Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Retina Bootcamp - CPU Speed , HIGH Temps, etc
- Thread starter OSMac
- Start date
- Sort by reaction score
I have a Xeon X3440 2.53 Ghz quad core CPU in my self built Windows PC. It's essentially a hyper threaded i5. Overclocked to 3.4 Ghz, and it's not uncommon for it to hit 97-100C under load very quickly. From what I have read, the i5 and i7 series get into the high 90's quickly. They are just hot running CPUs.
Heat has been an issue with Apple notebooks for a long, long time. Especially the high-performance metal ones (Titanium PowerBook G4 anybody?).
Apple prioritizes low noise over low heat, and has calibrated the fan ramps to that ethos. Thus the SMC waits until the temps are quite high before ramping the fans. The hope is that the heavy load will be short, so the fans don't have to spin up at all. For many casual users that is a safe assumption.
Users such as myself and probably many others here, this is not a good assumption. tjmax for most current Intel notebook chips is 105C, the CPU should operate fine at 104C indefinitely, and still be usable for its full lifetime. However, above 90C is uncomfortable from my point of view.
Using SMC Fan Controller for OS X you can manually ramp up before a game, a render, an encode, et cetera. Placing the fans at maximum speed will generally keep the temps at around 85C in my previous experience. We'll see how these mid-2012 MacBook Pros do.
Apple prioritizes low noise over low heat, and has calibrated the fan ramps to that ethos. Thus the SMC waits until the temps are quite high before ramping the fans. The hope is that the heavy load will be short, so the fans don't have to spin up at all. For many casual users that is a safe assumption.
Users such as myself and probably many others here, this is not a good assumption. tjmax for most current Intel notebook chips is 105C, the CPU should operate fine at 104C indefinitely, and still be usable for its full lifetime. However, above 90C is uncomfortable from my point of view.
Using SMC Fan Controller for OS X you can manually ramp up before a game, a render, an encode, et cetera. Placing the fans at maximum speed will generally keep the temps at around 85C in my previous experience. We'll see how these mid-2012 MacBook Pros do.
I remember reading a post somewhere that suggested if you set up SMCFanControl with the speeds you want, and then restart (as opposed to shutting down and then starting up with the power button) into BootCamp, the fan temp/RPM profile will remain loaded, and work even in Windows.
I've never used SMCFanControl, so I don't know for sure if it works or not.
And heeeeeeeeeere come the fanboy apologists. Nothing to see here, folks. Move along. Nothing to see.
I really prefer running windows in vmware fusion, the track pad works much better than trying to use the boot camp drivers.
joke as you like, macs do run hotter in bootcamp. Something regarding power saving or fan profiles or some other thing, but they indeed do run hotter.
I've been using Input Remapped as a dedicated fan control for older late-2009 15.4" Pro and few other models. They're not quad-core, but still quite hot guys.
Lubbo's is another option.
It's very important not to rely on automatic fan speed adjustment in Windows. As long as you run both fans at their max RPM, laptop should stay reasonably hot but it won't overheat even when gaming for few hours. Regular 3D game or benchmark is unable to stress CPU as much as Prime95, but if you take into account the GPU load and the design of cooling system, it makes more sense.
As far as RMBP goes, it should have even more potential, since:
1) It has slightly longer radiators that 2011 MBP which equals to larger cooling surface
2) Higher total number of fins (108 vs. 97)
3) Potentially higher performance fans
4) Larger gap between screen hinge cover and exhaust vents will result in less air restriction and higher total airflow. Turbine fans provide a lot of exhaust pressure, so it's not that big deal here, but it's certainly helps.
5) Two additional holes on both sides serve the same purpose. In traditional design the main and only air intake area is the middle part of the hinge vent hole. That causes the following issues:
A. Higher airflow restriction due to the hinge cover being very close to the vent holes now is applied not only to exhaust but to the intake as well - which results in leven lower airflow and therefore worse cooling performance. And since it's an intake flow, it's more important.
B. Higher amount of hot air from exhaust fans recirculated due to the hinge proximity (and it's the only intake!), which results in higher air intake temperature - worse cooling performance.
Both issues are addressed in RMBP by 4) and 5)
Lubbo's is another option.
It's very important not to rely on automatic fan speed adjustment in Windows. As long as you run both fans at their max RPM, laptop should stay reasonably hot but it won't overheat even when gaming for few hours. Regular 3D game or benchmark is unable to stress CPU as much as Prime95, but if you take into account the GPU load and the design of cooling system, it makes more sense.
As far as RMBP goes, it should have even more potential, since:
1) It has slightly longer radiators that 2011 MBP which equals to larger cooling surface
2) Higher total number of fins (108 vs. 97)
3) Potentially higher performance fans
4) Larger gap between screen hinge cover and exhaust vents will result in less air restriction and higher total airflow. Turbine fans provide a lot of exhaust pressure, so it's not that big deal here, but it's certainly helps.
5) Two additional holes on both sides serve the same purpose. In traditional design the main and only air intake area is the middle part of the hinge vent hole. That causes the following issues:
A. Higher airflow restriction due to the hinge cover being very close to the vent holes now is applied not only to exhaust but to the intake as well - which results in leven lower airflow and therefore worse cooling performance. And since it's an intake flow, it's more important.
B. Higher amount of hot air from exhaust fans recirculated due to the hinge proximity (and it's the only intake!), which results in higher air intake temperature - worse cooling performance.
Both issues are addressed in RMBP by 4) and 5)
I'm aware of that. Above 100C it will throttle back.
Check the sig - I've got a 2011 MBP 15". It has boot camp installed and I've seen 95C temps in Windows 7 while gaming on it. I wasn't appalled.
Whatever, if I wanted to monitor my fan speeds and tinker with a bunch of crap then I'd just buy a Windows PC.
This is why I purchase AppleCare and do hourly Time Machine backups. If the thing gets fried I'll have Apple replace it. I refuse to tinker with a bunch of grognard settings on a Mac. What's the point of having a Mac then?
This is why I purchase AppleCare and do hourly Time Machine backups. If the thing gets fried I'll have Apple replace it. I refuse to tinker with a bunch of grognard settings on a Mac. What's the point of having a Mac then?
Thanks mate, I've got my retina macbook on order, but I was just curious if someone could confirm firsthand that it is compatible with lubbos's and if so the temps under load while running it.
I've seen lots of people recommend trying it but not seem anyone confirm conclusively that it works with the new rMBP.
I ordered my rMBP but it is still processing
, i will try it for sure one it arrives. But i used that fantastic app on my last 2011 17" mbp and it worked very well, it was impossible to use bootcamp without lubbo's fan control.p.s: sorry for my english
I ordered my rMBP but it is still processing
p.s: sorry for my english
Your English is fine mate.. Mine is n order too, delivery due July 9th, if it arrives before yours I will try lubbo's and report back...
Lubbos fan control does "work" for the rMBP.. it always detects the cpu temperature as 90C, so the fans will stay revved up.. and by default it only revs up one fan to full and keeps the other at 3000 rpm.. but you can still max them out manually.
The only real downside that I've found is that you have to shut the machine off to have the fans go back to "normal", even if you exit Lubbo's fan control.
It works in a pinch, though..
Thanks mate, thats good to know...
Do you know how warm does the cpu gets under load with Lubbo's maxed out? Is the fan noise distracting?
It is not silent.. I find the noise much less annoying than the early 2011 macbook pro, though.
I haven't tried it under load with both fans maxed out.
And it does provide a meaningful reduction in cpu temps Im assuming?
Sorry just a little paranoid about frying my new machine lol
And it does provide a meaningful reduction in cpu temps Im assuming?
Sorry just a little paranoid about frying my new machine lol
When the fans finally spun up on their own (before I got lubbo's) the fans brought the temps down to around 90 instead of breaking 100 running 8 threads of prime 95.
I did the maximum heat test in p95.. and it did throttle the cpu before the fans kicked in on their own.
I'll definitely be using lubbo's before I do anything that will need more cooling.
I've got a question/observations for you guys. I've been playing BF3 w/ Windows 8 Release Preview bootcamped, which has been great. Anyway, I was curious about the temps, so I used the program RealTemp to see how high the CPU was going. RealTemp reported that my temps were hitting and staying around 102 degrees Celsius. This was a little alarming so I downloaded a different program just to see if I got different results. With CPUID Hardware monitor (I played in windowed mode so I could see each movement of the temp) I only saw the CPU temp reach about 80c and settle at 60-70c when the fans kicked in.
Obviously the second result is a lot less alarming than the first, but I'm wondering which is more realistic and which monitoring software should be trusted?
*I gamed for about 25 min in each test, and I'm playing in an unairconditioned room, so the temp is about 80F.
Obviously the second result is a lot less alarming than the first, but I'm wondering which is more realistic and which monitoring software should be trusted?
*I gamed for about 25 min in each test, and I'm playing in an unairconditioned room, so the temp is about 80F.
Run Prime95 (heat test) and compare CPU temps in real time, also can use HWInfo as a throttling detector (each core will throttle when hits 105C)