Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

What GPU heat did you get?


  • Total voters
    46



The rImac with 295x is fine, i get consistent results and i am happy with it. when the 295x starts to fail or catch fire, NONE have yet, then we shall call the fire department.

Future driver and SMC updates might offer improvements but until then who cares. this thread was to post results from this bench, thats what has been done. it seems the worst scores are from fans which don't throttle up when needed but perhaps a Smc reset will help those people. I have melted desktop PC GFX cards so its no surprise some will run hotter than others. but the most disturbing thing is people seem to want it to fail just to say they were right in wanting the 980m or what ever there pref is.

If you don't want to risk the 295x stick to the 290 no ones forcing anyone to buy the highend GPU, but until i see failures and lots of them i don't think the 295x is such a big issue as some would have you believe.

I say enjoy your rImac and let the others worry if there gonna melt or fail. life to short to keep going over the same old stuff time and time again.
 
Take it easy, my iMac is the same as all the others, I had the fan problem and considering that a lot of people here is struggling to deal with the issue of an expensive machine not working properly, I thought that was nice to point out the cause of the issue and to provide a solution. :confused:

i am with you, i never experience any throttling and i have all the modern games with max settings. mine never exceed 101 and this is a normal operation for and chips.
 
Today my rImac finally arrived :cool:

I've done the Valley test and i've noticed that the temperature never exceeded 100 degrees....most of the time at 98-99 !

This is my first score.



 
Today my rImac finally arrived :cool:

I've done the Valley test and i've noticed that the temperature never exceeded 100 degrees....most of the time at 98-99 !

This is my first score.

This gives me some hope, although it scored worse than a 2013 iMac.

Edit: I'm assuming no third-party fan control apps?
 
Today my rImac finally arrived :cool:

I've done the Valley test and i've noticed that the temperature never exceeded 100 degrees....most of the time at 98-99 !

This is my first score.

[url=http://s6.postimg.org/sgk7ai7l9/foto_1.jpg]Image[/url]

[url=http://s6.postimg.org/4oavyz5kd/foto2.jpg]Image[/url]

That's great, probably the first I see that didn't exceed 100C.
I'd be interested in knowing the firmware version of your R295 but I think that information is available only from the catalyst control center under bootcamped windows.
 
This is mine for comparison: Bios version 015.048.000.005 date 2014/09/29
 

Attachments

  • Capture.PNG
    Capture.PNG
    305.3 KB · Views: 180
This is really surprising! You might have gotten the "good thermal paste" edition :).

Perhaps he's just spent all his money on the RiMac and can't afford central heating ;-)

Seriously though, the GPU temp will be directly related to the ambient room temp (of course) so the same iMac tested in a 16C room will read 7C lower than same tested in a 23C room.

That alone will account for quite a big part of the variations we are seeing here.
 
Seriously though, the GPU temp will be directly related to the ambient room temp (of course) so the same iMac tested in a 16C room will read 7C lower than same tested in a 23C room.

That alone will account for quite a big part of the variations we are seeing here.

No matter the ambient temperature (within reason), the GPU will reach its maximum temperature during sustained 100% usage. The rate at which it reaches it will differ, and the rate at which the fans must spin to keep it from overheating will also differ, but it will ultimately reach the same temperature regardless of air-con or not.
 
That's great, probably the first I see that didn't exceed 100C.
I'd be interested in knowing the firmware version of your R295 but I think that information is available only from the catalyst control center under bootcamped windows.

Let me have time to install bootcamp and windows 8 ;)
 
I uninstalled all my fan control applications.

Reran the benchmark. Was mostly on 100 degrees, rarely increased to 101 degrees. This is better that original 104 degrees and 102 degrees after the SMC update.

I think there's another point to consider here - I think it takes time for the thermal paster to "catch on", several cycles of heat and cool down will make is "stick" better. So, since I have the machine for longer now it might be related.

Anyway. I decided to remove all apps and just enjoy my computer. Seriously, it runs great, excellent performance, and I have warranty for 3 years.
 
I uninstalled all my fan control applications.

Reran the benchmark. Was mostly on 100 degrees, rarely increased to 101 degrees. This is better that original 104 degrees and 102 degrees after the SMC update.

I think there's another point to consider here - I think it takes time for the thermal paster to "catch on", several cycles of heat and cool down will make is "stick" better. So, since I have the machine for longer now it might be related.

Anyway. I decided to remove all apps and just enjoy my computer. Seriously, it runs great, excellent performance, and I have warranty for 3 years.

This can be a factor to consider, for SMC update what do you exactly mean ? a Reset or a software update ?
 
They had this kind of problem earlier with Pro.

Mac Pro EFI Firmware Update v2.0

This update is recommended for all Mac Pro (Late 2013) models.

This update improves system reliability during reboot, resolves an issue with memory self-test, and improves graphics power management when using Boot Camp.

http://support.apple.com/kb/DL1714

http://images.anandtech.com/reviews/mac/MacPro2013/CPUGPU7zip.png

Take a look at VDDC with D700, 1.062v, M295X is freaking mobile card and somewhere in this or other thread i saw 1.13v for M295X.

D700 specs:

D700.png


M295X specs:

a4d29932.png


1.13v M295X:

attachment.php
 
Last edited:
Yes, the voltage of the M295X is 1.13V but compared to a D700, they have same SPs (2048) but the D700 4.3 Billion of transistor and the M295X 5 Billion.
We shouldn't compare different architectures, they're similar but not the same.
If they change power management settings of the M295X could result in a decrease od the TDP (W absorption) so less performance and heat or increase the TDP, more performance but more heat and 100/200 rpm for the fan to do the same job, so faster but noiser.

I don't think they would do any of the things above, to me the system is balanced as it is, fast and quiet on 90% of the productivity tasks, on heavy rendering (Final Cut or similar) the fan hardly get over 1800 and while gaming hardly over 2000/2200.

----------

A reset, just unplug the power from both ends of the computer for 15 seconds.
Well, very quick fix :)
 
Last edited:
No matter the ambient temperature (within reason), the GPU will reach its maximum temperature during sustained 100% usage. The rate at which it reaches it will differ, and the rate at which the fans must spin to keep it from overheating will also differ, but it will ultimately reach the same temperature regardless of air-con or not.

Sorry mate, but the above is simply and categorically not correct. I am afraid you are mistaken.

I don't say this as a matter of opinion, but as a statement of scientific fact.

In a freezing cold room at 0C, if you get a 80C GPU reading after say 20 minutes of running, then that same GPU will read precisely 100C if you do the same test in a 20C room. Or 101C is the room temp is 21C. It's absolutely directly correlated.

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

If you are interested to understand why, here's why:

Temperature is a function of heat and in a steady state, the temperature of a GPU (or indeed anything else in the universe) is such that the rate of heat input matches perfectly the rate of heat output. This is basic physics.

As you input energy into the GPU, its temperature will rise until such point that the rate of energy coming out as heat is the same as the rate of energy going in. (Rate of energy = power = watts, by the way.) It is then in a state of equilibrium.

But the rate of heat output is deteremined by the temperature of the object in relation to its surroundings. Heatsinks for example are measured in their ability to dissipate however many watts, per degree C... above their surroundings. The bigger the temperature difference the faster the rate that heat is emitted.

The difference between the GPU temperature and the surrounding room temperature is what determines the rate of heat dissipation, and therefore determines the temperature at which equilibrium is reached.

Therefore, in a steady state, the GPU temperature is precisely and directly determined by the room temperature.

Try it if you don't believe me! Take you iMac outside and run the same tests and see what results you get! They will not be the same as inside, if it's colder outside (or warmer!)
 
Last edited:
Sorry mate, but the above is not correct.

I don't say this as a matter of opinion, but as a statement of scientific fact.

Temperature is a function of heat and in a steady state, the temperature of a GPU (or indeed anything else in the universe) is such that the rate of heat input matches perfectly the rate of heat output. This is basic physics.

The rate of heat output is deteremined by the temperature of the object in relations to its surroundings. Therefore, at full load in a steady state, the GPU temperature is precisely and directly determined by the room temperature.

In a 0C room if you get a 80C GPU reading after say 20 minutes of running, then that same GPU will read precisely 100C if you do the same test in a 20C room. Or 101C is the room temp is 21C. It's absolutely directly correlated.

Not quite - this is different.

The GPU would reach beyond 105C given the chance. The only reason we think it stays at 105C is because of the Mac’s fan, and the throttling. This is the missing piece from your, otherwise factually correct, comments. The GPU will eventually reach the same temperature – again, within realistic room temperatures.

An example; in a room of 10C, the GPU will hit 105C in say 2 mins, in a room of 20C, the GPU will hit 105C faster. It’s all about time, in this case. It will still hit 105C, though, because without the fans and throttling it would far exceed 105C as an isolated component. Both of these variables do not kick in until absolutely necessary. The fans in the 10C room won’t have to work as hard as in the 20C room to then keep the 105C maximum. However, both GPUs will run at the max temp as that’s what the iMac/hardware/firmware has been written to do - optimum power, minimum throttling.

You are correct, however, if the GPU was completely isolated without any fans or throttling. But that’s not the case. And talking about 0C household rooms is specifically why I said ‘within reason’.

Edit: Just to be clear, I 100% agree with everything you said - that's science - but it's not applicable in the case with the fan/throttle variables.
 
Last edited:
Not quite - this is different.

The GPU would reach beyond 105C given the chance. The only reason we think it stays at 105C is because of the Mac’s fan, and the throttling. This is the missing piece from your, otherwise factually correct, comments. The GPU will eventually reach the same temperature – again, within realistic room temperatures.

An example; in a room of 10C, the GPU will hit 105C in say 2 mins, in a room of 20C, the GPU will hit 105C faster. It’s all about time, in this case. It will still hit 105C, though, because without the fans and throttling it would far exceed 105C as an isolated component. Both of these variables do not kick in until absolutely necessary. The fans in the 10C room won’t have to work as hard as in the 20C room to then keep the 105C maximum. However, both GPUs will run at the max temp as that’s what the iMac/hardware/firmware has been written to do - optimum power, minimum throttling.

You are correct, however, if the GPU was completely isolated without any fans or throttling. But that’s not the case. And talking about 0C household rooms is specifically why I said ‘within reason’.

Edit: Just to be clear, I 100% agree with everything you said - that's science - but it's not applicable in the case with the fan/throttle variables.

This is correct, a room with a reasonable higher temp might cause only the fan to spin faster once the GPU reach the target max temp.

Reading the "SMC Reset fix" that some users have done, all the others experiencing troubles should try this solution first to do anything, the problem could have been only a misreading from the SMC sensors, reading 3-4-5° less than the chip internal controller and triggering the spin up of the fan too late with the consequence of triggering the thermal protection (throttling).

But remember to keep in mind that the thermal protection have nothing to do with the clock ranging arount 800 (up and down), this is managed by the total power draw of the board (125W max), independent from temps.
 
Not quite - this is different.

The GPU would reach beyond 105C given the chance. The only reason we think it stays at 105C is because of the Mac’s fan, and the throttling. This is the missing piece from your, otherwise factually correct, comments. The GPU will eventually reach the same temperature – again, within realistic room temperatures.

An example; in a room of 10C, the GPU will hit 105C in say 2 mins, in a room of 20C, the GPU will hit 105C faster. It’s all about time, in this case. It will still hit 105C, though, because without the fans and throttling it would far exceed 105C as an isolated component. Both of these variables do not kick in until absolutely necessary. The fans in the 10C room won’t have to work as hard as in the 20C room to then keep the 105C maximum. However, both GPUs will run at the max temp as that’s what the iMac/hardware/firmware has been written to do - optimum power, minimum throttling.

You are correct, however, if the GPU was completely isolated without any fans or throttling. But that’s not the case. And talking about 0C household rooms is specifically why I said ‘within reason’.

Edit: Just to be clear, I 100% agree with everything you said - that's science - but it's not applicable in the case with the fan/throttle variables.

My sincere apologies. I now understand where you are coming from and you are of course correct.

My science stuff is correct too, but that's not the point is it, the GPU will throttle at pretty much a fixed upper temperature no matter what the room temp is.

My bad. :eek:
 


No. I installed that update on Windows on my 5K before returning. It DID slightly improve FPS, but not the heat.

The video I posted in the other thread of the 90 seconds to throttling was with that driver update...

I believe Apple driver/firmware updates will always come via the App Store software update.
 
Update: talked to apple card support, he said the temperatures on 104 for GPU is not normal, also 100 for CPU is too much.

Left the computer at an Apple Store (well reseller store to be exact) We'll see.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.