*Official* Alienware M18xR1/R2 CPU Overclocking Thread - Learn How and Share Tips Here

Hey guys,

Here are some news concerning the CPU overheating.

I finally managed to have Dell support (I needed before to transfer the support to Europe because the laptop was bought in Canada and I mainly use it in France).

Now that’s done, I had a technician by phone that asked me to flash the Bios (I already had the latest one but I did it anyway) and I launched again a simple Windows 7 performance test.
When I said him I still had 97C on my proc with that simple test, he said a technician will come tomorrow on site to change the cooling system of the processor.
Don’t know exactly what he is going to do but I will have a very close look on what it’s done tomorrow.

More news on thuesday…

 

Sounds to me like a simple repasting of the cpu. Maybe a new heatsink but fairly straightforward.

 

Brother Good Bear, while it is apart, be absolutely certain that the thermal pad that belongs on the CPU heat sink is there. We have seen issues where this is left off of the heat sink from the factory on brand new systems and it is not always sent with a replacement heat sink as it should be. Things will overheat like crazy without that. This pad contacts the voltage regulators. It's the beige-colored pad across the bottom in the photo that I am referring to. If that is not there, the technician needs to be sure to install one there before reassembling the system. If Alienware sends a new heat sink out, have the technician look before taking apart your system. If there is no pad, have the technician call Dell and dispatch a thermal pad before taking it apart, unless he has an extra in his tool kit.

Please post what you find in this thread. If you can take a photo, please do so if there is no pad on your heat sink from the factory.

 

Hey fox i am just praying that my replacement m18x has that PAD. All my replacment CPU heatsinks & OEM new from factory never had that PAD. I think its safe to say there is quite a few m18x's out there without this thermal pad installed. You can actually see if its installed through the grill at the back of the laptop (near powerplug) in bright light.

 

That's good to know, Dan. Thank you for posting the tip. Now I am curious and will need to look, LOL. I know mine is there because I put it there, but I want to see what you are referring to by looking through the rear grille. That would really simplify identifying the problem of the pad not being there.

You can buy a pad (if Dell cannot provide it) and install it yourself, but that should not be necessary. You've had more than your share of CPU overheating issues because of this.

I mentioned this to Bill, Chris and Louis on the recent M17x R3 conference call so they could look into the issue.

 

yeah i had some pads from my m17xr2 days so installed myself too as you would remember i will try upload a pic later tonight.

 

Please do... that would be great. +1 rep for the tip also.

 

My exchange system... my current M18x... didn't have this so i used a thermal pad that was green from a mobile gpu mem module i had as spare. I hope this was good enough... And the first M18x had a grey one like on our Nvidia cards... Not beige so early ordered ones are inconsistent. Not sure about the original M18x that was exchanged which went through a mobo replacement and more at repair depot... So im not sure on that. So out of 3 ive used 1 had it, 1 im not sure and my current as it came didnt even have a pad. I took care of mine like Dan but none of these should have got to us without one of these in place. Another thing i was noted which seems a semi rampant issue that by default the way these GT580m sit with back plate... The supplied 3M insulator does not keep the vram heatsink on the bottom bracket of either gpu. 1 gpu was touching in 1 spot and grounding out and another had 3 I repeat 3 areas it was grounded out. This was an on site tech who performed the upgrade but it is not at fault of the tech the brackets are shaped that they touch a few of those small rectangular metallic pieces of the gpu to the metal bracket that holds the gpu heatsink... i fixed the issue myself to avoid the hassle i had (not usual for an Alienware customer) from last few repairs. i put some non conductive electrostatic safe plastic from a bag that held sensitive electronic device... I put it between braket and the chips the braket was touching on both gpus. I shouldn't of needed anything to be added just like the thermal pad. I hope Bill can make note of that issue too? Dan or Mr.Fox one of you posted pic of your 580m with yellow plastic insulating is that something you had to add or are they supposed to be installed that way as in do they usually have that plastic or did you add it? It was frustrating and im still unsure how well my gpus are doing as at low gpu clocks in games i see artifacts where some are saying most handle 720mhz on gpu at .87v mine show artifacts at just a little bump up from stock. Still testing but stock may show it too. My system before i knew there were issues in game was just shutting off... Launch game 3d image would appear and a second or two boom system shut off. Since i fixed where gpus were grounding out it hasn't shut off but still unsure if gpus were afflicted by this issue at this point. No hardware is guaranteed to overclock nor is it supported by tech support... so on stock though we shouldn't see white dots and sometimes black missing textures? Also sadly temps were just outrageous before i checked on stuff some of the brackets weren't even touching the vram with thermal pad etc. I love my system dont get me wrong and I fully support the outstanding job Dell does in Tech Support etc... but I hope these from factory probs are identified and taken care of. For me now knowing whats wrong and knowing what i am doing inside a computer doesnt help me if i go to ask for support... We arent supposed to be in our systems like this atleast not cpu or gpu... As technically these are actions that void our warranty. I could have called and reported how it just shuts off but I had multiple repairs, a exchange in the first system, repairs... Depot then a replacement and the another motherboard replacement and gpus installed... Its just a lot to go through and for those of us like MrFox and etc that know what we are doing... what am i to do armed with this knowledge of the situation but stand at a mercy that if i state the facts i save time pointing out the issue but at risk of being told they have voided my warranty based on what i theoretically stated. Words of encouragement from Dell would make me feel better lol I appreciate Bill being here along with others to help support us in the forums too. Thank you Bill and company.

 

The transparent plastic insulators you are referring to on the GPU back plates are factory-installed. I posted photos of those before. The problem is, sometimes they get overlooked during assembly at the factory, the same as the thermal pads on the CPU heat sink. I think it is the Chinese factory workers that assemble Compal systems just don't care whether the work is done correctly, and I suspect most of them are non-technical people that are taught to do something they really don't understand. Combining lack of skill with lack of care will naturally result in an inconsistent and sometimes poor outcome.

 

Interesting. I noticed no difference with or without that pad. In fact, the temps went up a bit with it, so I removed it eventually.

 

My m18x is on the way and I will check this as well.

On my m17x R3 the thermal pad is not installed, but the temperature of the CPU at idle are about 51-52 degrees celsius.

Would it make a difference if I install one? I am a bit confused now, as Aikimox mentioned that the pad does not make any difference in his system?!

 

[offtopic]

Spoiler

that makes plenty sense to me. I just hope for others Dell can take care of the issues... seems since its all outsourced work they have little control over the quality control of the product and i definitely do not blame Dell for that but chalk it up to what you said it most likely is. I guess for now i will take care of it and if i have any work done ill point out how its supposed to be. Thank you thats strange the yellow plastic wasnt in the upgrade kit sent with the on site tech all was in the kit but that yellow plastic. It had both brackets, the 3M insulator for each bracket the left; right heatsink and the two gtx580m. Thank you Mr.Fox for clarifying how its supposed to be from factory. None of this has changed my opinion of my Alienware... i love this system. Mr.Fox do you know of a good method to confirm my gpus are operating fine? Despite good temps i still get issues what perhaps would be good test bed. I dont use furmark anymore on systems as soon as i learned of the issues that can arise a year or two back. Ive ran Vantage and benchmarks and even vantage has artifacts stock clocks, anyone get a flicker in the water during first water scene as the watercraft drives by? Just 1 example of artifacts im seeing. Games like Witcher and Skyrim are more noticable but BF3 doesnt seems to push them as hard and shows no real issues...

 

Did you fashion insulators to go under corners of the back plate in lieu of not having them provided? The artifacts sounds like the back plate may still be grounded out on the PCB.

 

unfortunately yes I did and I used only what I had available... it was an ESD safe bag with plastic that seemed it could take a bit of heat without melting. when it was grounded it actually was shutting the system off and or restarting it when i would enter a fairly modern game. its not anymore. I actually went back in and redid it so this is the second round lol and im sure its "electrically insulated" now. I am now on the latest Nvidia beta driver I may go back to 295.10 to test again on and it could be just fine and an issue with one of my games. I think I can confirm if they are just fine on a mainstream graphic test or game... do you know the best app I could use to stress these 580m in SLI (on the .87v vbios) to make sure they are fine stock... something you guys might know is strong but at stock never seen a single issue. I've done Vantage on other comps and seen the slight flicker as described above so could be just how it displays at that part of gpu test... so anything better than vantage thats consistent and would help determine if they are fine? at room temps on .92v they got to 840core atleast and with good temps since i repasted. Mr.Fox I was using my comp as normal for about 2 1/2 months like that the whole time not knowing it was missing from my system :| They didn't crash up to that 840 core not being really cooler so they seem they may be okay... before I had tried the Witcher 2 or Skyrim for awhile light gaming on BF3 didn't cause issue but eventually... when I tried the other two one day playing it started happening... not one shutdown or restart since. Thanks for the help thus far I hope there's some simple way of determining if they are fine.... maybe the 295.10 driver is fine and 300 beta is not showing as good performance on 580m or my cards they are just low overclockers on stock volt?

 

for the record...mine has none there.

 

the plastic to insulate or artifacts in vantage or the games mentioned?
if plastic that's interesting if it came that way? Maybe my heatsink just isnt sitting right and the corners touch like Mr Fox says to check.

but if you mean the games and or vantage then okay.

Either way thanks for the info. Im halfway to understanding lol

 

my cpu heat sink doesnt have the added pad...my gpus have a little piece of yellow tape right by where the screws go in. without the tape the gpus ground out on the board...why, i have no clue..

 

okay, yeah on cpu i wasnt worried about that pad either. yeah that yellow tape was missing from the upgrade package... I used what i had here and they are finally not grounding out anymore. being it survives vantage at 840 core on .92v and showed no further issues in that maybe they're just fine and really like voltage lol and run better at .92v... even on lower clocks

thanks John again

 

no problem rv, you more than make up for any little tidbits i share here.

 

I try but there's also so much knowledge you emit yourself Tends to rub off on some of us

 

good times with my friends on nbr.

 

Just for info, Dell has not come yet. I called them back this morning and they said the pieces were out of stock and that the tech should come monday or thuesday with the replacement parts.
Also, as far as I know, they are going to change the fan, the heat sink and the past, even perhaps the motherboard... (don't ask me why).
I had a look through the vents with a torchlight as said DanXbix but I could not see any brown pad (however all is pretty dark in there and I am not really sure what I was looking at).

 

Good news. Better late than never, eh? - Oh, and Dell seem to have a passion for swapping the mobo at any given opportunity - might be simply to ensure that its all working but I know that they do like to swap 'em whenever they can for some odd reason (working or not!)...

 

yeah its frustrating sometimes... they do excessive mobo swaps but it sometimes can be mobo issues when it just seems cpu or gpu related and can just be a hair fracture in pcb and movement sets it off or a single capacitor etc is gone, capacitor melt out etc and even some freak electro static discharge tweaked something wrong.

 

Hey guys, I've got an M18x with a 2960XM. I didn't get the $80 option for factory overclock because I knew it was already there in the BIOS.
I've manually set the OC level to 3 and it was fine at first. But after about a week the CPU just kept overheating, and no matter what level I choose now, the CPU always overheats after about 30 minutes of gaming.
When this happened the first time I checked the temps and noticed core 3 was 10C lower than cores 0, 1 and 2.
So I figured it was a thermal compound issue, which I replaced the next day.
Everything ran fine again for about a week with temps topping out at 92C. Yet at the end of the week it hit 98-99 again, and thermal throttling started again.

I figured I may have messed up on the paste, which I applied by a single dot. Yet after repasting with two dots the problem persisted in the same way. It was OK for a week, and then it overheated again.

So then I did some research and found out some users reported not having a certain thermal pad installed on their heat sink, which was the case with me too. So I tore the machine apart again, put a thermal pad on the sink and repasted with a single line.

This was last night and the temps seem to be a little worse than before after the repastes.

And throughout all repastes I still have this weird issue of core 3 being 10 degrees lower than the other cores, which seems to be the biggest issue in my thermal struggles.

I'm starting to think the heat sink applies pressure in a weird way, with more pressure being applied to core 3 than the other cores.
But before I start messing with that (and possibly cracking my die), I'd like some input from you guys first. Next time I tear the machine apart I'm going to take off the thermal pad, but I'm not sure what else to do here.

Oh by the way, all fans and heat sinks look great with no dust or debris. I'd call Dell about this issue, but it doesn't overheat without the OC level, and since I didn't pay the $80 extra for the factory OC I doubt they'll overnight me a new heat sink for free.

Any suggestions?

 

Grr... Why do I only get a 7.6 on WEI even overclocked slightly past level 3? Everything else is 7.9 but my processor is holding me back. I can't seem to get my processor to stay near 4.5Ghz. It is set to do so and very rarely does but mostly just goes as high as 3.8 under load.

 

7.8cpu for 2900XM overclocked around 4.2+/4.3/4.4ghz (8thread 4core)... 7.7 below that and 7.6 stock speeds... you need more amps and watts man... you have to give it more Power!

 

try not fully tightening a screw or two if you have too to get your right balance... under 8thread 100% load you may still have a few C difference when done not 10 but a few as the way the cores are built into the die and the way the metals of the heatsink are oriented you may never actually see all 4 cores the same exact temperature.

Just get them close to each other and cool and your fine... don't torque your screws and your fine too remember its lightly hand tightened not meant to keep heatsink engrained to cpu die forever. I noticed same thing but since you've changed them now you'll need to get them atleast in check... after a few you might have it just right

easiest to check system without fully reassembling I use a usb keyboard and mouse and reconnect the lcd cable + power. then unscrew the power button from the plastics and plug that in and you can turn on system without it all back together first to check your temps

 

Hey mate, thanks for the reply!
Yeah I was actually thinking about doing that, because that's how I did it with my M17x R2. It just seems odd that one core is that much cooler, even when under load. I didn't torque the screws down or anything, just tried to get a general balance on tightness. Yet it's still that one core that's always too cool. I may try some trial and error this weekend if I've got time, and if I can find my old USB keyboard lol.

Anyways, I just applied some more TX-4 through the 'spread it around with a glove' method. I've never had good results with it, but it seems to work pretty good on this particular die. Running Prime 95 resulted in a max of 89C on the hottest core, so hopefully this will be last time I repaste.

Let's see what it does in a couple of days, though.

 

Sorry to interrupt here, I have begun to adjust some settings with ThrotteStop, mainly using this guide for reference, yet I am confused as to the max TDP/TDC or wattage that can be safely applied. I assume that 100-120 W would be about as most you would want to go and not have thermal shut downs yet in ThrottleStop under TPL/Power Limits the default is currently at 300 W (Power Package Limit). Now, I believe this is because I have made some recommended changes via the modified A05 BIOS a while back that unlocked the long limit duration among other things, but should I turn this down?

Thanks in advance.

 

It depends on your multipliers but using 85-122 watts should be good range for everyday clocks... When benching it can require into 120-170 range at crazy values but for everyday 24/7 overclock you'll need to just run a stressful 8thread benchmark and lower watts or multiplier accordingly so your not reaching thermal limits. If your worried about safe watts stick to 122 or less

 

Oh I am also have a slight problem now, although I have the multipliers set to 44 in BIOS while running WinPrime 1.55 the multiplier will not go above 36! While troubleshooting I thought that maybe ThrottleStop is running in the background and on "internet profile" but no it is not the case.

It is very strange, after it does any serious work like Prime it returns to 4389 hrz at idle (very useful). How do I get it to stay at the max multiplier while working?

This is very strange, likely the reason for my low CPU scores, (although even in Vantage they register 4.4 Ghz, see sig)

 

you need to have the Pri @ 900+ and Flex VID @ 5+ in bios. I run mine with a 42multi 900pri 0flex for daily use.

 

yes I can confirm that and Dan that's a good everyday clock i use that too...

 

Hey guys,

Here is some news about my CPU overheating issue.
The tech came 2 days ago and changed the motherboard (with a refurbished one) and the CPU heat sink.

I had a very close look at what was done (nearly 2 hours of intervention) and I can confirm the following:
- The original CPU heat sink DID NOT have the brown pad on the bottom of the heat sink.
- The new CPU heat sink also DOES NOT have the brown pad (I told the tech and showed him the pictures Mr Fox posted but he wasn’t aware of that issue, so I let him proceed with the pieces he had).


Now with the result:

1 - It seems the overheating issue is solved : I have made many tests (at stock speed for the moment) and the CPU does not go over 72C during benchmarks like PC Mark 7 or 3D Mark 11 with a max difference between cores of about 6C. When idle, I have 50C. That seems a very good point to me and makes me think the brown pad is not a factor of overheating (at least for my machine).

2 – Unfortunately, I now have a new graphic issue with the PC. The screen goes red instead of black.
I explain: when I boot, all the background is red; under windows, everything has red or cyan stripes/shadows, very disturbing... This is only true with the built-in monitor, on a connected external monitor, all colors are OK. Also, I tried to deactivate the discrete graphic cards (with Fn F7), and when using the processor built-in GPU, all the colors are OK again on the built-in monitor (but the resolution won’t reach HD and the graphic performances are of course awful), when I reactivate the GTX580s (SLI or single), I again have the same weird colors.

The onsite tech has no idea of what’s going on (motherboard again?). I also did the built-in diagnosis (Fn F7 when powering up) and everything went fine. I am awaiting a direct call from Dell support to see what can be done.

If any of you have any idea concerning this new issue, I will gladly take it.
This is not the end of the story…

 

as far as i'm aware Red issue is the SLI cable is not seated properly. Yes i have multiple repairs on my current m18x and no heatsink i received had the PAD. I confirmed with @DellChrisM and he confirmed there is a Part No. for this. I think maybe Bill@Dell or @DellChrisM needs to report this up the chain. Hopefully my replacement m18x has the Pad installed on heatsink.

 

Using just any pad seems to work the same... But even no pad or using a dab of thick paste works too... But if the substitute pad is much thicker than the original it will make your cpu heatsink sit awkward. Otherwise it's not been a huge source of heat... Johns cpu doesn't have one and he still has some top records and runs cool. It would be nice if Dell takes care of it though.

 

Well as it is right now Primary Plane is at 2400 and Flex VID is at 10, which is more allowable current and voltage?

Have you or anyone else reading this experienced the multipliers not able to exceed 36 although being set to 44 in BIOS?

Also a correction, it doesnt idle at 4.4 Ghz, its just appears that way in Hwinfo32, in CPU-Z it idles at 700-800 Mhz.

 

This is very bizarre, until recently I have yet to run a Wprime 1.55 test so I had nothing to compare the phenomenon of it not exceeding 36x in both tests (yes I have selected 8 threads) so I ran another 3DMarkVantage test and the CPU results are comparable with previous tests, with the exception that I am experimenting with increased memory (800 vs. 750), I must say that although I am mostly troubleshooting this is my highest Vantage score so far, particularly the GPU's (broke 30k....wow!):

NVIDIA GeForce GTX 580M video card benchmark result - Intel Core i7-2960XM Processor,Alienware M18xR1 score: P28283 3DMarks

To monitor the multipliers, I used ThrottleStop and logged the following, reflecting that the multipliers are where they should be:

Date Time Multi C0% CKMOD Temp Power?

HTML:

04/18/12  11:07:01  22.62    2.2  100.0  100.0    51888   54   1.2560    7.7
04/18/12  11:07:02  20.38    1.6  100.0  100.0    51888   56   1.3911    6.4
04/18/12  11:07:03  22.48    1.8  100.0  100.0    51888   52   1.2560    6.8
04/18/12  11:07:04  13.86    0.9  100.0  100.0    51888   54   1.2560    5.0
04/18/12  11:07:05  24.27    1.6  100.0  100.0    51888   56   1.2560    6.7
04/18/12  11:07:06  43.08    6.4  100.0  100.0    51923   55   1.2610   18.4
04/18/12  11:07:07  43.51   11.2  100.0  100.0    51923   74   1.3110   27.6
04/18/12  11:07:08  42.88   27.0  100.0  100.0    51923   78   1.3160   50.5
04/18/12  11:07:09  42.25    6.2  100.0  100.0    51923   58   1.2660   17.8
04/18/12  11:07:10  39.02    3.8  100.0  100.0    51923   54   1.2610   12.7
04/18/12  11:07:11  23.05    2.3  100.0  100.0    51923   54   1.2610    9.3
04/18/12  11:07:12  16.29    2.2  100.0  100.0    51923   55   1.2610    8.8
04/18/12  11:07:13  27.28    4.1  100.0  100.0    51923   53   1.2610   12.2
04/18/12  11:07:14  41.09    4.1  100.0  100.0    51923   54   1.2610   13.0
04/18/12  11:07:15  24.51    2.9  100.0  100.0    51923   54   1.2610   10.6
04/18/12  11:07:16  41.24   15.3  100.0  100.0    51923   71   1.3861   30.3
04/18/12  11:07:17  43.90   15.6  100.0  100.0    51923   71   1.3511   41.2
04/18/12  11:07:18  43.76   15.1  100.0  100.0    51961   68   1.3511   38.2
04/18/12  11:07:19  39.98   30.4  100.0  100.0    51961   75   1.3861   46.3
04/18/12  11:07:20  43.36   22.2  100.0  100.0    51961   73   1.3511   53.9
04/18/12  11:07:21  44.00   17.5  100.0  100.0    51961   71   1.3110   50.0
04/18/12  11:07:22  43.74   22.9  100.0  100.0    51961   69   1.3961   53.9
04/18/12  11:07:23  43.90   17.4  100.0  100.0    51961   66   1.2660   49.7
04/18/12  11:07:25  43.74   25.2  100.0  100.0    51961   74   1.3561   56.8
04/18/12  11:07:26  42.93   30.5  100.0  100.0    51961   79   1.4011   59.5
04/18/12  11:07:27  42.66   31.6  100.0  100.0    51961   78   1.3561   59.8
04/18/12  11:07:28  41.72   37.9  100.0  100.0    51961   80   1.3811   59.4
04/18/12  11:07:29  41.67   35.1  100.0  100.0    51961   78   1.3561   58.4
04/18/12  11:07:30  43.29   29.9  100.0  100.0    51961   72   1.3361   59.9
04/18/12  11:07:31  41.13   49.4  100.0  100.0    51993   77   1.3461   58.9
04/18/12  11:07:32  41.70   31.5  100.0  100.0    51993   70   1.2710   56.1
04/18/12  11:07:33  43.99   14.5  100.0  100.0    51993   72   1.3060   37.9
04/18/12  11:07:34  43.70   11.3  100.0  100.0    51993   53   1.2560   29.5
04/18/12  11:07:35  12.14    1.4  100.0  100.0    51993   51   1.2560    5.5
04/18/12  11:07:36  19.70    1.6  100.0  100.0    51993   51   1.2560    6.0
04/18/12  11:07:42  22.76    2.5  100.0  100.0    51993   59   1.3911    7.2
04/18/12  11:07:42  26.50    3.9  100.0  100.0    51993   62   1.3711    9.8
04/18/12  11:07:43  34.95   18.6  100.0  100.0    51993   58   1.2610   38.5
04/18/12  11:07:44  43.86   18.3  100.0  100.0    51993   70   1.3961   44.7
04/18/12  11:07:45  43.81   17.8  100.0  100.0    51993   68   1.3511   42.9
04/18/12  11:07:46  43.14   17.3  100.0  100.0    51993   73   1.3060   40.0
04/18/12  11:07:47  43.81   12.9  100.0  100.0    52035   65   1.3060   32.7
04/18/12  11:07:48  43.92   12.5  100.0  100.0    52035   68   1.3060   29.7
04/18/12  11:07:49  43.93   15.7  100.0  100.0    52035   65   1.2610   38.7
04/18/12  11:07:50  43.33   17.5  100.0  100.0    52035   64   1.3511   38.6
04/18/12  11:07:51  43.92   14.7  100.0  100.0    52035   71   1.3060   35.8
04/18/12  11:07:52  43.76   13.0  100.0  100.0    52035   66   1.2610   30.5
04/18/12  11:07:53  43.50    8.8  100.0  100.0    52035   51   1.2610   22.5
04/18/12  11:07:54  43.65    8.9  100.0  100.0    52035   66   1.3060   22.8
04/18/12  11:07:55  43.65   12.7  100.0  100.0    52035   60   1.3511   31.3
04/18/12  11:07:56  42.21   14.7  100.0  100.0    52035   69   1.1158   28.7
04/18/12  11:07:57  42.85   21.4  100.0  100.0    52035   67   1.3961   43.8
04/18/12  11:07:58  43.80   19.6  100.0  100.0    52035   69   1.3511   45.4
04/18/12  11:07:59  43.46   19.7  100.0  100.0    52035   64   1.2660   44.4
04/18/12  11:08:00  43.92   16.2  100.0  100.0    52035   65   1.1158   40.0
04/18/12  11:08:01  42.12   22.9  100.0  100.0    52035   70   1.3961   44.7
04/18/12  11:08:02  43.92   17.8  100.0  100.0    52035   71   1.3110   43.5
04/18/12  11:08:03  43.96   17.3  100.0  100.0    52035   69   1.3110   43.4
04/18/12  11:08:04  43.98   17.8  100.0  100.0    52035   67   1.3060   44.4
04/18/12  11:08:05  43.96   19.8  100.0  100.0    52073   71   1.3511   47.3

The ThrottleStop settings, were: TPL: Power Package Limit 1: 120 W, Limit 2: 80 W, Turbo Time Limit: 56 (as per BIOS), Limit 2: 0, enable limit: Yes.

EIST: Yes, C1E: Yes.

Temps, CPU: 84, 87, 86, 81

Temps GPU: 74 76

It is my understanding that to actually run a 45 multiplier with ThrottleStop I have to increase it in BIOS first?

I am on BIOS A05 and 45 is the limit, how are some achieving 47, 48+ Multipliers? (not that I am that adventurous).

I am thinking of upping the current via throttlestop and would like to use a 45 Multi for some CPU dependent games (i.e. Skyrim) as my temps still have a wee bit of headroom. Any recommendations? It is my understanding that in the TPL settings limit 1 refers to short duration limit and Limit 2 is long duration limit?

Still bewildered by the 36 multiplier limit with wPrime 1.55.

 

Update:

I increased multi form 44 to 45 across the board via BIOS and TS and also Flex ID from 10 to 12 which I will now dial back as I got some very high temps this last Vantage run:

NVIDIA GeForce GTX 580M video card benchmark result - Intel Core i7-2960XM Processor,Alienware M18xR1 score: P29483 3DMarks

CPU Temps: 91, 99, 94, 88 C

GPU Temps (same as before) 74 and 76 C

Post test CPU idle temps: ~45 C across the board.

But my CPU score increased 3k!

In an attempt to decrease CPU temps I am tempted to reduce Flex from 12 to maybe 8. I may also dial back the power right now its at 150 W short, 85 W long via TS.

Still locked at 36x in wPrime 1.55, any ideas?

 

Update:

Experimenting a bit, I got a few crashes or system freezes rather while playing Skyrim with multipliers at 45, so I upped Flex VID up to 25 thinking it may be a voltage issue, but in retrospect I am not sure if the CPU shut down at 100 C, where do I find the error codes after a black screen?

Anyway, increasing Flex ID didnt alleviate the problem so I went back in BIOS after last black screen of death and reduced the Flex ID to 8, from 12 where it was initially. I also dialed back Short Power Limit to 130 from 150 W. Left long at 85 W and reduced multipliers to 44 via Throttle Stop (leaving them at 45 in BIOS).

With these settings it achieved a very similar score but slightly cooler temps:

Result

CPU Temps: 89, 96, 93, 86

GPU Temps: 74 and 76 (pretty consistent)

Im open to ideas or advice, right now I am thinking of dialing back Flex VID from 8 to 5.

Update:

Still getting black screen of death while playing Skyrim, unsure if it is a temp issue, I have dialed back Flex ID yet again, down to 5, will try to back out of the game this time to check temps, hopefully before another black screen of death.

I am now wondering if this is a power starvation issue, kind of like that experienced with a highly overclocked M17x R2, going to dial back CPU wattage from 130 to 120 W which I should do anyway seeing as how even with the reduced voltage I am still seeing quite high CPU temps:

89, 96, 92, 86 C.

Update:

I loaded and played through a similar segment of Skyrim where I had previously had two BlackSOD' and exited after a similar amount of time (before BSOD) and the CPU temps were only in the high 70's with "CPU package power" at 62 W according to Hwinfo32.

I don't know if the BSOD' were heat, voltage or wattage related, or something else, very curious.

Anyhow, I will continue to use these settings as they yield great performance without very high temperatures.

Will do some more benching later when I am at "the hideout" (not to be confused with the bunker (see blog), every true super-villain/hero has an industrial themed hideout, in this case I do all my extreme benching and such in an unheated storage facility, the 50 F ambient temps do wonders for PC performance, but not so accommodating of the bodily extremities)

Correction: The monitoring software I referred to is Hwinfo64 not 32, but I'm sure you already knew that.

 

Great Success!

I dialed back Flex VID from 5 to 0 after 4 consecutive runs that all resulted in a 29,4xx using different multi variations via ThrottleStop.

Its a learning process, thank heavens Intel Core i7 chips are built like Sherwin tanks.

Vantage 30,140

Oh and because I am benching in "The Hideout" (see previous post) the temps didn't get too high:

CPU: 84,92,87,82 C

GPU's: 73 and 73 C

Settings:

BIOS:

Primary Plane: 2400

Flex VID: 0

ThrottleStop:

profile 1: 44x (all 4 cores)
profile 2: 42x " "
DTS Alarm and switch to profile 2: 10 C under thermal limit (90 C I believe, may increase to 15)

TPL: 120 W short, 90 W long

56 second limit enabled.

long limit disabled.

GPU's:

800 shader/800 memory @.92V.

I know there are better benches out there, I wasn't shooting for a high bench, I am trying to find the maximum practical clock settings for some of the more demanding games out there, using throttle stop I have two other profiles that I use for other tasks with much reduced multipliers and one with them set at 27x for battery use only.

For comparison, my old Vantage scores after initially experimenting with overclocking:

26,934 (23k CPU) vs. 30,140 (28k CPU)

Correction: I believe they were Sherman Tanks, I think Sherwin designed ballistic underwear for the American Forces involved in trench warfare (I'm kidding, I have no idea where I got Sherwin from).

Update: Still crashing in Skyrim and only in Skyrim at this point (Crysis 2 taking it like a champ, CPU temps dont exceed low 80's either (only one core too), I wonder if it is the transition between C States as the author of this guide describes?

Going to "experiment" for a few hours with CE1 disabled via ThrottleStop. If anyone has also experienced the system essentially shutting off whilst playing Skyrim with a vigorous OC and discovered the culprit please let me know.

 

Just received my replacement m18x and the thermal pad is missing AGAIN in the new one! What is with this?? Can the dell colleges on the forum post the part number.

Sent from my iPad3 using Tapatalk HD

 

Brother Dan - send a PM to AW-FedericoH and maybe he can assist you with getting a pad sent your way. I don't understand why they keep leaving out that component.

 

In reference to the issues I have been having with Skyrim, I disabled C1E in ThrottleStop and still got a shut down within about 10 minutes, I dont believe it is a thermal issue as I have had the chance to check temps before this happens before and they are only in the mid to high 70's. Not sure if this is a PSU issue either, right now I have wattage set to 150 W short in throttle stop but it has happened even at 120 W, and again it isnt as though that much wattage is even being drawn.

Going to experiment with Throttle Stop inactive, although this was the whole purpose of even experimenting with Throttle Stop in the first place.

Skyrim PC shut down/reboot

Update:

Problem may be solved, possibly too high of GPU OC

http://forum.notebookreview.com/alienware-m18x/624048-m18x-skyrim-performance-problems-general-discussion-thread-25.html

 

because....if that pad is not placed right, it will cause the cpu to over heat as well...

 

thanks fellas back to square one with this replacement. missing PAD & bad DPC latency. will pm AW-FedericoH

 

Alright feel quite dumb new machine did indeed have 26x series Nvidia drivers (dell cert) installed from factory. I must have been looking at old unit and confused the two lol.

Anyway have contacted AWfrederico re thermal pad. The new machine is having trouble holding a 42x 900pri 0flex load which absolutely sucks and so sick of the DPC I'm getting no matter what drivers I'm Using. Don't really know what I should as I have both m18x's currently and not sure if I can request another??


Sent from my iPad3 using Tapatalk HD

 

Dan, silly question here Bro, but have you tried Nvidia 296.10 whql - no need for the Fox-Fix with them....

 

Hi guys,

Another technician mandated by Dell (not the same as last week...) came yesterday and changed again the motherboard (with another refurbished one).

It seems both my problems are now resolved (no more CPU overheating and no more weird graphics on the built-in monitor).
Now I will gently make some tests of overclocking my stock speed 2960XM and GTX580s.

There is so much stuff on the forum (110 pages only on the present subject) that I am a little confused on how to do that simply.

I any of you has some direct links to essential tutos on how to overclock the CPU and the GPUs I will really appreciate.

Anyway, thanks again to all of you for your help, especially to Mr Fox and StevieJones.