HD 3870 Crossfire vs 9800 GTS SLI

OCZ Whitebook (Alienware M17) and Toshiba X305-Q706 are about the same price, but what about the performance? We haven't seen Toshiba's benchmarks yet, but as far as i see and although there will be some objections to this, 9800 GT SLI (Sager 9262) and 9800 GTX SLI seems to be far better than HD 3870 Crossfire. (Equal in 3DMARK06, but 47 FPS vs 57 FPS in Farcry 2, http://forum.notebookreview.com/showthread.php?t=318343&page=4 ) What do you think and suggest?

 

on the game vs. benchmark thing: the "bad" performance of the ATI cards in games is mostly due to immature drivers i believe. so you can't really take that as granted yet.

 

Actually i can, at least for a little Of course nVIDIA had many driver problems too, but i've been waiting for ATI to make better drivers since my X800XT! nVIDIA is good at making driver business.

 

name one game it's very good at then. and we can all go from there....

 

Its usually good at any game that was solely built on a platform utilizing nvidia graphics, thus bringing u to the idea of the nvidia logos being placemarked on certain games. similarly, some games are just designed for nvidia.

 

that is understandable some of the time...but the question still sits.

what is the best game the ati (3870m x2) cards excel at. you know. like how they excel at 3dmark06 right now. (3870m x2)

 

i am not sure. the easiest way is to first find out all of the games that are "designed" and advocated by ATI. then, we just need to find out for ourselves by testing it.

 

I personally can wait until all those people who ordered alienware and Ocz, receive their systems and post topics here on forum on: "Why my $3000 laptop cant play games smoothly!?". I just can not wait until someone unboxes their machine and then gets his 17fps in Crysis I would be pretty mad

 

umm...I don't think any notebook can play Crysis @ 1920x1200 at very high. Sure the Clevo's 9800M GTX SLi performs better than the 3870 x2, but both are below 30fps at those settings.

 

Well, it performs up there with a 9800M GT Sli in UT3 according to CNET (factoring in CPU speed). DMC4, Call of Juarez and Supreme Commander usually favor ATi cards, but I wouldn't know because no one here tested those games.

 

yes, most of us tried to explain this though. you included. this is why everyone wants certain test done in all high and not very high. hell, most desktops can't play crysis so how is it that a laptop would surpass them is beyond me. (talking about a dual video card anything and up)

im all for testing those out as well when i get some of them. not sure when that's going to happen though...*LOL*

im all for this as well.


ichime: if you see that guy that scored that 16.6k 3dmark....can you ask him to show some of his game benchmarks since he seems to have an understanding on how to make the ati's work for him. now if we can get him to benchmark games or 3dmark vantage, then we would have another high end perspective.

mark said he will be retesting when he gets time.

 

Just about any game that actually has a ingame setting for dual gpu's should already work. COD4, COD5, lost planet, UT3, Gears of war, Quake wars, thats all I can think of right now. Anything else will probably require you to configure crossfire.
Also for Justin's test, its clear that the games are only running using one graphics card, but they are all still running quit a bit slower than a single 9800m gt. That alone is pretty much all the proof you need that crossfire won't beat out sli.

 

single card

 

Single card of what?

 

sorry, forgot you haven't been following us with this gt/gtx/ati stuff

it's a 9800m gtx

 

It's a 9800m gtx, not sli.

 

yes, i had to test both.
single card = 35 fps
dual card = 56 fps

 

u get 56 ?? I get only 50fps but I do use 2xAA. What drivers u re using?

 

180.43, but it's also ran at 1920x1200

i was getting like 60+ at 1680x1050. not sure because im not at home right now.

 

So whats the deal with the owners of these laptops. It's been two weeks since people started receiving these and still not a single thread with benchmarks other than the three obvious people who only seem to know how to run 3dmark06. Now we have a thread thats asking for comparisons and this is the only thread those same 3 people still haven't posted squat....

 

Ive thrown up FC2, Crysis, and I think I also put up Fallout 3..what else would u like? I have CoD 4, we all know Bioshock is gonna run 60FPS maxed at native res..just like CoD 4 will. I apologize, but not sure what you are expecting of me. Ive got Assassins Creed, but never seen a request for it..It will also max out flawlessly. I mean, what would you like to accomplish?

Maybe ya should visit the "Real in game benchmarks . . " thead?

 

Cod4 would be a good start, COD5 is another easy one to test, UT3 has a demo I believe. Sorry if I come off harsh, but all I see is bickering in other threads and no help for those who need it.

 

I should be getting my system sometime this week and will be more than happy to provide any info for you guys that I can.

 

Wrote this in mhooper's thread after it got hijacked but it really belongs here I guess.

I would like to see a comparison done of the AW M17 that CNet reviewed (or a similarly equipped OCZ Whitebook) vs. Toshiba's just released X305-706.

Specs:

$1999 USD pricetag on both.
2.26GHz Intel Core 2 Duo P8400 on both.
M17-3GB,1066MHz DDR3 vs. X305-4GB, DDR3 1066MHz (CNet screws up and says AW has DDR2)
M17-160GB 7,200rpm vs. X305-320GB 7,200rpm
M17-Intel P45 vs. X305-nVidia9400M
M17-ATI Mobility Radeon HD3870 (x2) vs. X305 GeForce 9800M GTS SLi
M17-17" WUXGA, 1920x1200 vs. X305-17" WSXGA+, 1680x1050

Skimping on the AW screen leaves enough cash to pick up a 320GB HD and 4GB of RAM and stay at the same price tag. The X305 will always have the 9400M Chipset though which is a pretty nice bonus.

To test the performance difference those would be the best notebooks to use.

 

E-wrecked, it would be really nice if you made a compilation of your benchmarks, along with their background info. (Patches, drivers, settings etc.) And if we can also find mhooper's benchmarks, we could do some comparisons too.

 

Boy, I wonder why I haven't gone out and bought a bunch of games so I could benchmark them for people who I don't know who call us names and seem to feel entitled to benchmark results.

 

What wound you like to see. I should be getting mine sometime this week (hopefully) and will provide you with what I can. I've got Fallout 3, COD4 and COD5, The Witcher Enhanced Edition and several others that have already been benchmarked extensively. I can probably get some other stuff too

 

The first 3 you mentioned would be very nice, if you like once you get your computer I'll give you access to my steam account and you can try several other games.

 

Will do. Now, with these games, do they have built in benchmark tools? What exactly are you looking for, like FPS?

EDIT:

I also have STALKER and STALKER: CLEAR SKY

 

Mostly I'd like to see if crossfire is even working. So a fps comparison between crossfire enabled and disabled.

 

How do you disable crossfire?

 

Disable Catalyst AI if that doesn't work you can also disable one in the device manager.

 

Nice benchmark Johnksss 56 fps average is very good. Those 9800m GTX surely punches out some power. Well I need to get a better CPU for my M1730, an X9000 would be sweet to unleash more power of my 8800m GTX SLI

Hmm any Crysis benchmarks at Very High John of your 9800m GTX setup? What is the average at DX10 Very High? I landed on 21 fps average with my 8800m GTX SLI setup at 1920x1200.

 

24 to 25 frames very high. it's posted in one of these threads.

it's a screen shot of the screen running
http://forum.notebookreview.com/showthread.php?t=318343

 

Nice indeed John.

 

i'm getting really confused. i was agreeing with you johnksss that SLI was so much better than Xfire, but i saw your benchmark of Crysis , http://img443.imageshack.us/my.php?image=q6600noocveryhigh1920x1qi2.jpg , and it is not so much different than mhoopers?!

 

lets re look at the picture...
1: it's a 2.4 ghz cpu stock no over clock
2: it stocked gpu with no over clock
3: it's 19 fps and not 17fps
4: also stock factory driver.

are you with now?

 

Q6600 vs Q9550.

Edit: nevermind. I was beaten to the punch. geek.

 

I saw all those , but 2 FPS is not so much, especially considering the huge difference between the desktop versions of these cards: 9800 GTX SLI, even 8800 GT SLI is much faster than HD 3870 X2.

 

he got 25 from q9550. what's 2fps?

 

then i guess you have your answer... looks like the ati is the way to go for you...

 

And I got 22 fps average in Crysis very high DX10.

TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 99.58s, Average FPS: 20.08
Min FPS: 2.38 at frame 734, Max FPS: 27.84 at frame 891
Average Tri/Sec: -5883641, Tri/Frame: -292948
Recorded/Played Tris ratio: -3.13
!TimeDemo Run 1 Finished.
Play Time: 95.06s, Average FPS: 21.04
Min FPS: 2.38 at frame 734, Max FPS: 27.84 at frame 891
Average Tri/Sec: -5873667, Tri/Frame: -279186
Recorded/Played Tris ratio: -3.28
!TimeDemo Run 2 Finished.
Play Time: 91.56s, Average FPS: 21.84
Min FPS: 2.38 at frame 734, Max FPS: 27.84 at frame 891
Average Tri/Sec: -6059178, Tri/Frame: -277388
Recorded/Played Tris ratio: -3.30
!TimeDemo Run 3 Finished.
Play Time: 90.60s, Average FPS: 22.07
Min FPS: 2.38 at frame 734, Max FPS: 28.38 at frame 876
Average Tri/Sec: -6115382, Tri/Frame: -277034
Recorded/Played Tris ratio: -3.31
TimeDemo Play Ended, (4 Runs Performed)

 

i didn't see that 25 with q9550. Then it is much better as i expected. I didn't expect cpu change will cause such a big difference.

ATI is not the way for me definitely then, as i said on the first page, i can't wait another lifetime during they make better drivers, but if it were 19 vs 17, i would be confused. What i guessed for 9800 GTX SLI was to be over 30, even 35; so you can understand why i was a little shocked with just 2 FPS difference, but still , 25 vs 17 is big.

 

onlyone check this out a custom config I have made for Crysis that runs on my 8800m GTX SLI at an average of 40 fps at 1920x1200. Might wanna try that one out. Here´s a vid of it, sorry for the bad quality it´s from a cell phone camera.

http://www.youtube.com/watch?v=s62lybcpt54

 

yeah, but that's dx9.

this thing looks very nice at dx10 1920x1200 high(32-40 fps) or even 1680x1050 high (42-50fps).

 

that same cpu (q6600)can get 24 to 25 fps on very high and get 32 to 40 fps on high 1920x1200.

edit: he was over clocked at that time. i was not over clocked. you didn't notice that part? the 19 frames was ran out of box with standard driver. nothing fancy.

 

Yes that config is DX9. The average framerate above is DX10 though. With an X9000 I would prolly hit 24-25 fps average considering I have 22 fps average now with a 2.4 Core2Duo.

 

yeah, im knowing magnus...you never flub scores. i see ya gained 1 to 2 frames from the last run though. *LOL* when you *cough* clowned *cough* my *super duper 9800m gts* *cough*. *LOL*

 

he he new drivers the Big Bang II does it´s job

 

You guys are lucky that you have machines that can do that!

-J.B.