SLI or single gpu?

Hello all, still trying to determine whether to get m15x or m17 or m17x. I know this is probably a dumb question but does having 2 graphics cards really make that big of a difference as far as frame rates and how well games play and can the eyes see the difference? I want portability yes but I do want to be able to run all the hottest games (crysis and cod 4 for example) as well.

 

Yes it does make a difference but not instantly twice the performance. The performance of dual cards can vary game to game.

So i suggest you check the single and SLi benchmarks for the games you want to play.

 

The benefits of having SLI is becoming more and more apparent, that is without a doubt, most games are taking advantage of it. From my perspective, however it is still not worth it. the price tag of SLI is substantial and most single cards today (higher end cards) will be able to run most games at almost max, almost crysis as well...
I would rather buy a single card and upgrade sooner than you would with an SLI laptop and save a few bob, depending how long your laptop should last you. When you do need to upgrade, there will probably be a single card out there that will still be able to max games.
Also I like my laptops to be mobile...

Just my opinion though, if you got the cash and really want it, the benefits are there. Just those benefits arent worth the price tag for me to be honest

 

Well you can't always upgrade your GPU. M9750 users got lucky with the HD3750. Anyways, personally if I had a choice I wouldn't do SLi. For the most part, SLi does increase performance a lot and can make your gaming experience better. However, when something goes wrong, it is a huge to fix and some games will even run worse with SLi on vs. just one card.

As you can see I wrote a SLi guide and this was mostly because back in the days of the M9750, the 7950GTX's were horrible with SLi and had a lot of problems (I used to have the 7950GTXs, long story). I think though with these new systems SLi is being utilized a lot better so if you want to go for it I'd say go ahead.

 

It depends on how much it costs. The Ati Crossfire solution is not as expensive as the 8800/9800 SLi setup, so it would be an easier decision to make.

I prefer single card setups, but theres no doubt that SLi (and crossfire) have improved a lot in recent years (drivers and game support etc). And if you're gaming at very high resolution, it can definitely be a big boost in some scenarios.

Its hard to give a single answer, you would have to weigh up price, budget, alternatives, screen res options etc.

 

After reading many horror stories about SLI/XF gone wrong, I've got to say that my XF setup is my first dual-GPU solution ever, and it hasn't given me any crap yet. Plus, it's very inexpensive to go from one GPU to two as far as the M17 goes.

 

On the subject, if he does go for a single card what is the cost "later" to upgrade to SLI?

 

Couldn't tell you. I do know that there is a Mother and a Daughter GPU. They're not two of the same card in there.

 

Guys Help me decied
Im lookin to get the AW m17
is it worth buying ?
or should i go for a dual Sli
in the m17x, but they say
the new m17 is faster

are the Ati cards ok ?
or should wait & buy latter
something with dual 9800 Sli

They do come in Clevo laptops
but the ugly lookin

 

I'd go with the M17 and I wouldnt even think twice about it.

It's easily the best machine of the 3, and when better components start dropping (cough... sig, cough), you'll be glad you picked the M17.

 

How do you Double HD4870 X2 1GB?

4870 x2 + CF? are you dreaming or are you hoping?

 

Until CrossFire or SLi scaling >=90%, 2 graphics cards aren't worth it.

 

When I did the test on the dual 8600 M17X.

3D Marks 06

SLI 7008
Single 4045
Stealth Mode SLI 3784
Stealth Mode Single 3223

I know the grapic card low end but I will check it out with the 9800gt Next Week. Upgrade.

 

If you can afford it SLI is the way to go

 

maybe you want to check my Sig for SLI ive added Dead space recently

 

I pretty much think of it like this

SLi or Crossfire For High Resolution (1680 x XXXX)+
or
Single for Low Resolution

The gains of SLi and Crossfire at lower resolutions its usually less noticable, I like to think of 1920x1200 as the sweetspot

 

Woah woah wait. Did I miss something?

Anyway, yeah; SLI only really kicks in at high resolutions for most games that support it. I feel some difference when hitting 1280x*, but if anyone could confirm, feel free to.

 

Ok so from what everyone saying is that SLI is the way to go if you can afford it but a decent single gpu would be able to run most games just sacrificing some frame rates correct?

 

SLI doesnt kick on only high resolutions nowadays, all the benchmarks done in my review of SLI are on 1440x900, which is considered a low res for SLI, and i'm getting over double the Frames

 

I doubt you're getting over twice the frame rate. That would mean a scaling rate of over 100%, which would be impossible. In desktops, CrossFire/SLi adds up to 80% more performance.

 

i think benchmarks dont lie do they? deadspace is getting over 150% -.- did you even bother to look at the link?

 

I dont think your right brainer. Mybe 140%. I tend to get 130% Which other M17X get. If the computer was clean of all the bloatware and Lowest setting for Vista. That where the 140%.

Please give us a idea how it was 150%. How that was done.

Most M17X have dual 8700gt or 8800gtx With a 3dmarks of 12,000 Plus or 13,000 Plus score. And a single M15X get 10,000 in 3dmarks.

 

Ive tried Dead Space Max settings, 1440x900 Single GPU, VS max settings 1440x900 with SLI, and got that type of increase, with the same scene, did you even bother to look at the results i have in my sig before telling me i'm lying?

Besides, i dont think that m17X's get above 70% yet.

EDIT: Oh, almost forgot, and whats the point of the 3dmark06 scores you posted? what are you trying to prove?

 

if you got the money go for the best if not go for the rest.

 

Then give us other Users that have the same scores as your self. Mybe that will back up your benchmarks. Or have a Video of your fps. Mybe it will show facts.

I seen people use a desktop to make there laptop look faster. I seen people who edit there fps to make there Laptops look faster. Why dont you post a video. That show both the Laptop and game with the fps. mybe that will give us facts.

One thing I found about games. There areas that have alot of info caculations into it so the Fps are lower and other areas that the info has lower caculations so the gpu has less work to do so fps are faster.
This is one fact brainer did not use.

Here how I would do a Benchmark. I would Post Both the Highest fps and Lowest fps and then caculated the avg benchmark. Not print screen it.

Let take a game I like. Lord of the rings Shadow of Angmar. Some areas I get 40fps. Other areas I get 5 to 15fps Like Bree, and Old Forest due to high gpu demand. This is with a dual 8600gt. setting max out. Some areas are really slow.

Depending on when I play and areas of the game I play. Most of the time I get 25 to 30fps. Give or take a few fps. I would rather have a person give both high and low of fps and a video to give facts then one screen shot. How easy is it to stay in one spot and get a high fps. Then take a picture.

 

erm, just to be clear with you here, i'm a known guy in this place, and done several reviews, why would i lie about benchmarks? besides, if i wanted to be a prick, i could pick the scenes that i get 100+FPS and tell you omg SLI rocks, take Dead space for example, i chose some place with smoke and focused the light on a CERTAIN SPOT, which reduced my FPS by 20 FPS than walking in a corridor for example, ive played the whole chapter, even in fights, m maintained above 88 FPS, so that scene i have the screen of is the most one that killed my FPS. same goes for lets say Bioshock, i can get up to 90 FPS when fighting freaks, but i chose the fire scene because it was intense.... rest of the games are about the same, i could do a vid alright, but when is it going to upload when you have a 256kb connection? huh?

the one in my DMC4 review took me 6 hours to upload, and that was a 4 min video i think

EDIT: Oh! ive almost forgotten about this, did you even see the fallout 3 benchmark? this scene is the most intense so far, which is the wasteland incase you havent played the game, and still i maintained more than double..

 

You're not going to get over double the frames with an SLI setup, that would be over 100% increase, and thats just not going to happen. If you've done a bunch of tests that show that (as per your linked thread), unfortunately it just invalidates the testing procedures/results and the knowledge of the tester.

You can't just take one screenshot with fraps in the corner and use that as the basis for a 'test result' unfortunately. The effort is appreciated, but the testing methods don't stand up to scrutiny.

 

and how do you like to test? see that UT3 Bench? that was taken with the 169.25 driver, ive did one yesterday, got 35.55 and with SLI got over 75, thats not questionable is it?

if you dont like what i'm doing, suggest me how to bench instead of giving statements that SLI cant get over 100%, for example, the DMC4 benchmark was taken by a benchmarking tool, so i cant be possible Lying about those benchmarks, because what i did was sit and watch, then take a screen of the benchmark,

and for the games that dont have ingame benchmark tool, ive got the FPS as screens, and i cant possibly record with Fraps for the connection speed limitation AND Fraps will kill both SLI and no SLI performance

Sounds like ive made a mistake here, i thought the effort is going to be appreciated, but oh well, we live to learn dont we?

one last thing, and thats for you V C , you cant be telling me that the whole benchmark is bull because of the dead space screen, other screens are the best i could get , and the most intense that i know.

 

To answer the original question you have to go dual cards, there is just to much advantage, present and future you will see more support in that direction. IMO
A review done at guru3D has a nice run on some cards running CODwaw...

guru3d.com/article/call-of-duty-world-at-war-vga-graphics-performance-test/1

Here's an overview of the cards that we used to test the game with:

Radeon 4870 X2 2048MB
Radeon 4870 1024MB
Radeon 4870 512MB + CrossfireX
Radeon 4850 512MB
Radeon 4830 512MB
Radeon 3870 X2 1024MB
Radeon 3870 512MB
GeForce GTX 280 1024MB + SLI
GeForce GTX 260 896MB 216SP + SLI
GeForce 9800 GX2 1024MB
GeForce 9800 GTX+ 512MB
GeForce 8800 GTX 768MB
GeForce 8800 GTS 512MB
GeForce 9800 GT 512MB
GeForce 8800 GT 512MB
GeForce 9600 GT 512MB

 

Thanks for the link mate, just look at the increase there, and thats on an OLD ENGINE THAT BARLEY SUPPORTS SLI

and the increase in FPS is steady even on the lowest resolution tested (which is very very similar to 1440x900) i think my point got proved.

 

Unfortunately, you can't just take one screenshot with SLI and one with single-card and call that a 'result' or a benchmark. Its just not a valid test, no offence intended, but your thread is pointless if that's the method you've been using.

At the very least you need several 60-second fraps runs with max/min/avg.

 

and how could Fraps be better if its going to cut performance by 3/4....

if you think the screens arent good enough for you, i think the ones made by benchmarking tools are going to be just fine,

i wish there were benchmarking tools for every single game i benched, id be happy to do them, but thats not possible as you can see.

 

A Fraps fps benchmark doesnt adversely affect performance by any noticeable degree, certainly not 75% .

I don't think you fully understand what im talking about (or what fraps can do); im not suggesting recording a video with fraps. Im talking about a simple 60-seconds max/min/avg FPS test (F11 by default in fraps). Its much more meaningful than just taking a single screenshot and using whatever FPS happened to be at that moment in time as your 'result'.

 

hmm.. let me check, i havent known that Fraps had such thing, will report back, thanks

 

EDIT: sorry i'll add the results in my thread, dont wanna Hijack this one