M15x & 6970M / 6990M - What you need to know - upgrading, problems, solutions

I have i7 620 that only have 2 cores (4 threads) and still having shutdowns.

News. Right now I have tested Crysis I, the procedure used was as the follow:

1. with ATT opened i blocked clocks on 100/150 0.8v
   
   
2. Run crysis in low: no crashes min 11 max 12 fps (very stable! xD)
   Run crysis in Medium: no crashes min 10 max 11 fps
   Run crysis in High: no crashes min 7 max 10 fps
   
3. go back to crysis menu. Then back to windows (crysis still opened but minimized)
4. increase clocks to 400/600 0.8v
5. select the same level and play with Very high settings and NO CRASHES average 20fps in 1920x1200

Tomorrow more test.

 

Alright guys, I need help. I cant get my core clock to go over 400. In GPU-Z, it shows everything else is fine but core clocks don't go over 400 and I get terrible frame rates...obv. It used to work completely fine, I had no throttling or anything. Heat never went over 75C. Anyone know what might be causing this? Thanks

 

did you try power drain and toggle steath button yet ?

 

Oh I have 620m too...so that means it won't work?

 

Certainly i think it shouldn't work. If jonnyflash thinks to limit the number of cores to decrease consumption he should to know that i7 620m have only 25 watts in the m15x. Maybe intel specifics 35wats in his charts but that is using a integrated graphic, that graphic is dissabled in our laptop.

 

I think we're running low on ideas.

 

User Account Control settings.

 

Here's my GPU-Z log with BF3 till shutdown.
Playing campaign mode, the drop in clocking is when the game changes map (the loading screen).

 

Well try configuring with ATT 720/950 at 1.05v, i tried this config a few days ago for a 2 consecutive hour without problems.

Ps: only proved with bf3 and sc2, don't works with crsysis 1 or 2

I think to have great news for all:

http://i.minus.com/1332613610/hPhfGusEW-tmRNEYhbw69g/dbw95QxvEjeEEP.xlsx

These are my own results using Ati Tray Tool to manage voltage and clocks.

Ok, until now I have had 2 dyas to put in proof the new clocks and voltage. I haven´t shutdowns at all with 580/950 0.9v (580/900 are clocks of 6950 wich have the same specification, i readed both and 6990m are the same graphic card, only change vbios/clocks and voltage). I played several hours at Starcraft 2, Battlefield 3, Crysis 1 and Crysis 2 and I haven´t any problem.

I don´t know if these clock works loading it in vBIOS, I only tried modifying it with ATT.

Now, to complete the work is necesary to someone would do test.

things left to do:

1.- Someone can use the DisplayPort? Someone can help to activate it?

2.- Why there´s no fan control? How we can activate automatic control?

 

This is a really odd problem. It seems many have this problem but some don't. Myself and mexic00ls don't seem to have issues with throttling.

Could it be the 6970M causing it vs the 6990M that we have or down to motherboard revision?

Perhaps we should start comparing hardware revisions.


If this indeed the case then perhaps I can trade my card with someone here for a very short time to see if the 6990M is throttle friendly... I don't really push it hard ever so I shouldn't miss it for a couple of weeks. Obviously means taking it out etc but I can go that far for the community if there are no ideas on how to solve this.

 

i would fix the shutdown issue first before the throttling issue

 

The 6970m and 6990m makes no difference, at least both of them work perfectly fine in my system, at least at stock clocks and voltage.

I don't think the shutdown issues some people experience are caused directly by the GPU. Otherwise it wouldn't work for some users (e.g. me) while others still are having troubles.

The CPU can be excluded from the troubles as well imo, it didn't make any difference for me, and i5 owners have reported that it works fine too if I remember correctly.
One of the last things I can think of is, that some other parts of the system overheats / has too much load on it. It's unlikely but still possible that some mobo revisions have a slightly higher bus bandwith, at least in some parts of the bus and that the additional load of the 69xxm (yes, that's a lot of additional load, at least when playing at high details / res) might cause a controller of some mobos to overload or heat up too quickly.

I can't remember whether this has been done, but the users who experience shutdown issues should check the event log, maybe there's some helpful information in there, I don't know.

 

i get voltage readings with the new gpu-z 0.6.0

 

I do have the old M17x PSU which is 240 watts. And to the guy who asked if I did in the heatsink mod, I have not, but ironically, my heating issues do not lie with my GPU. My GPU rarely exceeds 80C, it's my CPU. And I'm too scared to test it because whenever I run benchmarks, my CPU gets so hot my PC shutdowns to prevent damage, and the temps still increase even if I've closed the benchmarking program or something. I just feel like in games like BF3 and SC2, I'm not able to keep the settings I want to play the game at, because after about 10-15 minutes in game, my PC starts to throttle and I'm forced to lower the settings drastically. For example, I can run BF3 with all ultra, medium shadows, no HBAO, no motion blur, and no AA and get an average of 45-50FPS which is pretty darn good. However, as my CPU heats up, this drops to like 30ish, and I've repasted and done lots of things, but it seems to no avail. It seems that this CPU heat problem is only prevalent in games that utilize CPU as much or more than GPU, because in Skyrim I do not have this issue at all, I run the game at almost max settings with HD textures at 768. I've tried to "dial" in my chip but I've just had too many failures trying to find the right amount of OCing using throttlestop. Anybody willing to fly out here to Socal to do it for me is welcome, I'll cover your airfare gladly

@ Kurupt: No I have not, but like I said up top, GPU temps are not my problem, my 940XM is just getting way too hot.

 

Kazalor what are your tdp/tdc settings? Also what turbo ratios have you set?

You need to start at stock settings and tdp/tdc and work your way up. Also use ICD 24 I use it and my temps are really low.

 

@mexic00ls really huh? That will be nice to law to see if voltage fluctuates while gaming.

Sent from my SCH-I405 using Tapatalk

 

Does anyone know from where exactly on the card the memIO sensor is reading and if there is a way to specifically target that part of the card to cool it?

 

It's the memory controller, which is located on the die.

 

Currently I'm at 17x turbo rates, and 62 TDP and 60 TDC

 

I'm very interested in upgrading to a 6970/6990. I was just wondering if the shut down issues is affecting everyone?

 

Not everyone no. At least 2 of us don't have issues. Myself and Mexic001s. I don't game however ever. I have tested instead with furmark which is more exacting than a game and found out this way that I am lucky.

 

Furmark is stressing the GPU less than a game, it mainly stresses the power supply circuits, but not the graphic processing logic. It's an easy way to shorten the life of a GPU, but not to test stability.

Edit: Most people don't seem to have shutdown issues, a lot of users did this upgrade and never reported any issues, e.g. The Rev, widezu, Quadzilla, dttran, ... to name a few. Ah, and I don't have issues either...

 

is it just me that i see those without shutdown are using Solid State Drives?
i don't know, just throwing all the possibilities now since i'm stuck in a corner.

 

Just to mention.
I have 6970m and 920XM for over 6 months now.
After good repasting no shutdowns what so ever.
With heavy gaming Red Orchestra 2 - Fallout 3 - Serious Sam 3

 

@eXplosetoi can you post your specs so i can have a look at it? also did you overclock at all

 

Use to overclock for the fun.
But as I don't need it I stopped.
Got 60 Fps on all games with almost everything maxed up.

 

SSD or not eXplosetoi?

Could be a potential factor but then again I still run a HDD inside the optical bay so power draw should be greater than a single HDD and ODD.

 

No SSD just ATA

 

I had shutdowns even with SSD. What did it for me was the thermal pad covering the backside of GPU card completely. And no extra padding on the other side (heatsink side). Just stock pads on rams.

 

I just tried the Witcher 2 and before i could even complete the tutorial i got shut down. So far i have shut downs on

The Witcher 2
Skyrim
Alan Wake
BF3 (multiplayer)

I haven't had any shutdowns with SWTOR, Batman AC, Mass Effect 3, Modern Warfare 3

I'm so damn frustrated that i'm really thinking about saying screw it and just reselling this damn card on ebay and going back to my old 260m. I bought the card to play high end games and when half of them are shutting down it's just a waste of money.

 

Just an update, i disconnected my external monitor which is using a displayport-DVI adapter and haven't had a single shutdown yet with the Witcher 2.

I manually set the clocks to be the same regardless of external/built in display. Not sure why i don't get the shutdowns when running directly off of the alienware monitor, but it works.

 

Not read all 200+ pages so question: since you guys upgraded from a 260m/460m/5850m to a 6970m/6990m do you find the thermal cooling of the m15x stressed more i.e. fans kick in much more on idle and light gaming?

 

On idle - no. Same behavior. But as soon as theress some stress on the ah
GPU it gets clearly warmer than with the cards that are officially available for the M15x.

 

Maybe less power being used.

Iget shut downs on swtor when my psu gets real hot.

 

do you guys think the shutdowns could have something to do with the ac adapter?

We've looked at all other possibilities right? I don't know crap about power draw and stuff like that but perhaps the adapter is just getting too hot or something? I know if I disconnect my power adapter, even with a fully charged battery my laptop will have a hard shutdown after a few minutes or so, which is strange and i never thought it was related but perhaps it is.

 

Sorry to break your bubble but... *pop*

I use 240w adapter and still has shutdown. Anyways this suggestion has been mentioned before in this thread.

 

Hi everyone,

A few days ago I bought to @makwu a second 6970m card, this time from Dell (my previous card was from mxm-update), I received that today and I installed on my M15x without problems.

Previously, I uninstalled drivers and Catalyst.

When I turned on the laptop, I saw this: http://img696.imageshack.us/img696/618/imag0367zz.jpg

Windows appears with visual failures. Drivers don´t fixed the problem. Temperatures are low, arround 35 to 50 ºC.

Any idea to fix it, or return the card to @makwu?

 

looks like the card is faulty to me :/

 

there is no fix for those kind of problem , I'd you should return it.

 

Ohh, that´s sad because was only 300$, anyone know where I can find other card from dell?

 

Hello there,
I have just bought a dell 6970M and I am wondering if this backplate would fit properly to the card itself?

GPU Spreader / X-Bracket for MXM 3.0b AMD HD6970M/HD6990M VGA Modules | eBay

I went through some of the posts stating that the x bracket of the 6970m/6990m would cause some problems and just wanted to make sure.

 

Yes it works i am currenty use that backplate make sure you request the screws for it

 

anybody have any more ideas on how do i fix this shutdown issue? i don't want this thread to be left dead and new people thinking that the issue was fixed.

should i revert to 11.5 and see what happen?

SSD - people says still shutdown
RAM - do people with 8gb have shutdown too?

 

Use 500/900 at 0.9v. No more shootdowns. I have a vBIOS modded with this info if you want.

 

so the only way for us shutdown users is to downclock the card? i mean then what's the point of buying this card when you can't make full use of it.

@phobos don't worry i know how to downclock and mod vbios, been experimenting with it far too long.

to those who haven't had shutdown AT ALL with ANY game. can you pretty please post your specs here so i can compare? as in every computer specs. i know it'll be too tiresome to do, but please help me with this since i'm so desperate to get it working.

throttlestop settings, AMD graphic settings, BIOS settings would be helpful too.

 

To prevent shootdown without downclocking the only way is to do the heatsink mod but it works with dell card, i dont know if work with cards from mxm-upgrade or others.

My system is 620m 8gb ram bd-rw 500gb momentus xt and 6970m from mxm-upgrade (dell 6970m arriving).

With downclocking you can overclock to 550/950 at 0.9v but in many games, for example in crysis you can have shootdowns.

The problem is no caused by power consuption or amd processor overheated, but is caused by memory card overheating. The temperature measures from the memory must be wrong in this card.

Enviado desde mi HTC Vision usando Tapatalk

 

hmm, do you mean the RAM or GPU RAM
either way, i'm already looking in these 2, but the first choice can be ruled out since those who doesn't have shutdown use the same number of GB. still i'm unsure about my RAM brand which is samsung. i prefer kingston brand to be honest.

anyways, i already did heatsink mod since the week after i found out about the shutdown (last year), and still shutdown. oh and i have eurocom card (no soldered post one).

 

I mean GPU RAM is overheats.

My system RAM is kingston and I have shootdowns on stock clocks, but I haven´t any with 500/900 .9v.

your heatsink mod includes thermal pads?

 

I'll go on record as saying i've had my external monitor disconnected for over a week now and i run my card at 680/900 1.0v without a single shutdown.

I'm going to try plugging the external monitor back up and running it at 500/900 .9v as suggested and see if i get the shutdown.

 

I am looking to sell my system with these specs.

radeon 6990m
16gb ram
500gb HDD
1080p screen
i7 740qm

I am not sure if this is where I should post it, but since it is modded I thought here might be good. Also I would like to get some input on how much the system is worth... although I have a number in mind

PM if you have any questions or are interested.