*Official Alienware M18x AMD CrossFire Discussion Thread (re: Dell/AW Conference Call)*

Yeah, mine is the same currently. Just idling with a few tabs open atm and the temps are like this......with the primary at 52c & secondary only at 42c, there is no need for the fan to come on for the secondary......actually, just turning the cooler fans off (they are pretty silent at full tilt anyway so i normally just leave 'em on) there is absolutley no sign of any fan noise at all....from anywhere.....but then it will rise up on the primary very slowly and kick in.... at around 62-63c

I can sit at idle like this all day long with my cooler on and temps are rock steady at low 50's primary / low 40's secondary....and this is with external display on & internal display off. If its vice-versa, the temps do shoot up a bit but I guess that's down to using the built in display and direct heat emmissions (maybe) - weird as I thought they would run hotter feeding an external 24" monitor than the lappy screen...

 

Just done a quick test. Unhooked the U2410 and cooler and let the system idle with 3 tabs open and a bit of surfing. As you can see, primary got quite high before the fans kicked in....secondary still did not require cooling....

I reckon the fans could do with coming on a bit earlier than mid to high 60's but once they DO come on, they cool rapidly.

Also, with the external disconnected and no cooler, the temps will "cycle" where the primary cools down to low 40's - fan goes off - slowly rise back up and come on again....I have noticed a difference in this "cycling" since flashing A04....this being that the secondary will no longer get anywhere near as warm as on A03....for me, the temps on the secondary no longer idle up to mid 60's before initiating the secondary fan quite so often - they stay below 60c most of the time and the fan for the secondary really doesnt kick in...it will eventually hit a wall and the fan comes on but it just doesnt seem that frequent.....guess its not getting warm enough!

Before, on A03, the gpu's would effectively "take turns" in cycling up to mid 60's and initiating the fan before cooling back down again quite regularly and noticeably.

I no longer see this cycling quite as much for some odd reason.

Anyway, just my 0.02 on the temps of these cards with a view to maybe helping out.

EDIT: Still suffering from random in-game shutdowns though....this is obviously a problem under load.

 

when gaming...did you get the big 6C to 10C between gpus in crossfire?

also...while in idle...are your fans off...with cpu fan being the only one on?

right now...both gpu fans are off and cpu fan is on. temps are 46C on both cards.

 

The fan simply does not rev up. (GPU overheats and system shuts down). That is the central issue and it has nothing to do with extreme setups vs non. I think those of us with problems all have recent builds.

 

you just dont seem to have any luck at all.. every single laptop you have bought has had issues...

 

I did not have this issue with the 2630 ... and I could run room temps without fan control quite high ... like 850/1150.

Had the A02 bios

 

I agree. I think the heat thresholds on the fans need tweaking. But do i wait or return my system. No way I'm going to let them give me refurbished m18x or 6990's

 

same here shaden...i could run pretty high as well.

side note:
did a reboot and now my gpus (580M) are sitting at 60C and no fans are on.

 

Odd as I just noticed that both my gpu's were idling at 60c ish (no cooler, just M18x on the cooler and no external)....didnt reboot but plugged the cooler in and the temps dropped to 53 and 42 respectively....its almost as if they were "wanting" the fans to come on but just not getting to that level to "kick in" - dont think its good to have cards idling at 60c+ without fans coming on whatever way you look at it,, right?

Now I know that a cooler is beneficial but 18c better on the secondary? - no way.....something does seem slightly off with consistent temps with 6990's

 

mine changed when i closed gpuz and reopened it. the primary gpu fan kicked in and temps dropped to 46C....but now it's right back to 60C. no cooler or anything. just sitting there.

you played battlefield 3 right? what kind of temps in game?

 

Ive noticed that behaviour too with gpu-z...opening and closing gives different readings without just cause....mine are currently rock solid as per my last post on the previous page (cooler back on full) but that doesnt account for 18c of cooling!

 

Ordered mine on the 19th of October. It has the fan problem and the keyboard is not the UK version (i am in the UK). Sometimes after rebooting the taskbar is not transparent.

The spec is:

cpu = 2760
gpu 2x6990
4gb ram
320 gb hd
killer wifi
blue ray drive
win 7 hp

bios A04 standard settings.

Also after reboot, befor the fan comes on it makes a sort of noise (clicking) on the first gpu. It may also make the noise on the second gpu, hard to tell.

 

Yeah seems that way. Here I was all excited to finally have a system that was trouble free.

 

yeah, im actually very happy with this setup now. once i flashed the 620/620 vbios...all is well.

 

Interesting that so many are having issues with their Dell 6990M's. I have Clevo cards and all is great except the high fan thresholds. Regardless of what game I play, it will always max at ~80C.

 

throw back in your 2630 and see if it's still the same...im a bit curious....

 

I had it in for a few days and all was fine. Oh, there is one weird issue. Sometimes, the left/primary GPU fan would hit 100% for no apparent reason (temps below 70C) and stay there till the temp drops to 45C. Happens when I start stressing the CPU.

 

that's with clevo cards and dell vbios right?

 

that's with clevo cards and clevo vBIOS. flashing the cards with Dell vBIOS bricks them. You then need to pair each card with a Dell 6970/6990, boot to DOS and flash them back

 

Ugh that is horrible.

But again John I didn't have any problems with my 6990s on my 2630 ... and you didn't have any problems with you 6990s either ... so I don't know if that is the link.

I could throw back in the CPU but I thrashed that thing all over without any heat issues ...

maybe bios issues or just as some people are suggesting bad batch of cards ? Is it possible one of the heat sensors isn't working properly on them or some such thing ???

 

Do any M17x R3's that have a 6990 gpu in have this problem?

 

I got my m18x about 3 weeks ago, I havent had any overheating issues.... by the way i repasted with the Dimond paste.
it sounds like the paste has nothing to do with this, it sounds like a fan problem (not turning on)

What program do you guys use to check temps? I use GPUz.
Wondering if there is a program that would display temps on display, Kinda of like frams but instead of frams to have the temperatures.

 

I got my M18X in September 2011, no fan, thermal or shutdown issues (except the Black Ops MP BSOD with any driver newer than 11.9).

I'm in Ireland so mine was made in China, shipped from Shanghai.

I'm running W7-64, all dell drivers except newer IRST driver and Catayst 11.10v3.

Specs below.

 

Awesome. Glad you guys are getting ready. I think the thermal shutdown thing is going to have some news real soon. I know we are aware and working on a fix. I'd say if you can reproduce the issue on a call with tech support, you should have your cards replaced.

Next steps for you guys are to settle on a (loose) agenda, decide on your "speaking panel" and let me know via Mr Fox the list of invitees. I will work with Mr Fox on time/date and send out invites via email once we are ready.

Just to be clear, we (Dell) do not care how many people join the call. We'd just like the number who are going to participate by speaking be kept to an amount of names and voices we can remember and recognize. If you just want to listen in and not be part of the conversation, that's fine by us. Just mute your mic so we don't hear you breathe on the call.

 

Bill, by fix do you mean a bios based fix or a hardware revision? If its a bios fix I'll wait on that rather than replace the cards outright.

 

Last I heard it was a hardware replacement, but they may find something in the VBIOS that will address. I'd hold off until I get something more definitive I can share. I also hear the 6990's are on back order right now anyway. Not much you can do but cold boot and reboot before you use it if you are having the issue.

 

Thanks for the updates Bill. I am going to call back tonight and hold off on getting the cards replaced right now. Again I hate to have a tech come out and tear into my brand new system and put in refurbished cards. I am personally going to wait until we have a definitive answer as to what the problem is and how to fix it before I go any route.

 

Just so I'm perfectly clear, you can work around the problem by rebooting after a cold boot, right?

 

Hi Bill,

Thanks for the updates thus far!

Just to double-check...are you sure the 6990's are on back order at the moment?

Reason being, I have just ordered a new m18x last Monday with 6990's in crossfire...and I'm due delivery this Wednesday (note: I'm based in the UK).

Hopefully my system will be fine...but this thread/issue is making me slightly apprehensive!

Cheers

 

I'm not sure about inventories in all regions and inventories for manufacturing. We usually keep inventory separate by Production and Replacement (I made up the names and capitalized). Each region manages its own Replacement stock. I don't think there is a back log on Production stock. I think it's just on Replacement stock, and that's just what I heard for the Americas. I am not sure how we're looking in other regions. As far as I know, new orders are not affected.

 

So are new orders expected to have these issues? Reason im asking is because my m18x is in production phase.

 

You are correct that IN MOST CASES, a restart after a cold boot resolves the problem. I have found in my case however, that at least once it did not fix it, and in one other "trial" a restart only fixed the fan on gpu 1 (left side) however the fan on gpu 2 still did not work. Another restart allowed both to kick in.

The other work around is to use a 3rd party software such as HWiNFO64 to force the fans to work after a cold boot.

I believe you are also correct in the assumption that Production and Replacement stocks are ordered and logged seperatly. The reason I believe you are right is that my system was ordered and built after there had been people on the waiting list for replacement cards however my system was built and shipped with new cards in 5 days from order date hence the assumption that the backordered cards do not affect new production systems (or that new systems take priority and there were a limited number of cards avaliable that were "reserved" for new builds.

 

I think we can all agree that user-controlled fan profiles would be amazing. I don't appreciate my GPU's going up to 82ºc before the fans really kick in. (assuming they even turn on at all)

I also don't appreciate the problems I've dealt with since 2009, the countless hours on the phone with tech support, multiple replacement parts and systems - and yet I STILL have problems. I purchased an M17xR3 for my wife and it works flawlessly...

Now that my warranty is expired, and if these replacement cards don't fix the problem, you guys might get a surprise youtube video of me bouncing an M18x off the pavement.

Frankly I'm done being civil. I'm tired of being nice. The tech support people that you hire out of freaking costa rica don't know anything about computers, and your quality control is non-existant. The in-home technicians are borderline autistic. One dropped my M18x, another spent 3 hours figuring out how to take it apart - only to break half the connectors on my motherboard. I have yet to receive a system that didn't require immediate attention. Regardless of how the situation turns out, Dell has forever lost a (once) loyal customer. I'll take my chances elsewhere.

 

i can answer that for you....the answer is plainly put... NO! not gonna happen any time soon or the near near future.

 

Kinda like this huh? (they obviously knew nothing about computers ....AT ALL....) but I can see how problems can escalate!!!!

Alienware SUCKS!!! Laptop computer destruction! Alienware tech support made me KILL MY PC!!! - YouTube

 

Rofl.... that made my day. I almost feel better now. I might do something creative. Maybe I'll strap it to the front of an F350 at work and put it through a wall.

 

Glad it gave you a laugh mate - in all seriousness, I know how it is....kinda felt like tossing mine off the highest building I could find at times.....

 

It's quite frustrating. I suppose I'll try to give some useful info for the Dell techs though. Here's all the troubleshooting I've done so far with my new system that overheats:

1 reinstall of Windows 7 Home Premium (alienware disc)
2 reinstalls of Windows 7 Ultimate (personal disc)
Stock dell AMD driver, 11.8 WHQL, 11.9 WHQL, 11.10 WHQL
Reflashed the stock vbios from my cards. Flashed the stock vbios from my wife's M17xR3 6990m, and flashed a bios uploaded by another NBR member.
A02-A04 bios flashes - to include the 'unlocked' A03. I played with a million different settings in that modded Vbios with no luck.

At that point I decided it HAD to be a hardware issue (mobo or GPU), but I decided to play with some windows settings. Using Regedit to always enable crossfire and disable ULPS did nothing, as did disabling powerplay. I double checked to make sure my power settings were on High Performance (and they are). Disabling crossfire didn't help. Talking to tech support made me angry. Replacement cards were ordered 2 weeks ago... still waiting. My patience = gone. I feel sorry for the tech that shows up at my door.

Full repaste of the system with Arctic Cooling MX-4 lowered overall temps as expected, but the fans will still sit at idle until the computer overheats and shuts down (GPU TEMP ONLY). My cpu temps are fantastic... hey something positive ._.

HWiNFO64 is the only thing that keeps my system running cool.

 

I know you're frustrated Nicholaus and so am I but I think this is most likely AMD's fault. Dell should have halted production as soon as they found out or done more extensive tests on their builds to filter out the cards that are defective. Hopefully they can remedy the problem soon.

 

I agree. After taking a closer look at my Vbios, I noticed the vendor was listed as the first option (Apollo, Jetway, etc...). I tried changing the vendor to Dell and reflashing but my theory is that these cards were never designed for the M18

 

Correct. Manufactures make the components. Then send them to Dell and other sellers to build around the components. That's where you get ES/QS versions of components. The components are not made for any specific system.

 

Not AMD's fault since AMD only makes the GPU die. The same dies as those that go into desktop 6850/6870 line up. They speed bin them and the weaker clocking ones get passed onto add-in board partners for mobile MXM consumption. And that doesn't mean they are faulty chips by any means, they just don't clock very high, is all.

The fault is with the Add-in board partner who makes the PCB and they also do the base VBIOS (later configured by ODMs) for the board.

I can safely say its not the GPU die thats giving issues, its the quality of components used to make the PCB that houses the dies.

 

You're right but in this case I think the board maker may be AMD themselves. I don't think Compal manufactures the boards for Dell. If it is a 3rd party, I'm not sure who. Ultimately the blame still falls on Dell for letting this slip through QC.

 

AMD has pulled out of the reference board supply market long time ago. They only supply reference designs and its upto the Add-in partners to follow that or cut corners to save money (inferior quality as a result)

 

I can definitely recall my previous set of 6990's and 6970's being from a different vendor. I almost want to say the vendor was listed as dell. Can some people that aren't having the fan issue upload a screenshot of their vbios in RBE so we can see the vendor? I'd like to see how many cards from different vendors are floating around out there.

 

tell me how nicholaus and I will be happy to ...

** as in I know how to screen shot, just not how to pull up the vbios info you are requesting

 

Download GPU-Z and click the little icon that allows you to copy your vbios.
Drag the Blackcomb.bin file that you got from GPU-Z over to Radeon Bios Editor. The vendor ID should be on the bottom-right of the first tab.

 

I think this is what you are lookin for .....

http://i1227.photobucket.com/albums/ee425/Abetterplayer/6990vbios2.jpg

 

Hope that helps ... I have had no errors with my 6990s and have pushed them to their physical max ... only John I think has pushed his farther by making them collllld ... which I will be doing as soon as I get a free bloody day to take this outside now that its hovering around 0 !

 

waiting for that day!...