[Guide] Installing AMD Radeon Mobility 6970 / 6990 in your M17x R2 (Single GPU and CrossfireX)

Yep same here. I've got all my settings maxed out in the game (but don't typically touch the CCC settings). And everything runs as smooth as expected (minus the server lag, but that's another story).

 

I played about 3 hours of Diablo 3 and these are my temps. Here is a screen shot.

Bear in mind the 6990's are slightly overclocked and are running at stock voltages.

I wonder why D3 does not support crossfire...not as though it needs it. Perhaps in the future when a few high res texture packs come out.

 

Thanks for the responses guys.

One more Q...

Do you think the lack of warranty is an issue though. I'm getting it at less than the $1800 you would value it at, but even so?

 

Hi 2Ravens,

Personally I don't think so. I've heard that you can add a warranty after the fact, but the 6990's and the 940XM (assuming the 940 is an ES version added after the fact and not ordered with the system) probably wouldn't be covered anyway. So I wouldn't worry to much about it. The people here are very good at helping folks out, so if you do run into problems hopefully the folks on the board can help sort it out.

Just my 2 cents. Good luck on your decision either way!

 

If you're buying something with 6970's in it, then the warranty has been long void.

 

Ok, probably a dumb question here, but, got my 6990's re-pasted and did the cooling mods (haven't seen the temps hit 75C yet). However, I am getting BSOD with atikmpag.sys on a fairly regular basis and getting amd driver recovery messages from windows. Anyone seen this with 6990's? I have looked for updated vbios and i can't even find anywhere that has the version these cards are running (the versions i see are all lower by a large margin). Currently running 013.011.000.017.040794 (BR409794.004) vbios (Cards are Clevo/Kapok) 1558

 

Hi PWn3R,

Yeah there are sometimes problems with the 6990's and it can depend a lot on the drivers that you are using.

Firstly though, because the last thing changed were the paste\pads, make sure that the crashing is not temperature related. If temps are not deemed to be the problem then it is probably drivers. Make sure you are using something like hwinfo32 so you can monitor all the temps of the video card and not just the GPU temp

Are your 6990's undervolted? You want to make sure they are running at stock volts before beginning troubleshooting as this can really mess with your results.

I am currently using 12.4 previews, and there was some success with the 12.2 previews (not to be confused with the official 12.2 release).

Not sure what everyone is currently using now though, as I have not updated in a while.

I would start with a complete uninstall and driver sweep of your existing drivers and then try to reinstall the ones that you had success with before. Failing, try a different revision and see how you are going. Just to mention off hand (it may not be causing any issues) but perhaps turn off your overdrive and this has been known to also cause issues.

I would not mess with updating the vBios as I doubt this is the cause for your issues.

 

Ok, I just installed the 12.4 release drivers from the AMD site after sweeping and am STILL getting atikmpag.sys BSODs and even having my entire PC hang just sitting in windows after boot up, sometimes I can't even click on anything after a boot before it hangs. I am sitting at 46C on one card and 45C on the other right now in windows with nothing running -- I haven't been able to get a game to actually get to the login screen in the last hour -- i had diablo up for a bit earlier then it took a dump and haven't been able to get to the login screen successfully since.

I used driver sweeper, rebooted to safemode reran sweeper and then rebooted and installed. I did try toggling crossfire a few minutes ago and it turned off successfully but turning it on again resulted in a hang.

I'm not sure what the problem is here as I am not overclocked at all. -- Douse, what is your voltage at on your cards, and where should I go to check mine? (Overdrive is off)

 

Are you making sure to check all the temps on the cards? GPU, Shader, Mem etc?

Stock voltage is 1.1 volts

 

Douse, here are my temps from the application -- the PC has been up for almost 20 minutes without hanging this time but i haven't run anything that would actually make the cards work -- i am downloading the catalyst 12.5 beta drivers and i will give those a whirl -- if all else fails, i'll try to find a copy of the 12.2 preview drivers.

 

Ok, just checking

Hrm.........

If changing the drivers does not do the trick try disabling the crossfire and loading a game.

I can't remember if they were 12.2 previews or 12.1 previews......... hold up I will find out

EDIT:: I remember now, I was using 12.1 previews without an issue and these were also used by others. I also used 12.2 previews without any major issues.

 

Ok, i'm wondering if there is something wrong with the hardware itself -- I put thermal paste on the memory as suggested elsewhere and ducted the fans, -- everything was happy until i crashed out earlier this morning or so it seemed.

Edit: So, I booted in just now and opened CCC and looked, at the information in the OC section -- second GPU was showing 100% usage and no temperature information -- when it dropped to 0%, the system froze on the windows desktop. I will install 12.5 after sweeping a second time and then if that fails try it with no crossfire.

 

Did you put A LOT of paste on the memory?!?

Because if you do use paste, there has to be a fair amount as the memory does not make complete contact to the heat sink thanks to the components at the back.

This is the danger of pasting the memory as people often do not use enough and then they fry.

EDIT:: Try turning off ULPS

 

I put about a rice grain worth on each memory module on both cards.

 

From my understanding the paste must be pretty thick so that it's making contact on the memory and the heat sink (so a rice grain is probably no where near enough). I know it makes quite a mess which is why I've never attempted it. Maybe someone here has done this method, for me the pads have always been the easiest, cleanest so never bothered with the pasting method

 

I've always used pads on the mem. Never tried the paste since the gap is quite significant for paste.

 

Crap... well I would put at least the size of a pea. I would be re pasting ASAP and hope that no damage has been done. This is only for the memory chips along the back of the card

I found I had better results with paste as it makes sure that the heat sink is sitting as low as possible which gave me a better contact on the die. But yes it is risky and you need thick paste and a fair amount of it.

 

It was making contact, buying pads now and will redo it.

 

One problem with putting a thick layer of thermal paste on the memory is that, thermal paste doesn't have the same thermal properties as copper for example.

A very thick layer of thermal paste will actually insulate instead of providing a better contact between the heatsink and whatever component. Generally speaking it's never a good idea to do that.

Memory is robust, so I doubt it's damaged, but put the pads back on asap. But you also have to keep in mind, the memory has no way of shutting itself down when it gets too hot like a gpu can do - it will just fry.

EDIT By thick layer, I mean around half a millimeter or more.

 

True but it would just spread out and become as thin as required. Plus my heat sink is modified to make better contact.

I would imagine that the paste would be thinner than a mm.

I have never had issues and I have done it to both the 6990's and the 5870's. I found that the pads lifted the heatsink even more which reduced contact with the die

 

The thing is though I don't think PWn3R has modified heatsinks. If you can get the heatsink practically touching the memory then using paste is fine, but you can't do that with the stock heat sink.

 

True but even with a stock heatsink I think you can get a better fit and result using paste on the memory. Just my opinion.

Although I do recommend pads as it is safer.

 

Well, i went to 3 places today and no one has pads. Not even the Fry's electronics here -- i bought paste (Antec Formula 7 Advanced Nano Diamond) and am repasting the GPU/Memory right now -- I tried doing just one -- but the problem is still rearing its ugly head. I'm about to do the second GPU. I don't get it though becuase this thing ran fine the other day and evening for over 3 hours total with no problems and i was checking the heat levels the whole time. It just started acting up this AM.

Just got done putting this thing back together again -- still having problems -- i wonder if something died on the secondary card -- i'm about to try running a game on a single card.

The problem appears to be the second card -- running WoW @ 212FPS on 1 card right now on high settings -- the AMD overdrive software also shows that the second card runs @ 100% load on boot. The second card is also falsly reporting temp @ 45C and was just checked prior to being pulled with a meat thermometer and showed at approximately 65C.....

Now i have to pray i can get it exchanged from the ebay store i bought it from....

 

Hi PWn3R,

I believe you went through upgradeyourlaptop on Ebay? If so they have traditionally been pretty good about exchanging cards when they have been broken, but good luck to you.

If you are looking for thermal pads, I'd recommend going to Frozencpu.com they have a great selection (Fujipoly is the brand I use) and have been pretty quick with shipping.

Keep us posted.

 

i have mx-4 since a week on my gpu and mem and the result is very good.
diablo 3 max heat on hwinfo 75degree on gpu and mem

 

That's correct, thanks for the information on that. I'm assuming that no one is going to be in from their team on the weekend, not that it matters much since I'm getting married today and will be away from home for several weeks.

Hopefully I can get it exchanged. I flipped the card over and started looking VERY closely with a magnifying glass yesterday. I found several very small resistors near the heat spreader brace on the bottom of the card that appear to have been damaged when the heat spreader was put into place before i received the card.

I tried completely stripping the card down, cleaning everything and re-assembling earlier today. I still was unable to get the machine to boot and be stable just at the desktop with Crossfire turned on. I've removed the second card again for now and will wait for a response from the upgradeyourlaptop guys next week.

Thanks again to everyone who posted and tried to help out yesterday!

 

Sorry to hear it PWn3R. Always sucks when this sort of the thing happens. At least you purchased through a reputable company.

Hopefully it gets sorted quickly.

 

Yeah that's sad news when an expensive GPU goes to waste.
Hope you get good online support from upgradeyourlaptop.

I used paste with my 5870s fine and got perfect cooling with the temps never exceeding 75C overclocked.

I attempted the same trick with the 6990Ms and it just didn't work.
The compression you can get with the pads allowing for the use of the stock heatsink cannot be reproduced with the paste I think. It's a shame because I really wanted to use that method (I had my 5870s running like that for 1yr and when I changed my GPUs over to 6990s, the paste had not moved).
Worth noting the paste does leave a gray imprint on the shader and memory blocks, no matter how much you try to clean them up and it is a messy operation.

You could possibly use paste on some of the blocks that sit flush against the heatshink and mix it up with pads but I would not use paste on the shaders.

My opinion on this matter. Waiting for 680Ms to rear their ugly heads on online benchmarks but the one 3DMark11 score released so far is disappointing to say the least.

Cheers.

 

I don't mean to nitpick, but the shaders are part of the gpu die. They're not a separate chip. The only separate chips are the memory chips - which like you said, are the black chips surrounding the gpu die.

And yeah, once you put paste on them, they will forever be grey no matter how much cleaning you do.

 

Is anyone planning on installing 7970M Xfire on this machine?

 

Eventually. Pat, how you liking that X7200? I'd love one of those rigs.

 

Congrats on your wedding!! And hope you can get the card sorted out.

 

The 3 HDD bays are marvelous, waiting for the 1TB 7200 RPM drives since they are getting full . I'm actually thinking of selling it and going with Alienware. Those AAFES discounts are tempting.. very tempting... I'm waiting for the 2nd gen 28nm cards before I upgrade though.

 

Is that a desktop i7 in the X7200???

 

Yup, this thing is End Of Life though. The P270WM, the successor, was supposed to come out a month ago but Clevo is saying that Intel messed up on the motherboard design so it's delayed for a couple of months. They are redesigning the motherboard according to the resellers.

 

Someone has tested the new Ati 12.4b driver ???

AMD Catalyst 12.4b (8.961.3 May 24) AMD Official Hotfix - Guru3D.com Forums

Profiles 12.4 CAP 2 : AMD Catalyst Application Profiles 12.4 CAP2 (May 24) - Guru3D.com Forums

 

I will test when I get home although I doubt much will have changed for our cards compared to 12.4

Wow, what wattage is the PSU for these laptops? I saw that the CPU alone pulls 130 watts!

 

Here's the description for that driver btw:

300W, the brick is almost double the height of the Alienware's.

 

first post here, so sorry if this has already been answered/asked.

urgntly need help with my m17x! its an r2, originally had dual 5870's...but luckily got a 6990m from a buddy for free . problem is, its a dell card, from m17x r3..it works, i can game and stuff but it freezes/blue screens if i unplug/plug in power adapter, it doesn't go to sleep, shut down/hibernate....

tried different drivers but no luck same problem, it also over heats, hits 88*c in furmark then just shuts off in less than 5 minutes. i was wondering if i could flash the bios to a revision that would work?? if so any help on where to find that and how to do it. thanks

 

Uhh...6990M are beast cards, they are NOT meant to run on battery at all. Also did you try updating your drivers? I'm on 12.3 beta drivers, no issues so far. Also you may have to undervolt your 6990M's to run cooler, limitations of the 240 watt PSU.

 

running 12.4, also its a single 6990m.

 

If it's a Dell card, you're going to have to flash it to the second revision of the Clevo 6990m vbios. (This is different from flashing the system bios)

All the details you'll need have already been gone over in the 6990m/6970m Crossfire in M17x R2 thread.

You should keep in mind though, your fans will be stuck at 100%, but at least everything else will be fully stable.

 

thanks alot, was looking at that but i guess i didnt read through properly. will do the flash and hopefully il be good to go. and thats a beast of a m17x you have. 480w psu on it..?

 

No problem, there's a lot to read in there (500 pages) so it's easy to miss. Once you flash it, you should wipe your drivers with driver sweeper and re-install them.

I had some problems with stability on 12.4 so I reverted back to 12.3, but most people are fine with 12.4.

I needed the extra power from the PSU so I could keep the OC on my cards and cpu so high without tripping the PSU breakers. It's simply two 240 watt PSU's isolated from each other with a pair of diodes and then wired in parallel.

 

Tried the 12.4b's with the mobility mod.

Installed perfectly fine, also installed the latest CAPs which is now giving me crossfire support in Diablo 3.

Haven't been able to play for an extended period of time, but so far so good even overclocked.

You will need to do the dll mod to get overclocking software to work, such as Trixx: [GUIDE] Enable unofficial overclocking after Catalyst 12.2 - Guru3D.com Forums

EDIT:: Played for about 3 hours with no issues at all, cards run even cooler now that crossfire is working with temps peaking at mid 70's even with overclock. Obviously they are only being utilised at around 40%

 

so i flashed (well tried flashing it) found a clevo 6990m bios from this site and it didnt work, error i got was new ssid was different from old ssid. any ideas? and is there somewhere i can get another clevo bios from? been trying all day but nothing :/ . New problem now, i cant boot up when power is connected, it gets to the laoding screen then i just see my mouse pointer and a black screen and thats it...only way to get it to work is without the power plugged in, or with the power pluggedin in safe mode.

 

You need to force flash.

atiflash -p -f 0 nbios.rom

-f will force flash.

Sent from my Samsung Captivate/ICS using Tapatalk 2

 

that worked, thanks! solved almost everything, however, i still have one problem, overheating(i just received my thermal pads, so will replace the old one with these, that should help) but, according to CCC overdive my fans are at 100% but i can tell that it isnt, more like 30-20%. any fixes for that? also any programs that can monitor exact fan speeds?

 

HWinfo64 will monitor fans as well as GPU-z. Both will monitor them fairly accurately.

As far as temps, replacing the thermal pads is a great step, also the Clevo bios should allow the fans to be controlled automatically (the dell bios required manual control using HWinfo to adjust the speeds).

Let us know how the temps look and how the fans are going.

Good luck!

 

I was under the impression that the Dell cards required manual fan control, for example my ES cards.

I was not aware that you could flash the cards with the Clevo vBios as it still did not allow the fans to be automatically controlled and also gave graphical issues....

I may be wrong though...

All i know is that my fans are not at 100%, more like 30% so I need to use hwinfo32 for fan control.