P370em / np9370 overheating problems with 7970m crossfire.

Pretty much exactly what the title says.

I've noticed that my slave GPU tends to hit extremely unhealthy temps.

I decided to play Metro 2033 for a few hours while monitoring my temps/utilization with HWinfo.

to my surprise (and horror) i found that my slave GPU tends to shoot up in temperature very quickly, sitting way higher than my main gpu in temperature.

my main GPU can happily plug away at 70-75c for hours without anything happening. The slave GPU hits 80+ c in 5 minutes and 90+ in 15-20 D:

At around 95c, the fans kick up to maximum. This succeeds in bringing the slave GPU down to a healthy 80c and can keep it there, but unfortunately after 5-10 minutes with maximum speed, the fans will ratchet down again, causing temperatures to soar. after this happens, when the slave GPU hits 95C again, sometimes the fans don't even bother to spin up!

Is there any way to get the fans to stay at maximum during high usage? I don't think having my slave GPU hit the borderline unsafe zone for heat before the fans decide to do their job will help its longevity.

the depressing thing is, if the fans would just kick in and stay at maximum right from the get go, this issue wouldn't be occuring.

I have not overclocked anything, and am running catalyst 13.1 drivers.

 

It's a well-know problem of P370EM, Alienware M18X-R2 have a much better cooling with SLI/CF
I think my slave 680M is died because of that...temps was too much high...
i've buyed some copper ram heatsink (ram-sink), i'll install them all over the heatsink pipes when my 680M came back from RMA...let's hope heatsink + new thermal compound will lower temps a little bit...
Also maybe the pressure of the heatsink on slave video card is not enough, i've heard about a little mod to increase the pressure on the core/die of the video card (but i don't know if it's a good idea)

 

Out of curiosity, what happens when you use the -xcl command in afterburner to enable Overdrive?
Overdrive also has an option to force fan speeds. It works perfect on my 16f1 but it's completely unnecessary, since I already have custom fan profiles, but it may help out your machine.

 

Well... that settles it for me then. X570SE it is. >_>

I was contemplating on going for the P370EM but, if this is a well known problem, I will NOT spend $4k on a top end machine just to have it overheat on me due to poor cooling. The X570SE seems to clear that issue.

 

I cannot change the fan settings from the CCC even with the -xcl command unfortunately

 

Have you tried using an alternate program? Such as GPU-Z? A repaste might be in order (especially if you have something inferior to IC Diamond compound).

 

I will try repasting it as soon as I get my hands on some decent thermal paste, should be within the next few days.

At this point a bad factory paste job is the only other thing I can think of (It was IC diamond lol).

 

It happens, unfortunately. There is some debate on what the best application method is, but it's generally agreed that the center-placed, pea-sized drop provides the best spread once you apply the pad/heatsink.

 

am having the same issue like u ... i played far cry 3 for 20 min after i closed the game i found that my GPU has actually hit 102 C !!!
when i play for example world of warcraft i have reasonable temp around 70-80 but then am playing with xfire 7970m with 35fps :S
in 10 man raids 90, this just sucks , my 2 years old sager 7280 with 460m sli gives me way more stability and higher fps in wow ..
and it never jumps above 75c no matter watever i load it with...
sager 9370 in my own opinion isnt exactly a good laptop overall..just received it yesterday and i already feel more attracted to my old year x7200.
build quality is not superior.
so what to do now ! wait our cards to fry up ?
if u got a solution please do share with us.

 

I know you can use HWINFO64 to adjust fan tables for Alienwares, but I'm not sure about Clevos. Do you see that option in HWINFO?

Otherwise I would definitely repaste as others have suggested. If that doesn't fix it, I would return the laptop (assuming it's within 30 days of purchase) and buy an M18x instead. Cooling system is epic in Alienware laptops, hence the M18x's records for laptop benchmarks.

 

Okay, having returned my np9370 to the reseller for a checkup, they have confirmed that it is NOT a hardware or pasting issue. In house testing has also resulting in the 2nd GPU hitting 95+ celcius, even after multiple repastings + a different gpu.

The problem lies with the 2nd fan not going full bore until the 2nd GPU hits 90+ celcius. They state that it is a bios issue.

They have told me that they have contacted clevo and are requesting a modded bios that enables the Fn+F1 full fan toggle to see if that will fix the issue.

Honestly though, they also resell M18x's (Not available via Dell Malaysia for some reason). Tempted to just RMA and see if they swap me one

 

Hopefully a modded bios for the p370em. I'll keep my eye on this thread. Let us know if you get any details. I would love to have my fans in max.

 

my system fans throttle to full speeds when i hit FN+1 , but still i get the high temp .. so this solution doesnt really help ...
i know it sounds logical ( high speeds fans lower heat ) but that doestn seem to apply to this precise machine.
the only way i believe we can get lower heat is by having an app or something to reduce the usage of 2nd gpu to around
70-80% , could result in a bit of fps drop but might solve the issue.
currenty my cpu and gpu1 dnt jump above 80 watever i load , gpu 2 min is 90% in demanding games,regardless with FN+1 or without.
GPU 2 usage is always higher thatn GPU1 , if we can reverse that perhaps we can get to somewhere.

 

You just need to re-apply thermal paste (on both GPUs) and clean the GPU fans. No need to artificially cripple your machine.

 

Over in the Hardwareluxx forum in Germany there is a guy who had the same problem as many with the second GPU running hot ( Luna p370 em - Seite 8).

The solution was finally a new heatsink - but only the second one they sent him was finally working right.
It seems that the second GPU heatsinks are not the "best work". Maybe there are some faulty ones out there.
The problems can be that the heatsink doesn´t cover fully the GPU die, or that the pipes are crippled...
Just wanted to let you know since there seem to be many users with crossfire 7970m and the P370em that have that problem...

 

sirana i just recieved the machine couple of days ago ... no use of cleaning , lots of other have done a repaste which aint the problem ...
i wish i could just move everythin to my x7200 lol
then problem solved.

 

In that case it should really be the Heatsink. As it's so new I would send it back to your reseller and they should fiddle with the Heatsink.

 

Ok my newly arrived 9, 370 ran Kombustor for 15 min, full screen 8Xmsaa GPU0 went straight to 99% load and dropped halfway thru test to 96% and stayed there according to Kombuster. Gpu 1 stuck at 97% for first minute of test then got up to 99% and remained there til end of test. Temps GPU0 started at 39C ended test at 82C GPU1 Started 38C ended at 97C. Ok something else checking Afterburner which I also had monitoring it shows GPU1 usage maxxed out at 99% BUT gpu2 usage says it maxed out at 154%....... UHM !!!! That is not memory usage I checked that is GPU Usage? I know Kombustor is extreme usage and will have some game runs tomorrow. Just observation fans seemed to only really kick in as GPU 0 started to get hot they didnt seem to be affected by GPU1 temp at all. Once Fans kicked in things leveled out at 81 on GPU0 and 96 on GPU1.

 

While I know Kombuster is supposed to push the cards hard, 96 C on a stress test is still rediculously high when I see M18x users topping out at the high 80's with a 7970m xfire setup :[

 

Did the heat sink from the p375sm solve the overheating issue?