*OFFICIAL* M18xR1/R1/18 Owners "Are my CPU/GPU Temps OK Thread" - Ask Questions/Get Help Here

Trooper, do you have a high ambient temp in the room, or something that might be interfering with the airflow on the laptop?

 

I have no idea why the GPU is running at that temperature ...it seemed too warm for my comfort ....what can I do ???

I don't think room temps a problem but I do feel that I have my m18x kept on the table ...kinda close to a wall ....not touching it though ....maybe that

Also GPU idles about 40-45C

New Edit: PLayed the game at night ...........definitely the cooling in the apartment is a shade better and i did push it further away from the edge of the table and hence the wall too .........and after a hour plus ..on the same game ...max temps were 75C which are definitely better i guess

 

Yes, 75c is indeed waht I would call acceptable gpu temps for gaming with no overclock. Must have been a combination of lack of decent cold air flow to the fan intakes and warmer ambient room temps that were giving you high temps before. Keep an eye on things, but if it stays like that when you have adequate ventilation, you should be just fine.

 

Alright! alright! I will get a cooling pad!

 

Hope I am not starting a redundant thread, but I didn't see anything on this when I looked. I've noticed recently that when I have a 2nd display attached either to the VGA port or the HDMI out the temperature on one of my GPUs rises drastically. I have a M18x R2 that I purchased last summer. Quick stats are i7 3820QM, 2 680M in SLI, 32 GB RAM and 1.5 TB of hard drive space.

The temperature rise starts immediately when a second display is connected and rises up to about 90C on one card only from a starting point of about 50C with no additional display connected. The other card does not change temperature. I don't remember it doing this last year and I first noticed it about a month or so ago when the left side of the laptop was almost too hot to touch.

I was just wondering if anyone else has experienced this or has any thoughts on it. I'll attempt to attach GPU-Z screen shots.

Thanks.

 

Welcome to the NBR Community. Glad you joined us. Congrats on the M18x beast you have. Please have a look at the link in the right side of my signature and you will find an index of useful posts. Please take a few minutes to read the forum rules and how to search the forum effectively using Google. (Both of those are linked at the bottom.)

We have a thread for questions about CPU and GPU temperatures. I will merge this thread with that one.

If you haven't already done so, remove the bottom cover. Take the screws out of all three fans (no need to unplug the wiring) so you can lift them out of the way. Vacuum out all the dirt and lint from the heat sink radiators and fans. When you vacuum out the fans, hold the blades so they don't rev up to high RPM, as this can damage the fans. Needing to be cleaned is the leading cause of CPU and GPU overheating.

 

Thanks Mr. Fox. I cleaned out all the fans and vents while I had the cover off. I've been working on computers for many years as well and had thought about cleaning the vents and fans, but thought I would post first. What was throwing me off I guess was that it was only doing it with a 2nd display running and only one video card was heating up too much. I had figured clogged vents would cause them both to heat. After booting back up and hooking up a 2nd display the card started heating again, but this time when it got to 64C it started to fall and settled at 56C. Much better than 90C!

I've been lurking around here since shortly after I purchased my M18X. I don't think I've ever been happier about a computer purchase. Anyway, thanks again for responding. This is one of the best places I have seen for discussing various issues. I often come browse the forums even when not actively having a problem just to see what others are discussing. You can really never stop learning about these machines.

 

Cool! Glad that fixed it. Thanks for sharing the success story, and again, welcome to the Community.

 

Is there a way to run the laptop benchtop mode to tweak with temperatures (ie only external display connected) so I dont have to close the whole thing up to see what temps I have.

 

Sure you can. Terry702 runs his all the time only partially assembled. You can set the display hinges in their pockets and connect the LVDS cable to the motherboard (or even connect the rest of them if you want it fully functional with wifi, webcam, etc). Remove the power switch from the palm rest and connect it to the motherboard by the right GPU.

Below a photo of how Terry runs his system. Notice the absence of the I/O daughterboard, optical drive, etc. If you ever wanted to use dry ice for cooling, you could do it this way. I would imagine that having nothing covering the motherboard, CPU and GPU would make an improvement in and of itself. You could run a fan blowing air across the top of it to speed up the convention process.

Spoiler

 

Guys I am running 4.2ghz with 0 flex and 0 pri plane, 100.3 BCLK. Ambient temp around 50-60F.
When I run the TS 1024 benchmark I see temps around 93 with what may be comparable to a semi-raised back (the back is on the very edge of a table and air flows out freely).
I was under the impression these temps are about right but going to 4.3ghz (requires me to increase flex from 0 to 12) I see temps around 96-97 which is too high for comfort.

Is this a little high? I'm thinking maybe I should have screwed the heat sink tighter..? Or these are simply the shortcomings of having a QS chip. I should also state I am using standard fan profiles.
What do you think?

 

Did you tighten the screws until they stopped, or sooner? I'm not sure how much hotter (if at all) the 3920XM runs, but my 2920XM isn't even a QS, it's an ES. I doubt that's your problem.

If you're going to run those types of stress tests, it is best to force the fans to full speed. That is unless you're intentionally trying to see how well your system cools without the aid of fans forced to 100%.

If you didn't tighten the sink down completely, I'd start there.

I even went as far as placing a very thin washer on top of each of the CPU heatsink springs for slightly higher compression. I didn't do any before/after tests, so I don't know if this helped or not.

It seems like the thermal paste on the CPU has to be changed pretty often with these nasty XM CPUs. The heat dries it out quickly, no matter how great the paste is.

 

I'm certain I tightened it as as tight as they were before I unscrewed them, so I was wondering if I should have given it a little more. I used the Dell TIM so I won't have to go in and repaste so soon.
With my benches I am running them to see how cool/fast I can get it if on a standard fan profile.
With your help turbo,I think I've probably answered my own question here lol.
It seems that going beyond 4.2 requires a custom fan profile, cool weather along doesn't seem to cut it,
What do you run at 24-7 ?

 

I have the same triple pipe heatsink that I'm guessing you have, and I tighten each of the four screws until they come to a noticeable stop.

I run at 4.2Ghz with no extra voltage. During any benchmark or stress testing program though, I force the fan to 100%. I also have to force the fans to 100% when I'm playing Crysis 3 now too. Once I start adding voltage, the temperatures begin to skyrocket, which to me is anything in the mid-to-high 80s C.

Once my thermal paste starts to dry out, I usually turn off all overclocking until I decide to tear the machine apart and repaste.

We are talking 4+Ghz on a mobile quad core, which is more than enough for everything. Accept benchmarking, there's no limit there.

 

Thanks for that, yeah I am using Throttlestop to manage things. All stock voltage with a 100.03 BCLK in XTU, My performance profile is 46,44,43,42. My gaming profile is 45,43,41,40. Using the TS 1024 bench I see ave temp 93 and 89 respectively.
So I guess 43 + is off the cards until I improve airflow.

 

Hello everyone.

My Alienware M18x R1 has been acting up lately and I would appreciate if there's someone who could recommend some way to fix it.

Specs:

Intel Core i7-2760QM 2.40GHz
2x Nvidia Geforce GTX 580M
Windows 7
Bios version A05

GTX 580M bios version is 70.24.2B.00.06

The problem: Primary card heats quickly to 90 degrees celcius, leading to fps drop in games.

So I noticed the problem about 2,5 weeks ago when I was playing League of Legends and the fps would drop from 60 to 20 suddenly. After that I immediately updated my nvidia drivers to the latest, but the problem persisted.
I remembered that hadn't cleaned the fans in a long time so I decided to do that. There was a lot of dust in the fans and after the cleaning CPU and GPU temps dropped 20 celcius both and after I was able to play without any problems and primary GPU stayed around 70 celcius max 45-50 being idle temperature.

My friend had recommended that maybe we should replace the thermal paste in CPU and GPU's and we did that just 2 days ago. We used good quality silver paste and everything went well. I was able to play with decent temps but now the problem has started to come up again.

How long it takes to reach that 90c depends on the graphics settings, but eventually it seems to reach that number. After adding the new thermal paste the GPU cools down much faster than before after high load and idle temps are between 39-45c. Also CPU temps are good, 42c idle and going to 65c when playing games. So there is no crashing, blue screens or anything, but just the card getting heated up too fast.

For me it seems like some kind of bios problem or driver problem, but I'm not an expert.

Thank you in advance!

 

Which drivers are you using?
Try clean installing the latest 314.22 WHQL.
Uninstall everything, reboot safemode, clean the rest with ccleaner and driversweeper.

Do you use a cooler?
If not, try propping the back of the laptop up with a book, without blocking the airflow.
It would also help to know your room temperature.

---

I'm dumb. you're already on the latest drivers.
Did you clean install though? It's worth a try if you didn't.
Also, when you cleaned the fans, I asuume you also blew out the dust from the heatsinks.

 

Yeah I did the clean install again just in case, but no change. Yeah I also cleaned the heatsinks.
I use a coolermaster U3 cooling pad, which helps to slow the heating a little bit but doesn't really prevent it. Also my room temp is around 21c so quite cool.

However when you do clean install are you supposed to delete the PhysX too and with driver sweeper also clean the chipset etc? I only cleaned the display drivers.

 

I don't think deleting physx matters but no, you don't delete the chipset drivers.
It's weird that replacing the thermal paste solved it but only lasted for two days.
Do you know which brand he used?
Not saying it's the problem. Any thermal paste should last way more than 2 days.
Did you try other games? Does it only happen in LoL.
I don't know if LoL supports SLI. It's only your primary card overheating, right?
What are the temps like on the secondary card?

 

Actually the problem was solved after I cleaned the fans myself and about 3 days back we replaced the thermal paste and the brand he used was Arctic Silver 5. The problem has been with both LoL and Starcraft 2 so far and I haven't tested other games. With LoL it takes about 30 minutes to reach 90c and with Starcraft 2 only watching a cinematic for about 15 seconds get's the card to 90c.

When I tried Starcraft 2 with SLI the secondary card went to 67c max and otherwise stays at around 50c.

Now I think that when I cleaned the fans I think I had another NVIDIA drvier version. Maybe I should try going back few drivers and test if it helps. If I don't find an answer soon, I think I'm just gonna try to format the system.

 

I doubt reinstalling Windows is going to help.

Are you actually hearing the fans? With the GPUs hitting 90°C they should be running full blast and roaring loudly. This will be audibly unmistakable If you're not hearing them, then that's the problem. The GPUs should not be getting that hot, and if the fans are not running, they will quickly overheat under load.

If you are running HWiNFO64, if something is messed up with that the fans will never come on. If you are not running HWiNFO64, install it and manually max out the fans and see if you still experience the same overheating problem.

 

Yes both fans can be heard and I tested that with HWiNFO64. Even with both fans doing 3800rpm the card still heats very fast. The weird thing is that when I first encountered this problem and I placed my hand behind the primary card air hole I could actually feel the heat coming from there, but now even when the temp is supposed to be 85 degrees, the air coming out feels like it's around 65-70c.

My friend adviced that maybe we should open the laptop and check both GPU's and possibly re-paste them.

 

Possibly so... I would use IC Diamond this time. Don't spread it. Use a thin line the length of the die and allow the heat sink to "squish" it.

 

Is making a thin line the general way how the thermal paste should be placed? My friend covered the whole die with a very thin layer.

 

I wouldn't advise the "thin layer" approach as it could trap too much air between the heat sink and the gpu/cpu die. The application would have to be completely "glass-like" to avoid microscopic air bubbles when assembled. With the line or dot approach, the heat sink squeezes the paste outwards towards the edges as pressure is applied, ensuring no trapped air pockets. That's the theory at least. A bit like why scientists/doctors etc would place a drop of blood on a microscope slide and then place the thin glass over the top so it spreads out. That's the best analogy that i could think of.

Also, line method for rectangular die's and dot method for square ones - that way the paste should be squeezed out uniformly to each edge of the die.

 

Have you checked your environment? Try a laptop cooler or elevate your laptop on a bunch of books. Lowers temps for me by 10C easily. Also guys I have a question how often should a person perform a repaste on ICD paste?

 

I'd only repaste if temperatures started getting high, otherwise, why fix something that isn't broken?

 

Sounds like the heat isn't being conveyed to the heat sinks, which would likely be the result of poor thermal contact, so yes replace the thermal compound ASAP.

I recommend IC Diamond, line method on rectangular die, such as our CPU and dot method on square die, such as our GPU's.

 

Alright. After trying going back in Nvidia drivers, it seems that drivers do have a little effect. First I went back to 306 December drivers and I could play all games in SLI mode, but with 1 GPU it started to get a bit hot, but still playable. 310.90 drivers are a bit better, but still with 1 GPU card starts to get near 90c but I'm still able to play LoL and Starcraft 2. I'm going to try few other drivers soon too to see what they do.

 

Exactly as Brother Steve says. Repasting should be done only on an as-needed basis. GPUs very seldom require repasting when done correctly the first time. Once effectively applied, keeping the fans and heat sink radiators clean is generally the only maintenance needed to keep GPUs running cool.

 

Brother Fox and everyone else. I have a CPU question. My 580m GPU's are running cool (stock) at 70C on both. However my i7 2760qm is running at 80-85 on the cores through HW monitor after I run 3dmark11. Is this OK? Gaming usually has 2 corese in the 70s and 2 cores at 80. Any reason for concern? I may be due for a repaste as I have had the system for a few years (repasted maybe 1 year ago). I have no paste, and my time is limited. I will likely be upgrading in a year but will repaste if these temps are running too hot. Thanks in advance bros and broettes!

 

Hi Dave. For a 2760QM that is a tiny bit warm, but not terrible. If your heat sink radiators and fans are clean, I wouldn't lose any sleep on those CPU temps. Re-pasting might lower them a bit, but 80-85°C is not dangerous. If they don't go higher than that, I would not consider it a problem if you don't have a lot of spare time available to spend on re-pasting them.

 

Thanks Brother Fox. What are ideal gaming temps for this spec CPU?

 

I haven't used a 2760QM CPU for a long time and I don't know how to accurately answer the question about "ideal temps" because that is a relative scenario. I think any temp that is not "too hot" is technically OK. By "too hot" I mean that performance declines due to temperature throttling or a thermal shutdown occurs due to overheating. You will likely begin to see the 2760QM throttling somewhere between 85-90°C, definitely will above that, and your system will shut down at 100°C. [ ARK | Intel Core™ i7-2760QM Processor - 6M Cache, up to 3.50 GHz]

Temps are going to vary by the environment the laptop is being used in. Cooler is definitely always better, but I think no higher than 80-85°C with a 2760QM is fine. It should not be any hotter than that considering is does not have very high clock speeds. My impression is that if you begin to see more than 85°C in a cool environment (say 70°F ambient temps) that would suggest a problem that needs attention (i.e. dirty fan and heat sink or needing to re-paste). If you are using the laptop in a hot environment (more than 80°F ambient temps) 85°C under heavy load is probably something that could be reasonably expected under the circumstances. As long as the CPU does not begin throttling due heat and stays below the T-junction, I would say that probably can be considered "ideal" temps for gaming.

 

i think you're doing just fine, mine's 2670QM and it gots 83C on two hours gaming, the GPU got 78C after cleaning the vents.
before cleaning my cpu got 93C and my gpu is over 100C .
i just got thermal shutdown yesterday and decided to clean the dust on my vent and the temp is really reduce.

 

Good job, p4izanagi! Many people rush to re-paste their heat sinks instead of doing a simple cleaning and more often than not re-pasting is unnecessary. It is surprizing how little dust accumulation on the heat sink radiators that it takes to have a pronounced effect on CPU and GPU temps.

 

yeah Mr.fox, its all thanks to your post too .
i hope somebody try cleaning the vents first before re-paste their heatstinks.

 

Looks like I will need to get 3 pipe heat sink and repaste the CPU. My GPU's are fine and i did those myself but I didnt watch the tech do the cpu and under heavy load it's running 88C - 102C with the turbo hitting 4.4Ghz.

EDIT: UPDATE: I got it from dell cheap will be here next week.

Items Purchased

Item Total

$33.77

<form action="/support/orderstatus/us/en/19/OrderStatus/Order?page=0&perpage=0&s=dhs&SearchByValue=365589726&VerifyWithValue=127156449&SearchType=OrderNumber&VerifyType=CustomerNumber&InvalidOrder=False&SearchID=OrderNumber&VerifyID=CustomerNumber" id='grid' method="post">

Quantity	Item Number	Description	Unit Price
1	0MXJ5	Assembly,Heatsink,Central Processor Unit,QC XE,M18XR2	$20.99
1	11725	Warranty Longer of 90 Days or Remainder of System Warranty/ Svc Contract. 21 Day Return. Parts may be Reconditioned.	$0.00
1	375-6061	Thank you, Dell Outlet	$0.00

</form>

Item Sub Total	$20.99
Item Ship Charge	$9.99
Item Tax	$2.79
Item Total	$33.77

 

AlienMike, what are you running for flex? If you haven't already, drop it to 20 (or less if it is still stable with less) and that may help the temps. My system has been running at a constant 4.4GHz all weekend and has not gotten that hot playing games. I do have the 3-pipe, but dropping the flex should help your situation unless it is already set that low.

 

It is still at 25. I have noticed that I failed to properly ventilate the stand on my desk and today I switched everything around. I believe that adding the 3 pipe heat sink can only improve my temps and I can re-paste the CPU for my own piece of mind. It runs cooler now but I have found that a certain program I use is capable of running up the CPU/GPU loads far more than that of 3D mark tests, or anything else for that matter. I tested the temps for all benchmarking and some high end games and found just this one program able to stress it to over 100C. I throttled the program down to a max of 80% of processor and it stayed under 92C during this heavy load. I turned the BIOS setting back up to Performance 3 and my heavy program is no longer working for some unrelated reason. The system runs perfectly still and has yet to throttle for heat.

UPDATE: I went and set Flex to 20 after fixing my other issue with the program not loading. Now when I stress the system at 100% it is still hitting 102C on one core. Under any other normal conditions it is not getting above mid 80's.

 

The IC Diamond showed up today and I did the teardown and installation of the 3 pipe heat sink and the re-paste of the GPU's/CPU. I watched Mr. Fox's teardown video a few times and it was a fantastic help. The teardown and rebuild went perfectly and so far the stressed temps have gone from 102C to no higher than 86C after an hour of load. The machine is one happy beast I am gonna have beer to celebrate. :thumbsup:

Thanks again to this forum for all the great info.

 

Awesome work.

 

Yeah - nice job, Mike - gotta be happy with that!

 

I have recently had cause to watch all the temps on my system to look for anomalies in heat. I noticed, for some time before this, that one of my GPU is always 10C warmer than the other. They are not getting too hot but I went from stock paste to IC Diamond and Arctic Silver 5 before that and the temps are the same. I noticed the fan of GPU2 is running consistently 2000 rpm's slower than its partner.

Fans:
CPU min 2000RPM max 3600RPM avg 2766RPM
GPU1 min 3500RPM max 3800RPM avg 3845RPM
GPU2 min 1100RPM max 3700RPM avg 1806RPM

GPU Temps
GPU 0 min 42C max 60C avg 51.2C
GPU 1 min 47C max 66C avg 61.5C

I just had it apart but I did not remove the fans. I may start by taking them out, cleaning and re seating the connectors. The precision of the variance perplexes me, exactly 10C and 2000 RPM's.

Has anyone seen this and what are the normal fan speeds of the wild M18xR2? Mine is under GPU load of near 100% during this test.

 

Maybe some static is messing with your fan's functionality, try a power drain (Remove the battery and AC adapter, press and hold the power button for 30-45 seconds, reconnect and try it).

Other than that I don't see why the fan would start spinning slower after the tear down.

 

It was slow before the teardown, I will try the power drain now. Thanks Luis!

 

hi alienmike818, i have the same processor and heat sink just installed, what program/method did you use to track your temps, i'd like to do the same and compare results, i used the stock TIM that came on the triplepipe

 

I see, has it always been slow? It could be driver related.

Also, is your BIOS up to date?

 

Yes it has been slow for some time just not enough to get noticed before now. I am on A10 bios. I pulled the fans, cleaned and re seated the connectors. I drained the power for 2 minutes to be safe. Now it is running within 1000 RPM's and 8C warmer.

EDIT: After it warmed up its back to being 2000 RPM's slow and 10C warmer. Very Odd.

 

This is a fairly common observation. Guys like AlienMike that pay close attention often notice. I have a few thoughts that might be relevant.

• One GPU running a different temperature than other is normal.
  I have observed the same behavior on the XPS M1730, M17x R2, M18x R1/R2
• Chassis design has something to do with this. The left (primary) GPU is optimally positioned above the fan and runs cooler.
  The right GPU is slightly offset. The GTX 580M had a special support plate for cooling the vRAM for that reason.
  The right GPU is always a bit warmer than the left GPU on the M18x R1 and R2. It was that way for me with 6970M, 6990M, 580M and is with 680M.
• Load also comes into play with the fans. Based on the fan tables in the EC, their speeds are regulated by temps they report to the EC.
• Your temps are reasonable. If they are reasonable under load, then there is no cause for worry.
• Your observations show that both fans are working and able to achieve their full RPM when they need to.

That information, plus $2.00 might get you a cup of coffee at the local 7-11. Hope you find it useful.