Calling all GT70 OD owners

There have been some debate regarding the cooling capacity on the newest MSI GT70. Anandtech tested theirs and they found out that the CPU reached 98C during gaming. Notebookcheck on the other hand got pretty good results.
There are some numbers floating around, where owners say they get this and that, but no concrete evidence to really show the temperatures.

Lets figure it out

To do this:

1. Download HWMonitor Pro (14 day free trial). It must be the Pro version.
2. When you start HWmonitor, you can see all your hardware inside the notebook listed, and the temperature on them. HWmonitor PRO have some way to trick people so you can`t see the the temperature on all of the components. Some have "TRIAL" listed instead of the temperature. To remove this, just restart the program until "TRIAL" is gone. The results that is important here is "CPU Package" and "GTX 780M TMPIN0". Make sure none of them have "TRIAL" but show the temperature.
3. Go to "Tools" > "Start recording". This will allow HWMonitor to log the temperature during the test.
4. Fire up the games you want to play. Try to atleast game for 90 minutes while it logs.
5. Quit the game, Go to "Tools" > "Stop recording". HWMonitor will now create graphs and will put them in the HWMonitor File folder.
6. Upload the graphs to imgur or this website. Post your results here.

When posting your results, state:

- Factory or your own paste
- Turbo fan ON or OFF
- Ambient temperature
- What you ran when logging the temperature. What game. What settings. Benchmarks. etc.


Thank you

-----------------------------------------------------------

Here is how the graphs will look like.
My GTX 680M from 3 hours of gaming


3610QM 3 hours of gaming

 

What!?

You have been lecturing me about how low the temps on the GT70 2OD should be and you have the Ivybridge version?

I will contribute to this thread. But dont have time for 3 hours of gaming and am going offshore tomorrow. Instead, I shall run Heaven Benchmark for 30-60 minutes, which is actually less stressing than The Witcher 2, cos Heaven doesn't stress CPU, thus allowing the combined heatsink to remain cooler.

This will be my replacement GT70 btw (not that the temps are much better...no artifacts though....yet)

EDIT: Another thing that should be specified is what game was being played and what settings the game was on. Obviously, I can turn FIFA 13 on for 5 hours and never go above 54C. Also now running Heaven for first time on my replacement GT70 2OD.....seems that temps are about 10C lower at this stage than with the one I RMA'd. However, If it was the Witcher 2 I was playing, they would be by now hitting 86-87C instead of the 78 C it is sitting at.....this may suggest that much of the heat on the GPU, is actually coming from the CPU via the heatsink bridge.

 

Any data will help.

I reacted because 98C is too much no matter what notebook you are talking about. Its just not normal, which hopefully this thread will show.

 

Didnt work.

What I got was a big load of data stating 'TRIAL'....

I am not going to run all that again, but according to HWMonitor

MAX GPU 91C
MAX CPU 77C

That was with Heaven, having let it run for about 2 hours.

What I got was this think it must be for the CPU.....but I dont think it is quite right...I dont see why the CPU would heat up to 78C when Heaven is left running, and then cool down to 64C and stay there...whilst Heaven is still running......unless something else was getting throttled.....(dont know, wasnt watching it, was out doing some gardening).

 

Ah I forgot to mention that.
That "TRIAL" is a way to trick people to think you can`t use the software fully. If it shows "TRIAL" under "CPU Package", just close the program and restart it. Then the "TRIAL" may have gone away or jumped to something else.

Not sure what your graph is showing, "THRM Temperature". Please if you have the time, could you rerun it, to show the CPU and GPU temps? Atleast for 30 minutes or something. Thanks

EDIT: Just found out what it is. It is the temperature on the chipset. 60C in average is very good. But CPU package and GPU temp is much more important

 

Nobody else that want to prove or find out that the new GT70 is without cooling issues here....?

 

Have now done a repaste job and am finding that the temperatures are lower on the GPU, and much lower on the CPU.

I used the Noctua NT-H1 thermal paste..not Icey T Diamond or whatever it is called, but still decent.

Idle temps are 7 C lower on the GPU (idling after having run a benchmark). During the Heaven Benchmark, the GPU hit 78 C, which is around 7-8 C less than it would previously reach when running Heaven one time from start up. I also noted that the fans were quieter and seemed to be taking longer to kick than with previous runs of the benchmark. It seems that the fans kick in at a certain speed, at a certain temperature. I shall look into a finding a way of modding the preset fan speeds to kick in a little faster, at slightly lower temps. Whilst the Turbofan is way too much, it would be nice to have the option to sacrifice a little silence in exchange for lower temps and therefore better performance.

The really big difference was with with the CPU, where temps were around 15C cooler after running this benchmark one time, from start up.

I am personally much happier with these temps than with what I had before. From having done a repaste, I can vouch that the thermal grease, most notably on the CPU, but also on the GPU, was thick and all over the place. A bit too much, and the improved temperatures prove it.

MAX GPU 78 C°
MAX CPU 58 C°

 

So it really all is about the paste, it seems, huh? Interesting. I've been following these issues closely since I'm looking to buy a GT or GX 70 in the near future. A lot of good info. Personally, I feel some of the posters in the MSI forum tend to be a little too fanboyish about the MSI brand - THO they still are willing to help others. Matt seemed to be more skeptical and it sounds like his problems are getting resolved. That's encouraging.

 

Hey, if you went into the MSI forum and said that the member were a little too fanboyish, you would be risking a ban. The members there a not fnaboyish you see, it is just that MSI make so damn good products that there unwavering enthuse and support comes naturally.

Here is a further update after a more vigorous bench test (3d Mark)

Here is with factory default paste:
View attachment 97954

And After Paste:


So about 7C lower on the CPU and MSI Afterburner indicates a max GPU temp of 86C compared with the CPUID reported 90C on the first run through. However, I remember very well that MSI Afterburner reported 91C when I done the first run through, which is as high as the GPU will go before it throttles itself. So the fact that the max temp ceiling was not reached on this benchmark after repaste, is a better result than the figures themselves show.

Interesting that 3D Mark seems to like Windows 7 much better than Windows 8. I should also point out that the first screen came from the GT70 I RMA'd. You will just have to take my word for it that the second one in its pre repaste condition was the exact same as first as far as temps were concerned.

So in the final analysis, repasting provided for a very worthwhile improvement, but by no means a staggering improvement.

 

Ive ran battlefield 3 for 60+ min and the temp never went above 70C, granted I have a coolermaster U3, but I dont think that affects the temps too much.

 

Thank you guys.

Please try to use the log function within HWMonitor. A graph that shows the upswing and lows is very helpful showing the complete picture.

You can start the log function under "Tools" I think. You start it before gaming/benchmarking, then stop it under "Tools" again after you are finished. Then HWMonitor will create graphs from CPU and GPU and save them under the HWMonitor file folder

 

If people could also post the temperature of the room while they ran the test that would help eliminate one variable, and could be used as a (very) rough guide for people to estimate how effective the cooling would be in their environment... I don't have one btw...

 

Guys, good thread! for comparison we could so with also monitoring the frequency of the GPU and Core, as a laptop with better cooling will turbo higher or for longer periods because it can,

so while 2 system may run the same temperature, one might be running 5% higher clocks, or turbo for 10 seconds longer per turbo because it is able to better maintain temperature therefore 2 laptops, the same spec, with the same components don't equal the same performance, one will likely perform better thanks to GPU Boost 2.0 and Turbo Boost 2.0 in the Haswell's

These "Turbo" technologys and good old fashion throttling are blurring the lines, and a component will adapt it's speed to its environment!

also I would suggest,

no cooling boards just sat on a flat desk - so we know its equal
Add room temperature beggining, middle & end of test as this may account for some raise in temps

 

I will do that sometime today when I reach home cloudfire, with my GT60. Though I changed the thermal paste with ICD and the thermal pads. I will play for 2 hours continuously and post the log here (with turbo off).

 

Can we also please monitor the CPU frequency and GPU frequency and post basic system specs and TIM used, if you guys can send me the data I will collate and graph everything together, maybe get some of the guys from the other brand sections to throw some GTX 780m in to the mix and we can have a direct comparison! and compare average GPU and CPU frequency, temperatures and we can see which models can turbo the highest for the longest and if that is a direct result of temperature and cooling ability

 

It looks like for some reason did not pic up the GPU, I will have to do it all over again. I have the one for the CPU attached. This is continuous play GRiD 2 with ultra settings (Didn't mention the room temperature - 22 - 23 degrees Celsius).

And it was for about 1 1/2 hours because I got bored



And here is the GPU also for about 1 and half hour. Its shocking to see in the diagram how stable is the temperature and how quite this system runs... By the way, turbo fan is off (obviously otherwise how the system will be silent )

(If you see in between 37-44 min there is a bit of temperature drop; that is because I took a short break )

 

And here is the 3D Mark 11 score and temps (everything stock) NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-4700MQ,Micro-Star International Co., Ltd. MS-16F4

View attachment 98070

 

What drivers are you using?

 

Temps are now at 76-77C while playing bf3. I havent done anything major to this gt70 since i got it besides do a sfc /scannow in the commandprompt which revealed errors galore in windows, also did a windows update cleanup with it, and got around 8 gigs of freed up space. There are also a couple gigs of log files for the system scan, but I think the two are unrelated. I think I might need to repaste in the future. I've heard great things about IC diamond, and it will probably lower my temps by around 5-6 C.

 

Please see here post # 742 for the driver http://forum.notebookreview.com/msi/719029-official-msi-gt60-2od-owners-lounge-75.html

By the way, if you guys want, I can try tomorrow and run the same test / game with turbo on (but if you think about it, reduce the temp by 7-8 degrees and you will notice where the room for a little overclocking will be available; ofcourse I do not mind posting to prove what I am talking about). Now it's past 2 am and I'm going to sleep

 

Thanks juliant for providing excellent information. I repped you for the effort

We have one GPU temperature measurement now on the GT60. 87C in Average on the GTX 780M is pretty good considering the power in the GTX 780M. It looks like it runs 10C hotter than my GTX 680M, but I have a GT70 which is bigger, so that will help a little with the temperature.
Your GTX 780M max out at 90C with both gaming and 3DMark11 it seems.

I do have to ask you to run GRID again, and log the correct CPU temperature. You logged the chipset (THRM). You need to log the "CPU Package" which is all cores combined.

Thanks for helping out


@elmyo: Please download HWMonitor Pro and contribute you too. Graphs are so much better than saying a GPU run this and that. Really helps future buyers who got scared of the Anandtech review

 

Cloud,

Going back to your OP and to better achieve your goal here, I think there needs to be more established standards of testing. Without controls, the variables will make your objective somewhat a moot point. For instance, yes, a lot of folks are throwing around temps, but no one is telling their ambient room temps at the time of testing. It would be helpful if everyone where testing with the same methodology..... same test, notebook on a flat surface, and insuring all vents are unobstructed. Also, was testing done with factory paste, or has the notebook been re-pasted?

Just a thought......

 

Thanks for the advice Striker. I just edited the first post to make it a little more clearer.

 

I already mentioned that the system was repasted with ICD paste and the thermal pads have been replaced. FYI I have also tried to add extra copper in the thermal sink and copper pipes but I did not see any improvement in cooling. I have also said that the room temperature was / is 22-23 degrees Celsius. What I did not mention is that I had it on the dining table which is a wooden table (if this makes it easier, can be added to the questions); it may have been a bit better if it was a glass table...

 

Can you please also include frequency into that test!

I'm getting grid 2 tomorrow and currently have ambient around 22-23 so I will try and replicate this +rep if you can get those clock frequencies mate!
if you can give me the HWinfo CSV from the whole bench I'll work out the rest! (version 1930 will monitor the 780m, the later version 1945 & 1950 betas didnt, I dont know if 1960 fixes this as I haven't had time to assit Matin in working out what that bug was - maybe he sussed it out)

** I've got some dodgy RAM that's been RMA'd till then I'm on a really old single stock of 1066mhz CL10 4gb - I don't know if this will affect performance but I assume not?

pushing the boat out even further, I would like to get together with some of you guys if your local to me & also have a 4700mq & a GTX 780m so we can all do some select benchmarks FF, WoT, BF3 (the aero scene) and maybe some other benchmarks we can freely obtain (WorldInConflict demo) and 3Dmarks, OCCT &/or combined furmark & Prime95. I'd like to simultaniourls monitor these values.

CPU Frequency
CPU Temperatre
GPU Frequency
GPU Temperature
PCH temperature
Ambient Temp
overall noise emission at the users POV (ie microphone where your head would be)
FPS in game/benches

The idea of this is to evaluate the turbo performance and abilities and see if we have any variety between chassis'.

It would also be nice if we could do 1 or 2 with the 4800mq/4900mq to see what they perform and turbo like in comparison to the little 4700mq which has a significantly lower base clock (although I've never seen my run that low)

 

Temperature measurements of the 4700MQ inside the Dragon Edition. Pretty much confirms that the Anandtech review was faulty as hell. Exact same notebook as Anandtech tested.
The CPU averaged at 80C with max temperature of 90C. No CPU throttling found.



Source: VR-Zone China