MSI laptop cooling efficiency... Stock temps? OC temps? Cooling mods?

[alex-msi]

hi all

my msi gx60-3be (a10 5750m and hd8970m) without CPU load 42 degrees Celsius, under load 100% 78 degrees Celsius. GPU without load of 45 degrees, under 100% load temperature reaches 102 degrees . Replacing the thermal paste is not reduced the temperature of the GPU. In the review about gx70 wrote that the GPU they warmed up to 93 degrees

Review MSI GX70H-A108972811B Notebook - NotebookCheck.net Reviews

[Zbig0]

Yep, it's a close call between a 7970M (even overclocked), and an overclocked 675MX (as long as you flash that modified VBIOS). You'd be happy with either, the R4 is a good laptop though, a pretty good choice.

Thanks mate hope it will be runing smoothly similar as previous 17r4 Hmm 7970m with OC hits 8.5k in 3dmark11 675mx with OC can hit that much points?(Im not a 3dMark funboy but its some sort bench after all showing performance capabilities)... I would be happy with gt70 with 675mx as well with my use but occasion like I got just had to get finalized

hi all

my msi gx60-3be (a10 5750m and hd8970m) without CPU load 42 degrees Celsius, under load 100% 78 degrees Celsius. GPU without load of 45 degrees, under 100% load temperature reaches 102 degrees . Replacing the thermal paste is not reduced the temperature of the GPU. In the review about gx70 wrote that the GPU they warmed up to 93 degrees

Review MSI GX70H-A108972811B Notebook - NotebookCheck.net Reviews

what have you used to put the load? if Prime95 and furmark? if yes then check your temps while gaming as those synthetic test put un-realistic load that wont show up during gaming

If you checked those temps while gaming only then I would advice to read some more about re pasting techniques(And check if thermal pads are ok as well) if nothing will help this could be reason for warranty as if GPU gets 100C then throttling is happening(and its very unhealthy for GPU so its worth hustle if you are planning of using laptop loner)

[alex-msi]

I used a test msi kombustor 3.0 x64, games far cry 3, Metro Last Light is also very high up to 100C. Repaste arctic cool mx-2, did strip of foil - the temperature is still very high. Helped only removing the back cover, temperature dropped down to 87с. Now tried undervolt 0.975/1.025 at frequencies 900/950, instead of the former 1.050/1.100 - temperature tests fell to 90c and games to 86-88c

[Robbo]

Thanks mate hope it will be runing smoothly similar as previous 17r4 Hmm 7970m with OC hits 8.5k in 3dmark11 675mx with OC can hit that much points?(Im not a 3dMark funboy but its some sort bench after all showing performance capabilities)... I would be happy with gt70 with 675mx as well with my use but occasion like I got just had to get finalized

Ah, fair enough, that is a higher score than an overclocked 675MX, so it looks like the 7970M is a better choice from that point of view. A lot of it comes down to the luck of the draw on how your GPU overclocks. I've heard of some people who could barely overclock their 7970M's at all, but looks like if you get a good one then it's a better choice than the 675MX. Yeah, you can be happy you chose the 7970M!

[squareroot]

I used a test msi kombustor 3.0 x64, games far cry 3, Metro Last Light is also very high up to 100C. Repaste arctic cool mx-2, did strip of foil - the temperature is still very high. Helped only removing the back cover, temperature dropped down to 87с. Now tried undervolt 0.975/1.025 at frequencies 900/950, instead of the former 1.050/1.100 - temperature tests fell to 90c and games to 86-88c

Have you checked the sensors on gpuz or hwinfo? There are three temps for the gpu. I use those to determine which part of the gpu is not properly cooled(bad paste/thermal pad/contacts). The difference between the temperatures shouldn't also be far from each other.

Thanks mate hope it will be runing smoothly similar as previous 17r4 Hmm 7970m with OC hits 8.5k in 3dmark11 675mx with OC can hit that much points?(Im not a 3dMark funboy but its some sort bench after all showing performance capabilities)... I would be happy with gt70 with 675mx as well with my use but occasion like I got just had to get finalized

At what clocks did you reach on your 7970M? While some have hard time overclocking it, some also were able to go past its desktop counterpart. (7870 Core/Mem@1000MHz/1200MHz) 7970M & 8970M are underclocked desktop 7870 if you didn't notice.

[apprenticegamer]

I have the GT70 Dragon, i7 3630 with 680m, and I love it. I have only had a short time, but having temps creeping up there on my gpu, 84C playing farcry3 and dipping into the 90's playing crysis2 and cpu usually mid-high 70c, idle high 30"s. Was looking into repasting cpu and gpu. My non engineering mind has went to supporting the grain of rice sizing theory and letting the heatsink pressure move the compound around. Then there are also suggestions of tinning the heatsink surfaces, (uber fine layer to fill micro defects in the surface) seems logical in my mind, but then my mind thinks that may be both ideas together would be against each other as tinning covers whole area and the grain of rice method is more centralized on the heatsink. Would the advantage of the tinning be negated by the tinning on the exposed outer corners ? I would like to venture into overclocking, once I have my heats stabilized and to that point, can anybody suggest a guide to understanding oc of this 680m. Knowledge is good, I am just a little lite on it. All thoughts appreciated.

Ken

[Robbo]

I have the GT70 Dragon, i7 3630 with 680m, and I love it. I have only had a short time, but having temps creeping up there on my gpu, 84C playing farcry3 and dipping into the 90's playing crysis2 and cpu usually mid-high 70c, idle high 30"s. Was looking into repasting cpu and gpu. My non engineering mind has went to supporting the grain of rice sizing theory and letting the heatsink pressure move the compound around. Then there are also suggestions of tinning the heatsink surfaces, (uber fine layer to fill micro defects in the surface) seems logical in my mind, but then my mind thinks that may be both ideas together would be against each other as tinning covers whole area and the grain of rice method is more centralized on the heatsink. Would the advantage of the tinning be negated by the tinning on the exposed outer corners ? I would like to venture into overclocking, once I have my heats stabilized and to that point, can anybody suggest a guide to understanding oc of this 680m. Knowledge is good, I am just a little lite on it. All thoughts appreciated.

Ken

Hi Ken, I think tinting the heatsink is a good idea to fill those microscopic valleys. I used Arctic Silver 5 as my paste: old-fashioned but still a reasonable paste. That paste is quite thick, I used the spread method & tinted my heatsink. I think tinting heatsink & grain of rice/dollop method is still compatible to be used together. If you have a thick paste like AS5, then I recommend the spread method, if you have a thinner paste then the rice/dollop method might work fine too. Having said that, AS5 on my CPU I've tried both the line method (rectangular CPU) and the spread method with identical results; on my GPU I've only ever used the spread method.

[apprenticegamer]

Thanks Robbo for getting back to me, I appreciate your thoughts. But what kinda temps should I be getting ? Or is it a repaste and wait a few days and remove inspect and learn from what u see, then repaste. I like your idea of spread on the gpu , it is a rather large area. Again thanks for your brain.

Ken

[Robbo]

Thanks Robbo for getting back to me, I appreciate your thoughts. But what kinda temps should I be getting ? Or is it a repaste and wait a few days and remove inspect and learn from what u see, then repaste. I like your idea of spread on the gpu , it is a rather large area. Again thanks for your brain.

Ken

Hi Ken, I guess my brain has some uses sometimes! Here's the link that describes how to do the spread method:

http://www.arcticsilver.com/pdf/appmeth/int/ss/intel_app_method_surface_spread_v1.1.pdf

I can't tell you what temperatures you should be getting for your laptop - the cooling systems & cooling capacities are all different. You best bet is to Google your laptop to see what other users of your laptop are getting temperature wise. Generally, I think anything under 90degC is OK, under 80 degC is good, and under 70 degC is excellent.

[apprenticegamer]

Robbo, thanks again. Is there a breakin time for mx4 compound, I can't seem to locate any info . I will be doing this this afternoon, but I also would like to flash my bios in the near future. My concern is to have only one variable happening at a time.

Thanks again.

Ken

[Robbo]

Robbo, thanks again. Is there a breakin time for mx4 compound, I can't seem to locate any info . I will be doing this this afternoon, but I also would like to flash my bios in the near future. My concern is to have only one variable happening at a time.

Thanks again.

Ken

MX4 I've never used, in fact I've only ever used Arctic Silver 5, but that's because I've only ever re-pasted two laptops, and it's worked fine for me. A quick google revealed that there's no break in period for MX4. I would think in reality that temperatures would stabilise though over a couple of hours or after you'd put a bit of heat into it so it melts some more - maybe after spending just a 10 mins or so at 70 odd degC it might be stabilised. That's just me thinking about it, not based on experience or research.