Big difference between Crossfired temps.

[Fawkon]

Vortex 3.

Dual 7970m's.

Paid for Thermal compound.

Using zalmtec cooler.

Ran Furmark Benchmark

About 20% into said Benchmrk

GPU1: Max Degrees 77oC

GPU2: Max Degrees: 96oC

I shut it off as 96 was a bit worrying to me.

Question: Is it normal for the temperature to get that high and is it normal for a 20 Degree difference between GPU's?

Any advice. Thx.

 

[CEUOTC]

Vortex 3.

Dual 7970m's.

Paid for Thermal compound.

Using zalmtec cooler.

Ran Furmark Benchmark

About 20% into said Benchmrk

GPU1: Max Degrees 77oC

GPU2: Max Degrees: 96oC

I shut it off as 96 was a bit worrying to me.

Question: Is it normal for the temperature to get that high and is it normal for a 20 Degree difference between GPU's?

Any advice. Thx.

Thermal compound iirc only applied to CPU.

Look here Improving P170EM 7970M GPU cooling performance (foiled) for some advice on GPU cooling, l know it is for the P170EM none elite model, however the procedure should still be the same.

Regards.

C.

 

[LiddleP]

Hi there CEUOTC

Am I reading you wrong, are you saying the paid for thermal paste is ONLY applied to the CPU and do NOT include the GPU's !?

Pls. clarify

Thermal compound iirc only applied to CPU.
...

 

[tom_gr7]

just an fyi,

Furmark can be pretty bad for gpu's, im not sure why, but I recommend against running it at all!

 

[LiddleP]

just an fyi,

Furmark can be pretty bad for gpu's, im not sure why, but I recommend against running it at all!

Good advice, so what utility would you recommend to test the stability of your gpu(s) ?

 

[tom_gr7]

Good advice, so what utility would you recommend to test the stability of your gpu(s) ?

to be honest, I wouldn't bother... I'd just game on it, record/observe the gpu temps, with either gpu_z or msi afterburner, to check the temps. Other than that I honestly wouldnt bother.

I did run 3d mark once but that was a waste of time, it only ran a few minutes and gave me a score.

If you have a laptop, ensure the base of the laptop can breathe, or get a cooling pad to help the temps, they can get pretty toasty.

If you are gaming on a desktop,
- for single card setups you probably wont need to anything at all!
- for sli/crossfire rigs, I'd just make sure you have a side/bottom fans blowing towards the two cards, as well as top/rear exhaust fans getting that toasty air out the rig.

If you are interested in cooling desktops, there is a good article in this thread, post 7,- https://www.pcspecialist.co.uk/forums/showthread.php?21495-useful-tech-maintance-videos-articles

 

[Fawkon]

GPU no. 2 was reaching 96 degress while playing X-com also. Is this temp too high for a laptop which is on a cooling pad and why such a difference (nearly 20 degrees) between them both?

 

[CEUOTC]

Hi there CEUOTC

Am I reading you wrong, are you saying the paid for thermal paste is ONLY applied to the CPU and do NOT include the GPU's !?

Pls. clarify

When l got the thermal paste option it was only applied to the CPU, l took off the thermal pad on the GPU cleaned up the HS plate and GPU and re-pasted with IC 24.

Regards.

C.

 

[LiddleP]

When l got the thermal paste option it was only applied to the CPU, l took off the thermal pad on the GPU cleaned up the HS plate and GPU and re-pasted with IC 24.

Regards.

C.

Fair enough and thanks for your reply - I however already took it upon myself to clarify that this was not the case with PCS via live support and they informed me that the paid for thermal compound is used on both CPU and GPU(s) which is what I had expected to hear. Personally I would not pay that much for just having it put on the CPU as I can can get enough compound to do all of them and still leave me with spare for other systems, for under that cost.

I did intend to post clarifying the situation but I was otherwise distracted by other issues.


@Fawkon, I would personally not advice that you to take it upon yourself to make any modifications suggested by anyone to your laptop to resolve your observed problem as it will inadvertently invalidate your warranty - I am assuming here it is a new purchase, if so you are going about things the right way, unfortunately the XMas rush might be getting in the way of your receiving the appropriate level of support in the matter.

Personally, I think you have done more than enough by acquiring a cooler when there should not have been need for it - unless you are in an environment that might impair operation and even then as you have an external cooler to assist - what you have there is a system that is ideal for extreme gaming by design.

IMO, though temperature differences are to be expected due to design (the CPU more or less shares into the slave GPUs fan and some heat is inevitably exchanged) I would expect to see a difference of about 5oC to 10oC Max.

What you show there is IMQO excessive, it indicates a problem that would more than likely and relatively easily be resolved by a repaste of the Slave GPU (GPU 2) and proper seating/alignment of the cooler.

Unless you are competent at it and even then, if this is a new system I do not advice that you do this yourself, design constraints necessitate a repaste of the CPU as well -, personally, I would read up on the PCS Customer Terms & Conditions especially as regards RMA's and warranty -> http://www.pcspecialist.co.uk/terms/
while rightly continuing to seek resolution from PCS.

I will continue to follow this thread as I am interested in seeing how things get resolved as I am pretty sure others are too, so please endeavour to keep us updated.

 

[Fawkon]

@Fawkon, I would personally not advice that you to take it upon yourself to make any modifications suggested by anyone to your laptop to resolve your observed problem as it will inadvertently invalidate your warranty - I am assuming here it is a new purchase, if so you are going about things the right way, unfortunately the XMas rush might be getting in the way of your receiving the appropriate level of support in the matter.

Personally, I think you have done more than enough by acquiring a cooler when there should not have been need for it - unless you are in an environment that might impair operation and even then as you have an external cooler to assist - what you have there is a system that is ideal for extreme gaming by design.

IMO, though temperature differences are to be expected due to design (the CPU more or less shares into the slave GPUs fan and some heat is inevitably exchanged) I would expect to see a difference of about 5oC to 10oC Max.

What you show there is IMQO excessive, it indicates a problem that would more than likely and relatively easily be resolved by a repaste of the Slave GPU (GPU 2) and proper seating/alignment of the cooler.

Unless you are competent at it and even then, if this is a new system I do not advice that you do this yourself, design constraints necessitate a repaste of the CPU as well -, personally, I would read up on the PCS Customer Terms & Conditions especially as regards RMA's and warranty -> http://www.pcspecialist.co.uk/terms/
while rightly continuing to seek resolution from PCS.

I will continue to follow this thread as I am interested in seeing how things get resolved as I am pretty sure others are too, so please endeavour to keep us updated.

Appreciate your taking the time to give me a detailed and thought out reply. Currently tempted to take it apart myself and do the job and will contact PCS to see if it will void warranty. I will keep you updated.

 

[LiddleP]

Good Luck with that, will be watching out for your updates.

 

[LiddleP]

Well, back again to provide another update - DX11 definitely has major problems so I have abandoned that for the foreseeable future.

Using DX9, I have been able to achieve steady temps of 60oC & 80oC Master & Slave GPU, where like yours my temps had started exceeding 70oC & 98oC and at around 99% utilisation while monitoring - Set at High Performance, my current GPU utilisation is sitting at around 70% for both GPU's and FarCry3 is running in Ultra with the frame rate hunting between around 43 & 63 FPS but giving me around 53 FPS on avg - For me, this is manageable for now as I get a smooth gaming performance, no shakes or judders.

Let me explain it this way, When I come in game I am actually achieving frame rates way above 80 FPS and between 94% and 99% utilsation with relatively stable temps sitting at 66oC & 94oC. While this is way too high for my liking I feel I must mention that this is while monitoring temps via OSD by Escaping the game and having it sit in its menu, in actual gameplay everything falls away, the temperatures drop to 60oC and 80oC and utilisation too drops to about 69% giving me my avg of 53 FPS, don't ask me why.

If there was a software solution to aid limiting utilisation at between 75% & 80% or at least limit FPS to around 75 FPS the heat issue would be contained and a proper repaste of the Slave and CPU will assist things along even further - something I might try down the line if/when I can be bothered purchase some Shin Etsu G751 and finally open up the system (If PCS allow for this to happen without voiding my warranty), it has been reported that Clevo have a problem here but if the repaste is done properly, temperatures will drop for both the CPU and Slave GPU.

So far I have not really been able to monitor my CPU temps which is where I feel the real solution to this overheating problem lies. if one was able to underclock the CPU without suffering performance-wise I believe the temps will fall away even more, it is a pity that Clevo did not deem it fit to provide the facility in bios to underclock as this IMO would probably have been the easiest and cheapest fix to the current overheating issues.

All in all, I am not the happiest person with my build as it stands and especially as regards PCS assurances but dealing with things is better for me than simply looking to return a system that is not as one was led to believe, atm my system is getting along just fine (atm, "touch wood" says LiddleP, touching his bonce), I am playing away and not focusing on the overheating as being a major issue.

IF you can deal with the other problems you had with your WiFi I would suggest you simply get on with your gaming while of course still monitoring things for the interim and see how it goes, hopefully it works out well, if not do contact PCS in writing and by phone and get things dealt with.

 

[LiddleP]

Just thought I'd add, with the stable temps achieved above, even though it scales down to about 53 FPS @ 70% utilsation, when I need it it scales back up to 80 FPS @ about 97%/98%.

If I sit on the menu monitoring at 80FPS my temps go back up to 65oC & 89oC respectively with utilisation at 97% & 98%, which while still high is heck of a lot better than my previous 70oC & 98oC - the point to note is that when I resume gaming it all drops back down, at the cut scenes for instance it goes back up as at other points in the game, my game is smooth with no stutters.

I am using the Clevo drivers for Win 8 from the 24th of Nov where you would be using the Win7 ones.

The profile applied against Farcry3_d3d11.exe mostly uses the standard settings apart from OpenGL settings which has triple buffering selected and CF mode is at default where before I had tried the provided Farcry2_rt.

In Windows I am using the High Performance setting.

In FarCry3 itself, I have the option to set GPU Maxed Buffered Frames to 5, not that I feel it helps much but that is what I did, I do not know if you have any such option available to you in your game.

To install the Clevo drivers and CCC all I did was rename the Ati folder to whatever it was +"-bak" and ran the install, no cleaners where employed.

Hope something here helps and you can get your game on with no hitches, if you have any questions or feel I can be of assistance just PM me or post up here and I'll see what I can do.

 

[Fawkon]

Just done a repaste....

Just did a repaste job..... My temps are better when idling but the second GPU 2 goes under load it goes from 40oC to 103oC straight away.... I tried repasting again but still the same issue.... Please help....

 

[Karnor00]

Dual 7970m's are always going to run pretty hot. I'm not too surprised that one is hotter than the other - in an SLI setup one GPU tends to end up doing a bit more work, and the cooling to each is unlikely to be identical.

 

[Fawkon]

Dual 7970m's are always going to run pretty hot. I'm not too surprised that one is hotter than the other - in an SLI setup one GPU tends to end up doing a bit more work, and the cooling to each is unlikely to be identical.

103oC is too hot. Running at 90 odd temps is acceptably hot but not this high...

 

[edgez]

Hi Fawkon,

You managed to get your crossfire issue resolved then? what was the problem?

I notice when I monitor my GPUs with afterburner that one always runs hotter than the other, I guess this is common. I have also noticed the temp on the hotter GPU to be in excess of 100 degrees, so your not alone. I don't think I am overly concerned about it though, they are of course going to get very hot, and if both you and I are getting these temps then surely others are too.

 

[LiddleP]

@Fawkon, as you have taken it upon yourself to deal with all I can suggest is you look to people that make use the of the Thermal compound you are making use of. There could be a number of reasons for the increase in temp, too much paste for one. One thread I read indicated that some pressure needed to be applied to the GPU while it is cooling (after running it warm/hot).

***Update***

My prior posts were probably me being way too optimistic, after I had made them I my system started crashing more frequently and was frequently achieving temps of 70oC & 98oC (Master & Slave) respectively and subsequently crashing or freezing whenever I came to the more demanding sections in Far Cry 3, the longer cut-scenes being the most demanding.

The overly frequent crashes brought about the need to contact PCS and on informing them of the problems experienced with the laptop they say that my temperature readings are an oddity and the first Laptop with the 7970m's cross-fired that is experiencing overheating issues like this, despite the posts here that indicate the contrary. This leads me to believe that those reporting these same overheating issues in the forum have not made any attempts to deal them by contacting PCS directly either by phone or email, it would definitely be in their interest to do so.

I have since booked an RMA to have the system checked out which, although authorised, has not occurred yet as DPD, the collection service, have allegedly visited me twice to collect but missed me on both occasions - on both occasions I have been right there waiting on them.

DPD collection
On the first occasion there had only been one caller just before midday on the day, this was 3 hours ahead of the DPD scheduled time and I had not answered the call as I do not respond to unscheduled callers - I do have a fair idea of of just who that caller was and am now looking to confirm as I now do not think it was DPD.

On the 2nd occasion I had called PCS at 16:42 to verify if DPD would still be coming to collect for the scheduled time which was between 16:00 & 17:00, this as it had been snowing and I though it might interfere with the collection - Mustapha, the individual I spoke to at PCS, had contacted DPD to verify the position and had been informed by them that the driver was en-route and would be arriving shortly - My call had lasted just over 12 minutes but the only thing that arrived less than 2 minutes after I got off the phone was an SMS from DPD saying they had once again been to collect but I was unavailable - This raises a couple of questions, the first being "do they have the right address ?"

On both occasions no one approached my door or rang my doorbell at anytime within a 3 hour window either way of the scheduled collection date and time.

The Laptop has since suffered another crash out of the ordinary and with this crash I had static-like feedback through my headphones. The laptops GPU temperatures have changed from an average of 65oC & 89oC (Master & Slave GPU's) to one of 65oC & 76oC.

While the GPU's appear to be functioning okay and the Slave GPU temperature has taken the very welcome drop of over 10oC which would ordinarily be welcome if the norm, actual gameplay is not at all smooth and indicates that something untoward has happened which does not show up in afterburner or with any of the standard windows utilities.

Though concerns with DPD having the right address to make the collection have stopped me from re-booking collection for a third time, I am prepared to make myself available for whatever time PCS re-books the collection for and DPD is able to visit my address and actually make collection - this time their visit, or not, will be on record.

 

[vanthus]

It does seem DPD are going to the wrong address,If you missed them they would leave a card.

 

[LiddleP]

It does seem DPD are going to the wrong address,If you missed them they would leave a card.

Ahhh, that might explain things as no card has been left on either occasion, I do however find it strange as they had no problems delivering the laptop.