*OFFICIAL* Alienware M18x Benchmark Thread

chances are...they should fit...but due to the bios...may or may not work.

 

I'd love to try the Ati 7000 series especially next year when they release their top model in the 7000 series. We only get mid-budget class of the 7000 series this year.

 

figures...going to see if i can get some answers as well.

this was and probably still is the best buy, given what can be done to it, since the time it became for sale.

if you say it's so...i will concur. never had a m17x r2

mid will still cost an arm and a leg if dell doesn't allow us to upgrade from current models...with out having to have a problem first.

 

I have had a m17x R2 and this m18x. i would rather upgrade the 7 series in this laptop. Though the screen isn't as bright and vibrant; the extra real estate is better; especially comparing both side by side. The 330ps I THINK only works in the m18x which is going to be key for the 7 series (hopefully 330 is enough). The m18x has a better cooling design and it just looks much more aggressive and better when comparing the two in person.

The m18x R2 will surely have ivy bridge and probably a better screen; or 3D capability. However I can't see this machine not able to handle an extra GPU upgrade. if it doesn't then this was probably one of the worst purchases I have made.

 

What throttling?

A good run tonight... too bad my CPU doesn't perform as well as ruunning it as a standalone test... I guess the carry-over temps from the GFX tests gimps it. The CPU temps killed my 29k (since I hit 31k+ on CPU standalone tests)



I will say however that my GPU temps were completely under control (66/67) and never saw throttling at 740/1610 or even 750/1625, this last one froze, I suspect at this point the GPU's need more juice. If I try to go run anything over 724 with regular temps (73's or 74's) the throttling kicks in every time in both GFX tests. Temperature has to have something to do with it. I'm betting all my money on the fact that it isn't an Nvidia throttling (perhaps it has one but the one too, but less aggressive) but a Dell/Alienware one....

 

now that's looking allot more like right.

maybe i will grab some 580m's lol

 

I'm almost positive, that given the right pasting/cooling, these cards can manage a good 30k GPU scores without even hitting throttling...

I guess once I get the pasting job done, we'll have to give the new BIOS a try to prove this theory. But at this point, I think you are right John, this may very well be temperature initiated throttling, perhaps there's a couple of things that need to happen for the throttle to kick in.. load and temps... who knows, it's really confusing.

And what I am sure off, is that you would have been hovering 30k a few weeks back with the 580M and your CPU....

 

i think your right...because i would have set some dice on them and ran the .92V just to see what was going to happen.

 

Whoa! I got the Clevo 6990Ms working! Will have to join the benching campaign for the sake of science, lol. Need to know if there's any difference between the Clevo and Dell cards as far as OC'ing goes...

 

Here's the stock cpu+gpu Vantage (GPU - 24k)

 

How are the fans? Are they working properly? Congrats man, How'd you do it?

 

The fans are fine, I managed to flash the cards back with a Clevo vBIOS and it worked! I think the problem was with a system BIOS version (A00) after flashing to the latest it was ok.

 

Ahh nice. Glad it worked out. Clevo vbios is good in the R2 as well.

 

I had lower performance until I flashed upto A02 form A00

 

So your cards are also from Clevo?

 

No, dell

Think the bios had some added compatibility

 

I think so too cuz I didn't change anything else. Does anyone have a stock vantage run?

 

Not with a 2960 Aiki otherwise yes

 

Looks like the score is on par with a 580M SLI + 990X setup. Now, I'd like to run a series of stock gaming tests for a reference point.

 

I would very much like to see someone here run them ... or Aiki if you could run benches and compare to the posted benches on Notebookcheck or Tom's ... that would also be interesting....

 

Yes, that's the plan cuz their results look way lower than they should be

 

ok so here's the 1080p 4x AA very high run on Crysis - 45fps, everything stock. Very close to the 580M SLI + 990X here.

 

Very interesting ... 2 Questions

1) Are you using CAP profile 11.8 ?

2) could you (and I know this is doubling your work basically) also perform these with overdrive on in CCC ?

*I am generally pissed at AMD for not setting OD on automatically, which they could have done, instead you have to activate it. but it should increase performance 8-10%.

 

1) I didn't install any CAP's just ran the express install (11.8).

2) could you elaborate? Basically, I want to do the stock tests first with no OC', no tweaking of any kind.

 

Hrm, wondering about with CAP profiles, I was also wondering if Tom's et el had used the CAPs ... generally these improve performance, or at least thats what I understand.

2) Overdrive isn't OCing as such. In a way its a built in Over CLock if you want to define it like that. It just allows you to increase the clocks a slight amount.

Basically AMD is using it as a marketing tactic. OVER DRIVE ... a new awesome feature, raising the clocks even higher ... when In actuality they could have just set the clocks to that level. As far as I am concerned if its built into the drivers for the card, its not OCing. its a feature of the cards. All you ahve to do is go into CCC and turn Overdrive on.

But you might consider it a tweak ... I dunno
Semantics, perhaps. But it is what it is ... AMD is retarded

 

Sounds good, I'll see what OD does then. Meanwhile, I installed CAP4 and got same 45FPS. Now, I'm using the demo version of Crysis with no patches. Would it be different with the full version and latest patch?

 

difficult to say Aiki, I don't use Crysis ... would have to have someone else comment on this I am going to look through the release notes on the CAPs and see if there is a mention of Crysis ...

 

technically...it's over clocking...just on the laptop side they lock the sliders far lower than the desktop counterparts.... technically of course.

 

OK there was a CAP for Crysis, but I have no idea how long ago and I would hope/assume that CAPs are cummulative ? So if you have 11.8 CAP 4 it would have all the stuff that 11.5 CAP 3 would include ???

I don't really know though to be honest.

And again not sure if you would have to be running the full version to benefit ...

 

Ok, downloading the full version from Steam, might as well play it some day, lol.

 

LOL, thats going above and beyond Aiki

I don't think that was necessary.

Could of just moved on to the next bench

*although I did write a paper this summer I wrote a paper about virtual worlds, and included Civ 5 in it ... and so I downloaded it and played for like 2 weeks straight as 'reasearch'

 

as 'research' it is, hehe. Always wanted to grab the full version, needed the last straw to pull the trigger

BTW, looking at the available benches it is clear the 6990CF isn't polished, single GPU scores are very impressive. the 6990 > 580M in most cases, but as far as CF is concerned, even the 6970M's beat the 6990M's in many cases...

 

For the most part but not completely... the shader count makes a big difference. Otherwise my GPU scores would be up by the 6990m''s... they get around 8800 GPU I can just get over 8K.. in 3dmark11 that is and I can just manage 845-850 core on mine and i've seen 880-890 core 6990m. I wish lol.

Check some of Johnkss benches and Shaden and Scook etc... they are smoking the 6970m crossfires.

 

the problem with 3dmark11 is: it's a bloated gpu score in reality for amd. still better than the 580m when ran correctly though.(meaning: the way 580m users want it ran. )

 

John are you on 6990m if you are you probably have more core overclock room than 6970m in any bench...

 

Never used overdrive... used to use AMD overlock tool then when I got the 6970M I use Trixx and for more permanent overclock I flash my vbios..

 

I use trixx and have it auto start on boot up ... its so easy. Not that I require any OC for gaming ...

 

Was wondering if this was the case, because I have noticed some of that ... makes me wonder if AMD is working on better drivers or not then ... ?

580s just got a nice boost from driver support.

Guess we will just have to wait and see

 

At stock clock speeds there is not a humongous difference in performance between 6970 and 6990 CFX... minimal in gaming to speak of and a slightly wider gap in benchmarks... However, when bumping up clock speeds by just a little bit, the 6990 CFX totally kicks the 6970 CFX's butt in benchmarks. That's pretty clear by the differences in benchmark scores we see on this forum.

Here is a nice list of mobile GPU's sorted in descending order of performance: Comparison of Laptop Graphics Cards (This is based on Notebookcheck's data... YMMV).

 

Word record Overclocking oles!

AMD breaks world CPU frequency record Icrontic Tech

Cheers.

 

800/1100 guess there is room to go

 

Lots of room to go Drifter

 

My system exchange is in process, (authorized, but not "in production" yet). Unfortunately, my M18x started getting much worse about freezing in game and throwing a BSOD randomly. Temps have never been an issue. Around mid-week I started to experience a BSOD while the system was idle. All stop errors have pointed to my video cards. So, I decided to do some experimenting to see if I could do anything to make the system behave well enough that I could at least enjoy some fragging while waiting for my replacement.

I used RBE to modify the AMD Powerplay states structure so that all power states are permanently set at stock 6970 clock speeds of 680/900 and their maximum power of 1.1 volts. Now that I have done that and re-flashed the vBIOS so both cards run at fixed speeds regardless of GPU load I am having no issues. I have run multiple benchmarks with 3DMark11, 3DMark06, Vantage, Heaven and Furmark without my system crashing once. I am also no longer getting the frequent "Display driver amdkmdap stopped responding and has successfully recovered." error message with a crash to black screen (which was recorded 53 times in Event Viewer since 08/25/2011, including twice this morning before modding the vBIOS).

Clearly, this is not something I would consider a permanent fix, but it's something I plan to share with my Level 3 rep for their engineering team to research regarding the issues I have been experiencing with the 6970M cards. I think it's interesting that the serious instability problems my system has been experiencing seems to be cured with both cards running at a static clock speeds with static voltage. That may be related somehow to the fact that I was having some success with a driver-only (no CCC) installation with my previously installed 6970 cards. This could be an indication that these cards may have an issue working properly with AMD Powerplay in CFX. (I don't think there have been significant issues with a single 6970 GPU, but I admittedly have not paid close attention to the M17x R3 threads because I have no interest in the M17x R3.)

Today's results with the vBIOS tweak are posted below.

Fixed PowerPlay modes - at stock clock speeds

Spoiler

Fixed PowerPlay modes - with modest overclock and screen shot of RBE changes

Spoiler

Hopefully, no more of this crap...

Spoiler

On Fri 9/16/2011 5:00:41 AM GMT your computer crashed
crash dump file: C:\Windows\Minidump\091611-14227-01.dmp
This was probably caused by the following module: atikmpag.sys (atikmpag+0x7A48)
Bugcheck code: 0x116 (0xFFFFFA80127F24E0, 0xFFFFF8800302DA48, 0x0, 0x2)
Error: VIDEO_TDR_ERROR
file path: C:\Windows\system32\drivers\atikmpag.sys
product: AMD driver
company: Advanced Micro Devices, Inc.
description: AMD multi-vendor Miniport Driver
Bug check description: This indicates that an attempt to reset the display driver and recover from a timeout failed.
A third party driver was identified as the probable root cause of this system error. It is suggested you look for an update for the following driver: atikmpag.sys (AMD multi-vendor Miniport Driver, Advanced Micro Devices, Inc.).
Google query: atikmpag.sys Advanced Micro Devices, Inc. VIDEO_TDR_ERROR


[/QUOTE]

 

about time. haven't seen amd in the high cpu rankings in a long time.

 

If anyone else with 6970M CFX is having issues like freezing in game or during benchmarks, display driver frequently stops responding and recovers, or BSOD, you might want to see if my hack will work for you also. It's not a permanent fix, just like disabling OSD to prevent Aero from crashing is not, but you can at least get some serious gaming done without crashing. Since my current system has the third pair of 6970M video cards in it, I'm either just terribly unlucky or the 6970M video cards are garbage. The modded vBIOS is attached to my last post, above.

 

Ok, so I ran the test on the full version and got ~53-54FPS average, so it definitely makes difference! Now, there's a problem with the bechmark tool which doesn't show me the summary scores for some reason. Any insights?

 

very interesting Aiki, and that was with OD off I would guess ?

PS ... no idea how to solve that ...

 

Your benchmarkTool.ini should have the following as the set settings:

PHP:

path=c:\program files\test
DisableTips=False
ClearLog=True

 

Would be nice if someone with a 2960/2920 + 580MSLI could run the same tests with the latest drivers.

 

indeed we could have a real comparison to note if Tom's et el is off in the differentiation