M18x - GPU Temperatures

How come, bar-code, you were able to run Furmark with no throttle for the GPU's?

 

Hey Hey Redline!

I never said i didnt suffer throttle.

Just saying my max temps were 72 on both gpus in furmark 20mins after a repaste job compared to stock paste before the repaste of 78/80.

Vantage overclocked at 730ghz also reveals the temps to be at 72/73max no throttle which im happy with temp wise.

Yes i did suffer the throttle so my graph in furmark shows pyramids up and down, instead of a flat temp graph.

The funny thing with this whole throttle is, it seems to be power related not temp related. Same as when you run the m18x only in battery mode Vantage is throttle all the way from the start.

 

Question: Overall temps on HWinfo says 77 and 82. However both GPU's are at 92.5 on the memIO diodes. My GPU's are the 6970s with an 800/1000 OC. I am thinking that is fine as long as I am away from TJmax temps and away from 99-100. Is this accurate for these AMD GPUs????

 

*yes hwinfo is correct those temps you read were accurate. Not safe for 24/7 use at those temps repasting the vram modules is required for the best gameplay and or overclocking.
mine were the same then after repasting their cores and replacing the thermal pads too all with ICD24 the vram no longer reached 90+ more like 70+ its the pads that inadequately cool the vram and surrounding vrm regulator etc. Temps were like night and day... Plus ICD24 isn't conductive meaning it won't short out your vram if it ever baked out... Which is not easy for ICD24 to do since its very thick and very thermally conductive due to the synthetic diamond powder.

 

I wish fornax would be closer (los angeles) so I can have akimox do thw repasting on my machine :-(. I don't dare to open it myself. Don't wanna break anything.

 

I thought it was always known that the throttle was due to power draw and not temps......I pioneered this with my M17x r1 and did not realize people had thought otherwise on different models....

 

I think a little clarification might be needed.

Are you referring to removing only the memory thermal pads, or including the other pads on the VRMs and FET as well?

Because removing only the Memory pads creates a large enough gap that cannot be pressed down or bent. Even If I bent it it would be touching only the outermost edges of the RAM. The Gap is infact a little more than 1mm. removing the pads from the VRMs wont solve the problem either and is highly inadvisable to do so.

So I had to put the pea sized IC24s on the RAM and then place the pads to of them and then the heatsink.

For whatever reasons I have the same or better temps than you were reporting at my overclocked 800/1000 clocks.

Thanks in advance!

 

All I'm saying is you can remove the pads and use fair amount of paste instead.

As for the lower temps, there could be plenty of reasons (room temp, elevation, cooler pad, open window, etc )

 

Thank you for your reply Aikimox! +1 rep

So there is indeed a gap that will result, but you are filling it with a lot of paste. Almost like a sticky thick column that wont bleed and gets the job done of thermal dissipation.

I guess because of the sticky thick nature of the compound it works fine.

Would be great though if we could make them touch the RAMs a bit better. But gains might be minimal because its not a copper contact zone for the RAMs

 

This may be irrelevant, but does anyone know what the max rpm are, for the gpu and cpu fans? I'd just like to know the max rpm that I could set them to using HWiNFO64, without incurring damage

 

2900 I believe. set it to 3000 that's what Johnkss did.

 

I thought it was 4200 actually

Thats what my custom settings peek at ...

 

Pretty sure it's 2900... but I could be wrong (although It's never happened before lol)

Edit: Actually I think it's 3400..... can anyone confirm?

 

Wow. Actually my GPU1 peaks at 3600 on system auto so I they should be capable of 4000+?

Thanks for the responses guys

 

4100 rpm is the limit, tested and confirmed.

 

wait just to clarify these numbers I'm giving are for CPU not GPU.

 

What temps are some of you seeing for a certain RPM?

I ask because my GPU (560M w/stock driver, A02 BIOS) goes to 70-72C and the GPU fan goes on FULL BLAST even when playing older games like GRID.

It's a comfortable 46C at idle, goes to 50-60C when browsing the internet or playing 2D games, but just sends the fan into airplane noise territory at 70. Just a normal thing for hot GDDR5 or can I hope for a slower fan with an A03 BIOS or Verde drivers?

Also does the left edge of anyone else's M18x get warm/hot along with the GPU temps? The left side from the ethernet port to the audio jacks gets pretty toasty at 60-70C. The top part above the GPU0 gets pretty warm too. Do those of you with dual GPUs get warm edges on both sides?

 

On full blast the GPU fans run at 4100rpm and the temps for me atleast never exceed 73C on all 3 senors for main card, on the slave card its close to 70C on all 3 sensors (full load Crysis 2, Metro 2033)

My left edge is just fine. Whatever warmth I feel is on the dash board side on the top and its never toasty, just mildly warm. Maybe 33C-34C

You might want to consider taking a closer look at whats going on with your heatsinks, if they are touching certain surfaces that they shouldn't be.

 

Left edge near the audio jacks could be heat from your harddrives

 

Nah, the harddrives are both Seagate 750GB that run pretty cool and quiet. When they do get a little warm I just feel it on the palmrest, directly above the drives.

I know for sure the left edge heat is GPU related since it only gets warm when the GPU temps go up. The top left gets warm too, I just was wondering if it might have been an intentional thing, using the aluminum part of the chassis as a passive heatsink in addition to the fan cooling.

 

not that this is the case with you but its common for a platter based hd to get to upto 105 F in some scenarios

 

When I first got into upgrading laptops, I put 2 WD3200BEKT drives into my HP DV9260nr and they would both get upto 65c in there (which is QUITE hot)

It was very obvious if you felt around that area on the laptop

 

My M18x in the HD area of palm surface can get a little warm... two of dells 750gb 7k hd's can do that ... I havn't installed the 2 240gb vertex 3 max iops yet... been testing and benching and I don't want to degrade them for that, might bench just a couple tests once they are in but I won't till I am done testing the 2920. Many countless older laptops I have had and used or worked on... that had standard HD's... had bad heat issues if they were the 7k rpm type... the ones that spin up slower seem to produce less heat like a Dell studio here with 5400rpm and a Fujitsu Lifebook with a 4500rpm. they really can get quite warm and haven't had a single HD go from heat... but it could happen I guess. Sorry for the OTP

 

I have a WD 750 (think Dell uses Seagates) as well as my Intel SSD. They both feel like they get hot but I know that there is a lot more heat coming off the mechanical drive. The newer stuff is a lot more efficient than the older 7200rpm drives.

 

I know some 7200RPM hard drives run hot, but the 750GB Seagate Momentus Dell uses (I added a second one aftermarket) runs very, very cool. The area directly above on the palmrest has been slightly warm a few times, but that's it. They are almost whisper quiet, too.

I know it's the GPU that causes the left side of the chassis to heat up, since it gets really warm when the GPU goes to 60-72*C when browsing or gaming, and goes back to being cool when the GPU goes back to 46-52*C at idle or low activity.

Just wondering if this was a normal occurence for the CPU0 location in the M18x chassis before I start thinking about taking stuff apart or repasting just yet. At least the CPU runs cool.

 

Yes both sides of my crossfire get warm... usually my main gpu the left like you mentioned.

 

Im a designer and I use Cinema 4D regularly which can streen the CPU 100% when rendering. My temps get to about 78-85 max 89 and Im overclocked to 3.4GHz with a 2760qm and 6990CX.

Rendering can take hours at a time. So are these tempreatures safe? Im guessing no right?

 

m18x crossfire 6990m

 

have idle of 48
normal gaming at 54
intensive gaming at 58

im glad i chose the right one

 

got some more new things here i did few tests on samsung 256 gb ssd without sandforce controller as well as ocz 120 gb solid3 with 2280 sf. neither has stutters/bsod so imo its probably not ssd related issue that laptop freezes.

changing brightness still freezes for me once a while and i came to the fact that im accepting it as part of the pc, however this time around without the kernel driver installed, hwinfo64 still hangs if i play too much around it (only changing values of fan speed as well as switching between systemauto, set manual, custom auto)

i have had a few shut down hang with or without hwinfo64 freezes during window usage, however i no longer have any problem with kernel & system cpu usage.

another note here is that ive updated to bios A05 so im sure its got to be related somehow. with that bios update, it seems not all fans working at the same speed, most left fan on my keyboard seems to be for the 2nd graphic card and its just running its own speed, while cpu and maingraphic fans run at the custom auto speed.

i do know the fact that main gpu used to be the one on the left, however ever since i got a replacement machine cause my previous one was obviously bad, looks like some settings been changed =/

 

I recently bought 2 6990 OEM cards and instsalled into my M18X. I had a single 6970 before and temps were very good. I updated to the A05 bios for the 6990s and repasted with AS3 as that is what I had on hand. The temps with the 6990 were pretty high and I was kinda worried I was going to have troubles., I would see over 90C on the GPU. I continued monitoring and after about 1 week the highest temp I have seen is 89C in one spot on the primary card. The other temps are way down to respectable levels around mid 70s So I conclude the AS3 took a while to settle in, not sure why but if anyone else runs into high temps it might be wise to run it for a while to see if the heatsinks mate better after a while of use.

 

Unlike some compounds, AS3 does have a curing phase to achieve optimal thermal transference. Back in the day, it was worth the trade-off to use AS3 because it was the best compound around by far. Now days, I keep MX-2 for everyday testing/systems with no thermal issues/cool running (because somehow over the last few years I have enough to probably apply to over 1000 systems at this point lol), and MX-4/IC7 for hot and heavy system long term use.

 

guys sorry to hijack the thread, since i don't own an m18x.. but still i have dual 6970's...

Maybe i'm being "over protective" with my AW .

I have both cards undervolted@ 0.96V and i have paid a lot of attention in repasting, my temps are good, ranging from 68 to 73 degrees in games (depending on how demanding they are), the hottest sensor is the shader one wich stays variably 3 to 5 degrees hotter than the others.

I was thinking in upping the voltage to keep both cards @800Mhz core, 1V should be enough.. should i be worried about temperature increasing?

Is, for example, anything around 80 degrees under load suitable for everyday usage?

 

CptXabaras - If my memory serves me correctly, isn't the spec like 1.1v in the performance power state? (I think stock specs are 680/900/1.1v on the 6970M.) That being the case, you shouldn't have any issues with heat at 1.1v. If you stay at or below 80°C under 100% GPU load you should be fine. If you OC your cards while under-volted, I think your temps may actually increase because more current draw could result from decreased voltage.

 

Yes you recall it correctly, the 6970's @stock are 680/900/1.1V. i have mine flashed @ 715/900/0.96V

I will do some test tonight and see how temperatures are at stock voltage.

Thank you as usual Mr.Fox

 

Temperature isn't the only factor in determining what's safe.

Placing an equal load at a lower voltage will raise the current required by the GPU and is more likely to cause damage, even though the temperature could be slightly lower.

As Mr Fox has said however, in many cases this scenario will actually cause electronics to run hotter.

 

And, I learned that from you, bro. Thank you.

After you said it, I remember actually seeing that happen and not understanding why. My GTX 580M cards run ±5°C warmer while overclocked at Dell's factory under-volted 0.87v than they do at the same clock speeds running at NVIDIA reference 0.92v.

 

Preamble: GPU's have been running under 80 deg C since pretty much the day I got the M18x. Running A05 since pretty much the first day it was released here. Stock clocks/mem speeds. Fans are cleaned pretty much every time I move it which is a lot.

Noticed now that the temps are starting to creep up higher and performance is starting to degrade as well. Not sure what is going on and I'm starting to wonder if something is wrong. Seeing 82 deg C on GPU 1, and 83 deg C on GPU 2. Still below the threshold but unnerving either way. I'm not really able to break the machine down as I'm working massive amounts of overtime. At what point is a call to dell/alienware in order?

Edit: Fans are still working properly, these are sustained temps as well not just peaks.

 

use hwinfo64, i dont really trust dell/alienware drivers or bios. im still using bios A03 lol. i know for a fact that when i reinstall fresh windows, theres no drivers, none at all and i donno if its because my alienware cant detect intel HD 3000 (not showing in the bios) vga default graphic runs hot even during installation. first thing i installed on my alienware is ALWAYS winrar and hwinfo64 and force the fan at atleast value 2200-2600