Previous 1 2 Next 17 Replies Latest reply: Jan 9, 2014 5:38 AM by Pianoanna
scottpcs Level 1 Level 1 (0 points)

Hi Guys,

 

I'm currently on my 3rd replacement Retina MacBook Pro after having image retention (IR) issues with the screen and has been much-discussed in another thread.

 

This one is actually going into the Apple Store on Friday for a screen replacement but the subject of this thread is to do with the sheer amount of heat that this thing is putting out.

 

Of all the rMBPs that I've had so far, this one seems to run the hottest of them all, so much so that for the first time I've noticed just how hot it gets on the underside to the point of being uncomfortable.

 

I run Parallels in my day to day tasks and aside from this I just run iTunes music and Safari. This is enough to get the core CPU temp up to around mid-60s (Celcius) where it generally stays and hitting 70 / 71 degrees as soon as I do something such as go to compose an e-mail in Outlook for instance.

 

Without Parallels running i.e. just having regular Mac apps open, the core CPU temp drops to around mid-50s but again this can reach early-60s if doing anything remotely taxing on the CPU.

 

I don't particularly play games and to be honest isn't the reason I bought this MacBook but I do like to play Portal 2 and Half-Life 2 every now and again, which isn't (or shouldn't be?!) that demanding for a modern computer. Within a minute or two of playing either of these games, the CPU temp rockets to nearly 90 degrees before the fans kick in and bring that down to around early to mid 80s. GPU temp isn't far behind with these temps either.

 

As a final test, I opened up 8 x terminal windows and ran the following command:

 

yes > /dev/null

 

This apparently taxes all 8 CPU cores (4 x physical and 4 x virtual) to the max so that temperatures can be measured at full capacity. It didn't take long for all 8 cores to reach anywhere between 100 and 105 degrees during this test and I could've cooked dinner on the computer it was that hot. Obviously I didn't run this test for long due to the extreme temperatures, however, within a few moments the temperatures had come down to a more respectable level once I quit terminal.

 

The main reason for this thread is that this heat 'problem' is the last straw in my continuing struggle with Apple and the rMBP. Firstly it was the screen, then I noticed that the keyboard was not quite evenly backlit (the '2' key for instance is noticeably brighter than any of the other keys) and now I'm noticing how hot this thing is.

 

As this particular machine is my 3rd replacement, my only option really is to take it in for repair on Friday as scheduled for the screen and keyboard or demand a refund as I'm still within my 14 days I believe.

 

Trouble is, there isn't an equivalent machine to the rMBP if I got a refund as no other computer can run OSX but a Mac. Even the classic MBP isn't really an option as with the specs of this one (2.6 / 16 / 512) it would actually be more expensive!

 

I do wish I could just go out and buy some other computer but I'm hopelessly entwined with OSX and no other computer from any other manufacturer can run it, not officially anyway. This leaves me between a rock and a very hard place. I also need this machine for work so if I'm to get a refund, I need to have a replacement lined up .

 

Any thoughts at all? Is this machine a toaster in disguise?


MacBook Pro (Retina, Mid 2012), OS X Mountain Lion (10.8.2)
  • MarkRHolbrook Level 1 Level 1 (135 points)

    Well I have a Mid 2010 core i7 2.66 and I'm waiting for my retina to arrive.  All of these MBPs run hot in my opinion.  My 2010 gets unconfortable on my lap even when not CPU intensive.

     

    I expect the thinner MBP retina might even seem hotter but I would assume the chips inside are running about the same as my 2010.

     

    As an electrical engineer/sw engineer by trade I think it is safe to say that these MacBooks really should not be pushed to the max CPU because they simply don't have the cooling capacity to support that.  I think short to moderate length bursts of intense CPU activity will be fine.  But doing your test will probably overheat the machine if you let it go too long.

     

    I took my 2010 in to Best Buy the other day and I did four terminals with yes > /dev/null and my core i7 had all four cores at 100%.  I did the same, four terminals on a retina 2.3 core i7 and the four cores were at 50% and the machine was MUCH cooler than my 2010 after about 3 minutes.

     

    Clearly the machines were doing the same amount of work in four terminal windows but the retina was doing a better job of managing the CPU activity.  As you mention, I had to open additional terminal windows to get cores going at 100%.

     

    As a test on my 2010 MBP with four terminals doing yes I really can't watch a youtube movie.  It will become a little jittery in places.   I opened the same movie on the retina and it never even gasped. 

     

    But back on the subject of heat.  I think we can all do away with our toasters and just fire up our MBPs in the am to heat our coffee and crisp our toast!  LOL

  • scottpcs Level 1 Level 1 (0 points)

    As you say, it wouldn't do any computer any good by running the CPUs at maximum for any length of time. My particular test was only run for about a minute or two before ending it but in that time the core temperatures rose to an alarming level.

     

    Even when running at nowhere near capacity, however, it isn't unusual for the core CPU temps to rise to about 90 degrees celcius which is troubling.

     

    I never noticed such heat issues on any of the other replacements I've had, even when doing exactly the same things, including a couple of gaming sessions on said titles. I'd even go so far as to say I didn't notice any heat build up at all on my lap which leads me to think that this one has thermal issues.

     

    Regardless of any of this, I'm ringing Apple now to initiate a full refund. I've had too many issues with this first generation of retina MacBook Pro and I'm going back to my old workhorse for now, a 2010 13" MBP and wait for the second generation instead.

  • dmdimon Level 3 Level 3 (840 points)

    I run Parallels in my day to day tasks and aside from this I just run iTunes music and Safari. This is enough to get the core CPU temp up to around mid-60s (Celcius) where it generally stays and hitting 70 / 71 degrees as soon as I do something such as go to compose an e-mail in Outlook for instance.

     

    Without Parallels running i.e. just having regular Mac apps open, the core CPU temp drops to around mid-50s but again this can reach early-60s if doing anything remotely taxing on the CPU.

    This is normal temperatures

    I do like to play Portal 2 and Half-Life 2 every now and again, which isn't (or shouldn't be?!) that demanding for a modern computer. Within a minute or two of playing either of these games, the CPU temp rockets to nearly 90 degrees before the fans kick in and bring that down to around early to mid 80s.

    Most Mac games are just Windows ones, running in virtualising environment like Cider,

    thus demanding much more resources than native Windows versions. Thus Windows in bootcamp is much preferred for gaming. Especially with incredible Retina resolution.

     

    Heat goes from CPU and GPU proportionally to the taskload - so there are no escape, any notebook with same power will give you more or less same amount of heat despite of manufacturer/os. Only heat may be hidden from user behind plastic instead of aluminium - but it will be there.

     

    and I'm not sure that MarkRHolbrookis inevitably right on not pushing to constant 100% load. I used to run Mathlab routines almost at full 100 for hours, some media encoders at 100% for hours, some renders - and my 2011 MBP still feels good.

    It goes really hot, but runs and runs and runs. Actually I avoiding this as far as I can (thinking same as Mark) - but there are situation when I need all the power - and it just works.

     

     

    Almost forgot - Core CPUs will go overclock under high load - like from 2.2 to 2.8 GHz - it happens with a lot of yes > /dev/null's definitely. Thus heat dissipation goes high. Cider also tend to use just one core up to max - thus forcing it to overclock. Having some percentage free gives possibility to less overclock and dissipate less energy.

  • MarkRHolbrook Level 1 Level 1 (135 points)

    I'm not interested in being right or wrong but the fact remains that heat destroys components.  That is why we have heatsinks and fans on our electronics.  If for some reason the heatsink/fan system in the MBP can't keep up with the heat generated then it is possible to do damage.  In my electrical career I've fried many a chip pushing it too hard.  When they go, they go in a big way.  IE total failure.

     

    I too run VMWare fusion.  One of my main purposes for getting a retina is not the screen but the 16gb of RAM.  Obviously more screen real estate will be helpful but I'm probably not going to be as bothered as the guys in the burn-in forum if I have some IR. 

     

    My typical work day consists of three VMs running:  Windows 7 32, Windows 7 64, and a Windows 8 VM.  Each is running a compiler/editor so they are not churning away on rendering an intense graphic.  But each is eating up some CPU and at times when Windows decides it needs to get busy my MBP 2010 gets hot fast and the CPU goes up rapidly.  I try to be proactive and set my fans to run faster when I expect to do this.  I'd rather have the noise than the heat I guess.

     

    Is it a good idea to push your MBP at 100% on all four cores for hours on end?  I really can't say.  One would hope the Apple engineers did their jobs and made sure that there was no way injure things doing that.  But if they didn't then you are running the risk of your own personal meltdown so to speak!

     

    Back when I used a Windows laptop I had a toshiba and one time I started up a Window sprocess to scan the disk for viruses while I was watching a movie in the hotel on a trip.  I noticed it was getting hot but figured it would be ok.  About 30 minutes later the screen screwed up then went blank and I smelled this awful smell which electrical engineers hate... The burned chip smell.  The laptop never ran after that.  Toshiba offered a replacement motherboard at almost the cost I could buy a brand new machine for.  I lost almost a weeks work time sitting onsite without a computer.   Since that experience I just don't push things that hard.

     

    If I really need that much CPU I try to offload it on to my desktop (MacPro dual Xeon).

  • dmdimon Level 3 Level 3 (840 points)

    I'm not interested in being right or wrong but the fact remains that heat destroys components.  That is why we have heatsinks and fans on our electronics.  If for some reason the heatsink/fan system in the MBP can't keep up with the heat generated then it is possible to do damage.  In my electrical career I've fried many a chip pushing it too hard.  When they go, they go in a big way.  IE total failure.

    Absolutely with you on this.

     

    Just it happens that I periodically use software that loads my MBP at full for hours. It happens that my MBP does not have cooling system manufacturing defects.

     

    May be for OP it will be better to run hour or two at 100% just to find cooling problems (probably frying his MBP) and exchange it for good one - or just to know that all is ok at any load? It is a bit scary and unhuman proposition, I know, but what if problem with cooling quality exists on his MBP?

     

    It have to survive full load for hours - as there are a lot of math-intensive multithreaded Mac software that can load any computer up to 100%. What if OP will use one of that in future? Apple propagates multithread everywhere - so there will be more and more of this... Take a look at loads when iMovie does encoding. When Aperture works. And so on.

     

    I know smell you're writing about - but I can't control software I use. It just grabs everything it can.

  • scottpcs Level 1 Level 1 (0 points)

    OK so I've just got off the phone to Apple UK.

     

    After seeing the number of notes on the system and sympathising with all the problems so far, including the heat issue, they've offered a final replacement.

     

    The rep did say that I would still have the option of a repair or full refund if this one turns out to be unsatisfactory so at least my options are kept open at this moment in time. Let's see what the 4th replacement has to offer!

     

    I would imagine that the cooling system has to be able to cope with the system running at full load for a considerable length of time, otherwise it's not fit for purpose. At the same time, however, this is why CPUs manufactured in recent years have had the throttling and thermal shutdown features which are both designed to prevent damage in the case of excessive heat. With this in mind it should be near impossible to fry a modern laptop to the point that it destroys itself but of course theory is a wonderful thing.

     

    In response to dmdimon, I don't believe the temperatures that I'm seeing in normal use are normal at all for this particular machine. The other 3 laptops I've had didn't get nearly as hot when doing the same tasks and weren't even noticeably warm to the touch either. It's only when I felt my lap getting particularly warm with this one that I realised something wasn't quite right and the temperature readings seem to confirm this.

     

    Whilst not exactly a scientific experiment, I played through the same level on both Half-Life 2 and Portal 2 on each of these machines to see how warm they got and also to see if the fans kicked in. Only the macbook I currently have warmed up to the point of being uncomfortable and also is the only machine so far to have invoked the fans due to high temperatures.

     

    Like I say, not a scientific experiment by any stretch of the imagination but it gives me some kind of yardstick to work with. Up until now, heat hasn't even been a problem with the other replacements I returned, I just appear to have received a Friday special in this instance.

  • dmdimon Level 3 Level 3 (840 points)

    In response to dmdimon, I don't believe the temperatures that I'm seeing in normal use are normal at all for this particular machine. The other 3 laptops I've had didn't get nearly as hot when doing the same tasks and weren't even noticeably warm to the touch either.

    there was reports on bad assembly of cooling system in MBP's. May be this is your case. Just from my experience - 70 C is definitely nothing unusual when computer works. If you're interested - can check temperatures at full load when at home, in 2-3 hours.

  • scottpcs Level 1 Level 1 (0 points)

    dmdimon wrote:

     

    there was reports on bad assembly of cooling system in MBP's. May be this is your case. Just from my experience - 70 C is definitely nothing unusual when computer works. If you're interested - can check temperatures at full load when at home, in 2-3 hours.

    That would be useful to know and thanks for the offer.

     

    I'm willing to put this machine down to a bad day at the factory and I hope I'm right. None of the other replacements I've received had this issue, at least not that I've noticed.

     

    There isn't a whole lot of info on the internet about what's considered normal temperatures for the rMBP so it's difficult to know what's normal and what isn't, particularly at extreme ends of the spectrum. Terminology differs from site to site too and some people work on core CPU temperatures which are usually very different from the CPU proximity temperature. All I know is that this machine gets noticeably hot and the others didn't, hence this thread.

  • dmdimon Level 3 Level 3 (840 points)

    acceptable temperatures for core and GPU cores are listed in Intel, AMD and nVidia specs for every model and are easy to find if you're interested.

    for example see answer from Intel representative here:

    http://communities.intel.com/thread/29229

  • chronomane Level 1 Level 1 (0 points)

    As I read once in a while some of the rMBP related threads I come aware, that I must be an outstanding lucky guy, because my machine is picture perfect. The screen, a samsung, is perfect. No dead pixels, mura or light bleeding. This is the coolest machine ever. For me. In simple use, like safari or iTunes, I never gets above 35 degrees Celsius. When working more demanding like with Aperture or HD videos they get to max 50 degrees. Only if I really kick the thing, it might goes up to 60 0r 65. But never more. And at that point, the fans really don't kick in noticably. So, how come, that obviously so many rMBP's are not working right? Cause it can't be the design. Why would mine be so wonderful?

    Apple, do something!!!

    It can't be right, that people spending a fortune on this thing, and than are only struggeling.

  • dmdimon Level 3 Level 3 (840 points)

    here we go:

    CPU Die temps in Celsius reported by iStat

     

    lazy browsing - 1% load - CPU Die 48 C

    4 x "yes > /dev/null " - 51% load - CPU Die 88-96 C in 30 seconds and keeps at it

    8 x "yes > /dev/null " - 100% load - CPU Die 90-96 C in 30 seconds and keeps at it

     

    I'd say that cooling system is adequate - as with load from 50 to 100 % it keeps stable high temp. And as you see - from thermal point there are no difference between 100% and 50% load.

  • scottpcs Level 1 Level 1 (0 points)

    dmdimon wrote:

     

    here we go:

    CPU Die temps in Celsius reported by iStat

     

    lazy browsing - 1% load - CPU Die 48 C

    4 x "yes > /dev/null " - 51% load - CPU Die 88-96 C in 30 seconds and keeps at it

    8 x "yes > /dev/null " - 100% load - CPU Die 90-96 C in 30 seconds and keeps at it

     

    I'd say that cooling system is adequate - as with load from 50 to 100 % it keeps stable high temp. And as you see - from thermal point there are no difference between 100% and 50% load.

    Many thanks for that, much appreciated.

     

    Looking at those figures I'd say mine definitely has poor thermals. Even at 100% load, your machine produces temperatures that are 9 - 10 C cooler than mine at both maxium and minimum values.

     

    With 'lazy browsing' as you call it, mine produces temps in the 55 - 60 C range and that's with nothing else running i.e. ONLY safari.

     

    This machine is going back anyway for a number of reasons but at least I've got a yardstick to measure by when the replacement turns up.

     

    I had a look at Intel's specs for this processor but it doesn't really give me the information that I'm looking for. The max operating temp is listed as 105 C and the CPU in this machine runs extremely close to that at full load, if not at the limit. At the same time, it seems to be accepted that Ivy Bridge is a hot running CPU, certainly hotter than Sandy Bridge that it replaced. Not sure I'm comfortable with the chip running so close to its limit though, even during normal usage albeit demanding usage at that.

  • maratus911 Level 1 Level 1 (10 points)

    Dear dmdimon:

     

    Could you please specify which processor do you have in your RMBP? It would be also helpful to indicate stepping (you can use MacCPUID for that) and approximate purchase time. I'd really appreciare if you provide those details.

  • dmdimon Level 3 Level 3 (840 points)

    Could you please specify which processor do you have in your RMBP?

    sorry but you misread me:

    my 2011 MBP still feels good

    but if you're interested I still can take a look at stepping when at home.

Previous 1 2 Next