*** Official Clevo P870DM2/P870DM3 (Sager NP9873/NP9872) Owner's Lounge! - The Phoenix 2 is here! **

This machine is a huge upgrade from my fx6300 and hd7850 potato rig. I guess my games also need some upgrading

 

Try overwatch at lowest settings with the fps cap at 300. That will push the temps pretty high. Or if you have The Division, give that a shot.

And CSGO and DOTA 2 are great games. Way more fun than some of the AAA's out there.

 

don't forget the fake 21:9

 

Oh yeah. I think Overwatch's render scale also helps in stressing the cards. Its like in game DSR.

 

Yep, despite me not playing Overwatch much anymore, that games resolution scaling and uncapping fps is what I used to stress test game performance after a recent repaste on my 1080. That and The Division.

Whoops meant my 980 that I can't upgrade, honest mistake I swear!

 

Well now that's what I'm going to use for testing.

 

Please see my post at http://forum.notebookreview.com/thr...oenix-2-is-here.794530/page-282#post-10324341

It will make you less concerned.

 

Company of Heroes 2, Xcom2 and tomb raider will have you well on your way.

 

Please don't take this as rant, but to me it looks like this statement from nVIDIA means as much as "Great the card will die sooner and they'll have to buy a new one" therefore they're of course not "worried" per se.

It can't be healthy for the card to run at such high temps over an extended period of time. I guess in 2-3 years time this thread will either prove me wrong and laugh at my concerns (what I sincerely hope) or right and nVIDIA's happy.

Just my 2 cents, please don't take it the wrong way, I'm still going to upgrade either way

 

90 is my cut off for normal, that doesn't really leave much wiggle room for normal extra heat like from things like dust. If I had a system that normally ran at 90 all the time I'd have to carry a compressor around with it.

 

Unless you're benching 24/7 you wont hit 90. Idle temps are 40-50

 

How much would an extra 330w and converter box from xotic be?

 

On that same note, was it ever determined what differences, other than a longer cord, are between the AC-100 and AC-200?

 

In regards to this, the 1080 GTX is a Clevo product, right? What I mean by this is that the GPU is a Clevo PCB with nVidia components, but I don't know who is actually manufacturing it. So, is it Clevo stating 111C is the sweet spot for their version of the 1080?

If that is correct, what has my head spinning is that the suggested normal temps for the desktop 1080 is ~84C*. That's a pretty big difference.

* -

Spoiler

Nvidia GeForce GTX 1080 Pascal Review - http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-11.html

 

Oh you most definitely will. Try playing something like Overwatch/The Division / Witcher 3 / DeusX / Tomb Raider for about 30 mins.

 

... Sorry, but what? Those temps were measured from doing a simple Firestrike run wich only lasts a few minutes.

Of course games like Witcher 3 will put your GPU usage at 99 percent as well, unless you are bottlenecked by your CPU, so temps won't be any lower.

I sincerely hope that some other users here receive their beast soon and measure lower temps (~85 max) to calm us down

 

This.

Gesendet von meinem SM-G928F mit Tapatalk

 

Now that I think of it, I notice a heat increase just sitting on the OW title screen on my rig.

 

Owerwatch pushes the GPU pretty nicely, specially with the render is set to anything over 100%. In game even more so. Specially if you reduce the details and let the FPS fly at a higher render scale.

 

Exactly...thank you for posting reality.

The only other way you would get there is through stress testing, like with Furmark or other such software.

 

I was messing with the settings before, didn't realize that had an effect on temp that much. Will test tonight. It's high noon somewhere.

 

Basically pushing more frames at a "virtually" higher res. Results in higher temps, than those at Ultra/Epic at the same render scale, since the fps is lower.

Spoiler: BWHAHA

 

As per NVIDIA development guidelines for mobile Pascal the system has to trigger an emergency shutdown if 97c are sustained for more than 30s.
(The same trigger value for mobile Maxwell was 94c)

EDIT: For the CPU it is 100c * 30s

These are the recommended values. Actual trigger temps depend on EC as well as (v)BIOS revision. Throttle values start long before the above.

 

No one said 111C is a sweet spot...it is TDP when it will shut the computer down for protection. At 92C the card will undervolt to cool, and shut down at 97C. So at 90~91C it is below the undervolt setting.

Please be patient, and wait for the Prema BIOS/vBIOS to see what that will do to the temps for anyone concerned about 90-91C.

 

Thank you Prema...I stand corrected.

 

I don't know man, but a GPU running at 90c+ is pretty terrible if you ask me.

 

More worried about the ball grid melting/cracking in the BGA thats connecting the chip to the board.

 

BGA on an EVOC P870DM2 or EVOC P870DM3 ?? You must be talking about some other model that is not consistent with this thread

 

Good thing I waited before ordering, because the reported temps are too high for my liking. I am thinking long term though. If any of my computers go over 80c for an extended period of time, I do a complete breakdown of the tower and components. Mind you, that's on my desktops. I am even crazier when it comes to my laptops.

Hopefully Prema's Bios does the trick but until his Bios are out and the panel I want comes in, I'm going to sit back and watch.

@Bill56 You have the 1080p 120hz panel, how does it look to you? Is it a TN panel like I read on this thread? Do you think you're going to miss having G-Sync on it?

 

Please note that this hardware whether we are talking the GPU and or the CPU are designed to withstand high 80's to 90c without damage to these components.
We have checked this with the factory many times....

 

The die is still soldered on the PCB of the MXM board. I think should still be using BGA in that regard.

I guess I am just extremely paranoid about temperature. I try to keep my components below 90c most definitely if possible.

Edit: I guess as long as Clevo is saying its fine, its fine. They probably have done their testing to ensure a minimum level of RMA.

 

Hey everyone I got my machine but I'm pretty busy and trying to tweak this thing at the same time!

If you have feedback on any particular agents you have dealt with then PM me.

 

The GPU core is still soldered to it's board using solder joints. I'm not sure if they are better than BGA, but regardless of that, the entire board could suffer from being exposed to high temperatures.

I wouldn't be too worried either way.

80C is not too high, don't worry about it.

I mean, 96 on load would be really too high.

 

Yes, ask if he as another unopened unit, and go with that. Let him side line the partial broken unit and let him wait for parts.

Then you have a clean start on a new unit, and he can sell the other one to someone else.

 

That's probably not going to work

DOA parts happen. Something jounced in shipping, and pop went the 1080, happens to the best of us.

Just let them handle putting together a unit they want to ship you, and you have the 2 year warranty to cover you.

 

I'm looking forward for some numbers (temp - benchmarks and clockspeed)
I know you already have bench software already downloaded

 

Nice!
Mine shipped today.

 

Okay, time for me to chime in here...

I had a D900F for four years. My 280M ran 95c+ a lot of the time, and even at 105c for HOURS a day sometimes (I was unable to get better cooling, I used max fans, I was under 60fps in the games I played, etc). And it worked fine, as far as I know. I got no failures from the card itself straight up until the machine died, and this includes four years of awful voltage fluctuations. I am not worried about cards sitting at 90c every day, even though I personally would like to see an average of 80c or less without overclocking, whether or not a machine needs max fans for it.

HOWEVER. If YOU are getting 90c in a simple firestrike run, which at night in my still-extremely-hot-and-humid room I don't even approach with my 780Ms overvolted at 1006/6000 (up from the superclocked 850/5000 svl7 vBIOS, which is up from the 721/5000 plus boost that is the stock 780Ms speeds), not even hitting 82c should I run it from idle, then that is going to cause SERIOUS issues for everybody in hot areas. I'm going to be honest. I can make games push my video cards FAR hotter, especially in the daytime, at stock speeds/voltage, than Firestrike from idle will ever do, due to the short nature of the run. If you're hitting 90c and throttling in firestrike at stock speeds, that machine's thermals are broken. If you were to give me it, I'd likely have to apply CLU on the GPUs to take advantage of the vapor chamber contact points, prop up the machine more, modify the backplate at the least slightly for more airflow, and basically live in max fans, especially with higher FPS gaming at 120Hz 1080p, or at 4K 60Hz.

I'm not a big-time overclocker. I'm not a BIOS modder. But I'd like to think I know enough about PCs and these machines in general to understand that something has to be up here.

So, here's what. I want some information. If you're hitting 90c in firestrike, is this unmodified cooling? Is the dust filter in place? Are the vents open? What were the temps you had ambient when you ran it? Did you try maximum fans (fn + 1 in case you are unaware)? I gotta get some more information, but I can tell you certainly, if he isn't living in Dubai or the Sahara Desert or something and that was with max fans and a good contact with the thermal paste? I won't even be able to use that machine at stock. I will never achieve proper stock speeds, far less overclocking. And if I expect that machine to last me 3 years or so, I can guarantee you when games start getting difficult to run and the cards are getting actually pushed to their limits, we're going to run into problems.

 

When I said 80c earlier, I was meant to say 90c. 80ishc on certain things is probably fine imo.

The only thing I would really worry about is the pascal's AGGRESSIVE clock throttling. At least until @Prema fixes nvidia's stock vbios.

 

OIC. Those were ceiling figures. Thx for the clarification.

 

I'm not too familiar with all the different panel types, but I don't notice any difference from the laptop monitor when compared to my benq xl2411. For my purposes (gaming, coding, general use) The panel works great. I've never used G-sync before so I can't comment on it, but for most of the games I play on a daily basis, I turn on V-sync and because this machine can reach a constant 120fps, games are smooth enough for my taste.

 

You wont notice too much of a difference going from 144Hz to 120Hz or vice versa. But going from 60-75Hz to 100+ , thats a VERY noticeable and nice difference.

 

Y

yeah they also remove the two black parts aswell

 

http://www.3dmark.com/fs/9908593

9587 firestrike extreme

+100/+100 gpu
4.2GHz cpu at -134mv

Max fan, temps in firestrike peaked at 66c gpus (both same) and 74c CPU.

GPU frequency bottomed out at 1823mhz and peaked at 1949mhz.

 

So the other two owners so far might have warped heatsinks you think? Or undervolting the CPU helped draw enough heat out of the systems to help.

 

Nice! Now those are some realistic temps. What is the ambient temperature like ? And were you using a cooler under it ?

 

Read the thread or use search http://forum.notebookreview.com/thr...oenix-2-is-here.794530/page-284#post-10324370

 

@Meaker@Sager

That's what I was thinking, what are the other guys doing that you aren't

You don't have it sitting on top of the output vent of the air-conditioner do you??

Can you see if you can try to load the GPU's into that temperature range? And, then let us know what can do that.

I have been able to push GPU's into their pain zone with various CUDA jobs, and some games and benchmarks, so I know it can be done with those power hungry Pascal MXM GPU's.

Let's see'em running hot and bothered

 

those are some nice temps, what's your secret? It seems unlikely that 4 users would all have faulty heatsinks. Im getting low 80s on gpu with max fans, and cpu mid 80s when running firestrike.

 

What voltage and clocks you running on CPU?