiro_uspsss]

ive been preaching that story the whole time, fortunately he has now been banned, as noted in his rank thing under his name

is it permanent? please say yes! :D:D:D

[xMrBunglex]

i don't think it's that big of a deal. i run almost all my games at 4xAA anyway, even the games that don't push my hardware at all. there isn't much difference in IQ between 4x and 8x.