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

marcel500

macrumors regular
Nov 18, 2006
213
42
I think the "lag" is more related to the Yosemite driver / coding issue. Not really hardware hardware related.

That might be true and I don't exclude this option but the fact is that the computer lag's out of whatever reasons. Also I don't understand why they can't put in a proper graphic adapter instead of using these mobile variants. We are talking desktop and not notebook.

Anyway... this is just my personal opinion.
 

MrNomNoms

macrumors 65816
Jan 25, 2011
1,156
294
Wellington, New Zealand
http://barefeats.com/imac5k9.html

He only got his to 100C, the 780m was 82C.

The room he was testing in was 22C.

Meanwhile, he found fan speeds literally DOUBLED in the 5K.

One of the compelling reasons that is motivating me to purchase an iMac non-Retina over the iMac 5K but it'll be interesting to see what happens when the skylake CPU is made available and whether Apple continues with AMD or whether they swing back to nVidia.
 

dexterbell

macrumors 6502a
Jan 29, 2015
855
16
Why did Apple switch to AMD anyways? That was the main reason I didn't get the 5K and have kept my late 2013 27 inch. I have the GTX 775M in this.
 

MrLeek

macrumors newbie
Mar 17, 2015
11
0
http://barefeats.com/imac5k9.html

He only got his to 100C, the 780m was 82C.

The room he was testing in was 22C.

Meanwhile, he found fan speeds literally DOUBLED in the 5K.

But this just confirms what even the most die-hard 5K fans know - the M295X runs hotter than the Nvidia option.

The real question is this: does it run too hot? If it ran too hot then would Apple push it out of the door regardless? I don't know - I'm a relative newcomer to the world of Apple. From the various documents linked on this thread it seems that the GPU is designed to run at a high temperature.

The other question I have relates to unexpected 'peaks' in temperature. I've had an application running and seen the GPU temperature rise (following by the fan ramping up) when there is no change on screen. At first glance there appears to be no reason why the GPU would suddenly decide to generate 6-7 Celsius for what was essentially a static image - which had been fine for the previous 10-15 minutes running at a steady temperature.

I do not understand why the GPU could suddenly ramp up for no apparent reason.
 

UniDoubleU

macrumors regular
Aug 14, 2014
160
1
Thailand
Solutions?

Okay, Apple shot themselves in the foot for not using GTX 980m or at least redesign the cooling system to accommodate the R9 m295x.

Anything helps lowers heat peak to sub-100 degrees? Anyone tried to reapply thermal paste? Or other cooling hacks? Software side?

For such a gorgeous Mac, have MVC gotten his hands on one? I'm particularly interested in possible eGPU solutions. I know, it's stupid to want to buy some £700~ brick to complement an already expensive computer but it would definitely very much prolong the life of the machine as the Display and CPU part is already excellent.
 

Georgio

macrumors 6502
Apr 30, 2008
369
38
Essex, UK
Why did Apple switch to AMD anyways? That was the main reason I didn't get the 5K and have kept my late 2013 27 inch. I have the GTX 775M in this.

Apparently AMD off-loaded the 290/295 chips after they weren't wanted for the PS4 allegedly, so probably got them on the cheap as well.

The 980 chips weren't ready for when Apple would have needed them, so unless they delayed the launch of the RMac to accomodate the 980 supply, they pretty much had no choice but to go with the AMD chips or delay for 6 months.
 

Astelith

macrumors 6502
Oct 10, 2014
250
13
Spain/Italy
http://barefeats.com/imac5k9.html

He only got his to 100C, the 780m was 82C.

The room he was testing in was 22C.

Meanwhile, he found fan speeds literally DOUBLED in the 5K.

I have a question out of that test, double RPMs but only 7dB of difference ?
It's not that loud after all, or... the old iMac is not that silent :confused:

----------

And... I'm not the only one in this planet with a "100°C" iMac 5K :D
 

andy9l

macrumors 68000
Aug 31, 2009
1,699
365
England, UK
AMD Radeon R9 M295X Core Clock Throttling, Heat, and Performance

I have a question out of that test, double RPMs but only 7dB of difference ?

It's not that loud after all, or... the old iMac is not that silent

7dB would sound, to a human, almost twice as loud. 10dB would be almost exactly twice as loud - so 7 is about right.

---

Edit: The 5K iMac follows an identical design to the 2013 and 2012 iMacs - so a faster fan = a louder iMac. If the fan is spinning twice as fast, it will be pushing towards being twice as loud.

That article, for me, sums up the shortcuts Apple have taken with this product. Truly disappointing.
 
Last edited:

matreya

macrumors 65816
Nov 14, 2009
1,286
127
That article, for me, sums up the shortcuts Apple have taken with this product. The i7/M295X 5K iMac is an absolute shambles, in my opinion. Truly disappointing.

For those of us not challenging the GPU, we should be able to get a useful lifetime out of our rIMacs :)

I'm in an airconditioned room set at 24° C, and my rIMacs cores range in temp from 50 to 60° C with CPU load at 14% according to iStat menus...

----------

It's crazy, I don't understand why 10dB sounds 2x louder, I just know it does. Some weird human thing :p

It's because dB is a logarithmic scale.

I'm not particularly phased by fan noise, I have external drives with their own fans that make more noise.
 

andy9l

macrumors 68000
Aug 31, 2009
1,699
365
England, UK
For those of us not challenging the GPU, we should be able to get a useful lifetime out of our rIMacs :)

Well, even those pushing the CPU/GPU will ultimately get good use out of the machine. I'm not suggesting that these i7/M295X iMacs will fail - we've seen that they are not overheating they're just thermal throttling.

Those who bought the i7/M295X as an 'overkill purchase' are indeed unlikely to notice any issues discussed in this thread or elsewhere online - at least with Yosemite.

For those who push the CPU/GPU, it's just going to be a noisier iMac. If it's your first iMac, you'll know no different, and if you're in a loud/large environment you may not notice it. It's only a niche who will notice their components are actually too hot and are throttling performance - even then, only a subset of those people will actually return their iMac. Apple would have known this, and it looks like they've taken advantage of that.

If you did a cost analysis of designing, engineering and manufacturing a new physical design (or at least a new cooling system) vs. expected number of returns from customers who will notice the thermal throttling, which one do you think would win?

Annoying. :(

It's because dB is a logarithmic scale.

Ah, thanks!
 

andy9l

macrumors 68000
Aug 31, 2009
1,699
365
England, UK

MrLeek

macrumors newbie
Mar 17, 2015
11
0
I've been mentioning the sudden spike in GPU temp that was happening for no reason....well, now I've been able to replicate the effect in a game.

Orient the camera in a certain direction and the temperature spikes from 95-96 to 105 within a matter of seconds and the fan works overdrive.

Orient the camera away from whatever GFX effect is causing such a problem....and the GPU temperature drops back to 96-97 just as quickly.

I've tried to replicate the effect in different ways to try and narrow down the specific GFX effect(s) that causing it.....not had much success yet. But it's a very strange thing.
 

cynics

macrumors G4
Jan 8, 2012
11,959
2,154
Apparently AMD off-loaded the 290/295 chips after they weren't wanted for the PS4 allegedly, so probably got them on the cheap as well.

The 980 chips weren't ready for when Apple would have needed them, so unless they delayed the launch of the RMac to accomodate the 980 supply, they pretty much had no choice but to go with the AMD chips or delay for 6 months.

Have any sources for that?

Reason I ask is first the obvious, PS4 uses an APU. Plus aren't the dates way off to be just something laying around? Without me researching I was under the impression the PS4 was released long before that AMD series.

There is also the amount of R&D AMD themselves had to do to work with Apples proprietary TCON. This is clearly not a case of just any fast GPU works at least not without the manufacturer of the GPU's hand in the mix.

For the last reason I think we'll see AMD around for at least one more generation if not more. And because of bandwidth requirements I doubt we'll see an Apple 5k standalone monitor until Thunderbolt 3, so a very long time. I just don't see Apple going the path of dual cables, maybe but that just doesn't seem like something they would do to me.
 

Georgio

macrumors 6502
Apr 30, 2008
369
38
Essex, UK
Have any sources for that?

Reason I ask is first the obvious, PS4 uses an APU. Plus aren't the dates way off to be just something laying around? Without me researching I was under the impression the PS4 was released long before that AMD series.

There is also the amount of R&D AMD themselves had to do to work with Apples proprietary TCON. This is clearly not a case of just any fast GPU works at least not without the manufacturer of the GPU's hand in the mix.

For the last reason I think we'll see AMD around for at least one more generation if not more. And because of bandwidth requirements I doubt we'll see an Apple 5k standalone monitor until Thunderbolt 3, so a very long time. I just don't see Apple going the path of dual cables, maybe but that just doesn't seem like something they would do to me.

Just the jungle drums at work; obviously the PS4 was released way before the AMD's but like any manufacturer they'e always performing subtle upgrades over the life-time of a product to try and keep competitive and apparently that is what these were destined for in their raw state.
 

ioxakoyba

macrumors member
Jan 5, 2015
31
0
Well, again some observations from my tests:

I have been playing some Steam gaming titles both on OS X and Windows 10 (like Alien: Isolation, Duke Nukem Forever and Metro: Last Light) and at least with relatively short gaming sessions it seems that with my i7 & m295X combo the cooling seems to be able to keep the maximum CPU temp pretty consistently on 106 C degrees.

This does not seem to cause any noticeable thermal throttling, but GPU frequenzy fluctuates between ca. 760-830Mhz - which seems to be the default operating frequenzy for this chip (as the 850Mhz is the so called Turbo-Boost frequenzy).

I have been monitoring the temps with GPU-Z on Windows and as it also shows the GPU core clock keeping consistently between those numbers no matter is the GPU temp 90 C or 106 C, so I would say that as long as the temp stays on those numbers, the thermal throttling should not kick in (until in 108C which causes massive throttling and dropping GPU frequenzy to half or so).

However, after 30 minutes of Alien: Isolation I noticed that 107 C was the maximum the session reached, so this makes me worry a little bit. After all I am using Mac Fan Control to automatically push fan RPM to 2700 rpm on games (both on Windows and OS X) - and even with that spin the 107 C was reached - and this is very close to 108 C temperature causing massive thermal throttling and loss of GPU performance.

But what can you do? After carefull thoughts I am still planning to activate the 3-years Apple Care and manage with these temps. Hopefully they know in Cupertino what they're doing...

I am sadly missing those temps with my watercooled old gaming rig of keeping maximum pushed HD 7850 under 65 C :) However, I must admit that this Retina iMac is much more silent even with 2700rpm than my old rig.

I am playing with small headset possibly muffing some of the cooling noise, but I still doesn't understand people claiming this machine noisy :) Maybe the old iMac was complitely silent, but even with 2700rpm when running games, this riMac outputs just a comfortable humming of moving air. There is no fan rotation noise or bearing noise at all. It's very silent to me, comparing the most of the Windows gaming laptop / desktop PC's I've experienced.

Still, 107 C... :/
 

Mac32

Suspended
Nov 20, 2010
1,263
454
However, after 30 minutes of Alien: Isolation I noticed that 107 C was the maximum the session reached, so this makes me worry a little bit. After all I am using Mac Fan Control to automatically push fan RPM to 2700 rpm on games (both on Windows and OS X) - and even with that spin the 107 C was reached - and this is very close to 108 C temperature causing massive thermal throttling and loss of GPU performance.

But what can you do? After carefull thoughts I am still planning to activate the 3-years Apple Care and manage with these temps. Hopefully they know in Cupertino what they're doing...

Wow, that's not good. My Nvidia 680MX overclocked at +250/+375 goes up to max 85C (unless during very hot summer days, then I lower the overclock), and that is with a static fan speed at 2300RPM. Hopefully they'll come to their senses, and choose Nvidia next time (and improve the cooling paste application too!). Even the RiMac CPU is much hotter than my late 2012 iMac, again max temp 20 C+ higher. That's a lot!
 
Last edited:

CrankItUp

macrumors newbie
Feb 18, 2015
21
1
Germany
Hi folks, i ve made a 24h stress test marathon. Yesterday on the Bootcamp partition (Windows 8.1) i started Tessmark64 and PixMark Julia FP 64 at the same time and let them stress the m295x for 24 hours. I guess thats more interesting than a 30min-2h short time test.. here is the todays result:

Everthing seems ok, the stress test was still running - no restart, BSOD or sth else ;)

Same 107°C reached like ioxakoyba but obvioulsy not the heavy "throttling" point of 108°C (i know its a dynamic clock rate - no throttling in the original meaning) so i wouldnt have noticed any framerate drops in this specific run :)
 

Attachments

  • 24h stresstest.JPG
    24h stresstest.JPG
    308.3 KB · Views: 224
Last edited:

blufrog

macrumors member
Dec 19, 2014
97
10
Hi folks, i ve made a 24h stress test marathon. Yesterday on the Bootcamp partition (Windows 8.1) i started Tessmark64 and PixMark Julia FP 64 at the same time and let them stress the m295x for 24 hours. I guess thats more interesting than a 30min-2h short time test.. here is the todays result:

Everthing seems ok, the stress test was still running - no restart, BSOD or sth else ;)

Same 107°C reached like ioxakoyba but obvioulsy not the heavy "throttling" point of 108°C (i know its a dynamic clock rate - no throttling in the original meaning) so i wouldnt have noticed any framerate drops in this specific run :)

Wow!! Thank you!

I wonder if there is something about fabrication processes that mean it can sustain higher temps?

I know that lead-free solder has a higher melting point meaning that components are subjected to higher temps during soldering...I wonder if it means they changed the way the ICs themselves are constructed to survive this process, and as a side-effect they can now tolerate higher operating temps??
 

cynics

macrumors G4
Jan 8, 2012
11,959
2,154
Wow!! Thank you!

I wonder if there is something about fabrication processes that mean it can sustain higher temps?

I know that lead-free solder has a higher melting point meaning that components are subjected to higher temps during soldering...I wonder if it means they changed the way the ICs themselves are constructed to survive this process, and as a side-effect they can now tolerate higher operating temps??

Only time will tell.

There are people returning their AW15 w/ m295x because they see ~85-90c playing games. Lol. So it certainly runs hot but there isn't a lot of comparison out there currently. We'll be the first to find out how well it holds up with time.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.