Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
So you are pushing the CPU to it's limits and are then surprised it throttles. May I ask what you expected it to do instead? Continue at max clock speed and burn itself into oblivion? That doesn't sound too good for me:rolleyes:
Every single CPU is doing this since a few years already. It can happen with a desktop cpu as well if you push it beyond your coolers limit. Since this is a fanless laptop it is just easier to do in the MB. You can see this with all the Core-M laptops released and it is expected. The thing the OEMs can adjust is how long it can go at full speed until it hits the thermal wall and that is dependent on the CPU used, the thermal management in the laptop design and the target temperature that the OEM sets.
 
The chips in use have publicly available data sheets and it's easy enough to see the 1.1 runs at a higher wattage where the 1.2 is .5-1w less. But when my 1.3 arrives I'll be doing some extensive testing.

Thing is there are a couple of throttling threads now and people are under some dumb founded assumption that a cpu at 100% usage is consistant. It isn't and that's where modern cpu's work well. A millisecond unloaded the cpu will clock down below its base clock. This could be seen as throttling but it really isn't.

Thanks. I have had the Intel Power Gadget and its Windows counterpart installed on my laptops for a long time but haven't really paid a lot of attention, so my ability to compare the Core M is not great, but I have the impression that the Core M in my 1.2/512 is much more nimble, ramping up and down extremely quickly on all three monitored readings.
 
"just like the MBA did when it throttled. "

Let me state what he posted again

"just like the MBA..."

That's MBA for Macbook Air!

So what he is saying is this, if you want a bigger CPU then don't buy either the Macbook Air or the 12" Macbook.

Also running the VM view is a big deal on any computer!
 
Lots of interestingly useless dialogue in this thread. "It's throttling" is a funny notion here considering the screenshots. The OP has icons in the menu bar almost across the entire screen, has a ton of applications open, and to top it all off is running a Virtual Machine that is running and also updating the OS which is very intensive work!

The computer is fanless. Gee I wonder why it's throttling?

This workload would be big for any computer and one with a fan would likely be running at full fan speed for the virtual machine alone. If it didn't throttle, the computer would be dead.

This post amounts to "no duh." 99% of users would never do any of the things shown on screen.

Tons of apps? Wow.. thats funny. Take a closer look.

The point here is that if we are truly at the point where fanless computers are a USABLE option, then it shouldn't immediately throttle within under a minute. The case should have quite a bit of heat soak potential before hitting saturation. I have a few other fanless tablet/combos running Atom CPU's that don't experience this.

I'm getting fanboi'd. I'm a mac lover, but won't let myself not see this issue when it is there.
 
Buying a New Macbook and complaining that it throttles is like buying a new motorcycle and complaining you get wet when it rains. Throttling is expected, how else will a passive CPU regulate it's temperature? Just enjoy it, and as others have said, if you do more CPU intensive tasks for long periods then get the Macbook Pro Retina 13".

It shouldn't throttle immediately. I am going to test on OS X only apps and try to capture when this happens more accurately. I also have an IR Temp sensor and can sample how wide the heat soak is. Obviously throttling is the ONLY way to prevent an overheat, but it should tolerate full speed for at least a reasonable time period. Reasonable is the key word. Should it be for under a minute? Over a minute? 5 minutes? I'm going to find out.

I think that this needs to start being a metric on fanless designs. We are seeing benchmarks that show the rMBP to run at about a 2011 MBA speed, but it can not sustain it, so its actually quite a bit less capable than a 2011 MBA.
 
Last edited:
netbook behaving like a netbook.

News at 11.

****ing expensive netbook.

No, Core M is not a "netbook" quality CPU by any stretch of the imagination.

I remember the purgatory that was inflicted on the human race when netbooks were proliferating faster than rabbits breeding on a sunny afternoon in the English countryside.

They were a curse bestowed upon PC users the world over.

The new MacBook should not be lumped into the same category. Let's hope OP comes right.
 
It shouldn't throttle immediately. I am going to test on OS X only apps and try to capture when this happens more accurately. I also have an IR Temp sensor and can sample how wide the heat soak is. Obviously throttling is the ONLY way to prevent an overheat, but it should tolerate full speed for at least a reasonable time period. Reasonable is the key word. Should it be for under a minute? Over a minute? 5 minutes? I'm going to find out.

I think that this needs to start being a metric on fanless designs. We are seeing benchmarks that show the rMBP to run at about a 2011 MBA speed, but it can not sustain it, so its actually quite a bit less capable that a 2011 MBA.

It is also surprising to see it throttling at such a low temp. 65c is hardly a temp that warrants throttling. You could easily push intel processors to 90c before needing to worry.

My guess is they are being extremely cautious on temps due to the fanless design. The last thing that Apple needs is a bunch of people bringing in burnt out laptops under warranty.

The fact that the cpu runs at 60% under throttling is pretty horrible.
 
I think it is important to understand what this device is targeted at and what it isn't. Yes you can run any application on this that you would on a MBP but that doesn't mean you should expect the same performance. This device is meant for light loads on average and short bursts of heavy loading.

I agree, and the only way I can determine if this limitation is acceptable to my needs is to test. I am then sharing those findings here to help everyone. People (not you) need to stop taking offense to that and assuming I'm a moron for not understanding the user target.
 
It shouldn't throttle immediately. I am going to test on OS X only apps and try to capture when this happens more accurately. I also have an IR Temp sensor and can sample how wide the heat soak is. Obviously throttling is the ONLY way to prevent an overheat, but it should tolerate full speed for at least a reasonable time period. Reasonable is the key word. Should it be for under a minute? Over a minute? 5 minutes? I'm going to find out.

I think that this needs to start being a metric on fanless designs. We are seeing benchmarks that show the rMBP to run at about a 2011 MBA speed, but it can not sustain it, so its actually quite a bit less capable than a 2011 MBA.

It's not about what you find with your IR sensor. It's about what is happening at the die. That die will heat up very quickly under heavy load. Unless you can change the laws of physics all of this data collection is futile. Do you think Apple hasn't modeled the thermals of this to the nth degree? I'm sure the throttle set points are also there to keep the aluminum temps below a certain threshold. Assume they can let it run long before throttling. That heat is going to the body of the laptop. You will feel it.

Throttling is necessary within the boundaries of the 12" chassis.
 
It's not about what you find with your IR sensor. It's about what is happening at the die. That die will heat up very quickly under heavy load. Unless you can change the laws of physics all of this data collection is futile. Do you think Apple hasn't modeled the thermals of this to the nth degree? I'm sure the throttle set points are also there to keep the aluminum temps below a certain threshold. Assume they can let it run long before throttling. That heat is going to the body of the laptop. You will feel it.

Throttling is necessary within the boundaries of the 12" chassis.

True, but wouldn't you think that Apple would also design a way to allow the unit to work for a "reasonable" period before exhausting the passive heat dissipation abilities of the system? I also noticed that simply blowing (with my mouth) on the bottom of the case near the board area, the CPU speed increased quite immediately. This means that the external thermals of the case does quickly impact the CPU throttling and that the throttling is very responsive to thermal conditions. This goes against your thought that the outer part of the case isn't a good sampling point.
 
Does is still exhibit throttling after your VM has been installed (and you have the VM drivers)? I don't find it particularly surprising that the CPU will be pushed during the installation, but it should be more or less idle once the Windows VM is up and running normally.
 
Apple could have added more mass to the chassis but then you wouldn't have a nice thin small laptop. Of course blowing air over the bottom (aka heat sink) shows a change. What that tells me is that an more delay in throttling would make for a very hot base.

It is what it is. Apple can't change physics. They can only change how much metal is there.
 
Tons of apps? Wow.. thats funny. Take a closer look.

The point here is that if we are truly at the point where fanless computers are a USABLE option, then it shouldn't immediately throttle within under a minute. The case should have quite a bit of heat soak potential before hitting saturation. I have a few other fanless tablet/combos running Atom CPU's that don't experience this.

I'm getting fanboi'd. I'm a mac lover, but won't let myself not see this issue when it is there.

The bigger point is Windows Update is running on a virtual machine. As I said in my post, that action ALONE would make any other Macbook run at full fan speed, and even some iMacs!

The "ton of other apps" only contributes to an already hostile measurement.

And yes it would absolutely throttle in a minute. Why is this surprising? It's a Core M chip designed to burst at high speeds for performance. You are doing a sustained and large workload in a virtual machine. That is not typical performance and additionally not ideal for a Core M chip.

A tiny bit of thought makes this entire thread a useless endeavor in arguing a completely expected and unsurprising result.

Don't be Agnes Flanders. The bag can't hold all your groceries AND not be heavy. Consider me the possible police.
 
Heres some food for thought... I was on the fence getting the MBR, ultimately for me it was too small for "me". I'm holding off for a 14" version hopefully. But I'm still getting one for my wife down the road.

I used to have a SP3 i5 version with the 4GB of ram, great little machine. I read TONS of articles about the throttling issues and people putting little fans behind them to game :roll eyes:. Anyway, i never had a problem with mine, sure it got a little warm, or the fan would pop on when i was downloading something intensive. But untimely it served me well, but i just wasn't into the hybrid thing i wanted a full solid non flip machine.

So of course if you push it to the limit its going to throttle, isn't the base 1.1 model actually a 900mhz but overclocked to a 1.1 by Apple, I'm pretty sure thats what i read. Either way if you don't do intensive tasks that you would probably need a Macbook Pro for you'll be fine i think. Im not sure even running windows on it is a good idea. Windows is far more demanding compared to OSX when i comes down to the specs you need. No?
 
Heres some food for thought... I was on the fence getting the MBR, ultimately for me it was too small for "me". I'm holding off for a 14" version hopefully. But I'm still getting one for my wife down the road.

I used to have a SP3 i5 version with the 4GB of ram, great little machine. I read TONS of articles about the throttling issues and people putting little fans behind them to game :roll eyes:. Anyway, i never had a problem with mine, sure it got a little warm, or the fan would pop on when i was downloading something intensive. But untimely it served me well, but i just wasn't into the hybrid thing i wanted a full solid non flip machine.

So of course if you push it to the limit its going to throttle, isn't the base 1.1 model actually a 900mhz but overclocked to a 1.1 by Apple, I'm pretty sure thats what i read. Either way if you don't do intensive tasks that you would probably need a Macbook Pro for you'll be fine i think. Im not sure even running windows on it is a good idea. Windows is far more demanding compared to OSX when i comes down to the specs you need. No?

My testing is to find where those limits are so I can set expectations. I already own two 15" rMBP, 13" rMBP, and an 11" MBA. My purchase intent was not a do-all computer, but I do need it to do basic tasks, which to me is run a vm idle to do light windows-only asks. No review tested this particular use case, so I have to. If those limits are below my needs, I will return it.
 
Last edited:
65-70C is nothing. Guess you guys missed my thread where I was getting temps of 84-88C and all I was doing was exporting a 5 minute 1080p video in iMovie.
 
My testing is to find where those limits are so I can set expectations. I already own two 15" rMBP, 13" rMBP, and an 11" MBA. My purchase intent was not a do-all computer, but I do need it to do basic tasks, which to me is run a vm idle to do light windows-only asks. No review tested this particular use case, so I have to. If those limits are below my needs, I will return it.

If I had to deal with 5 computers I would want one to burn up and take me with it. How do you manage 5 laptops? Are they all for personal use?
 
So did we come to a conclusion if Fusion will be better than Parallels? or use bootcamp instead?
 
Due to the amount of batteries the chassis houses, heat dissipation might be a bit more difficult than one would think.

After all, the logic-board is quite small, so the heat concentrates around a small area...

Install and patch W8 somewhere else, import it into Fusion and create linked clones of it...
That should set it.
 
Due to the amount of batteries the chassis houses, heat dissipation might be a bit more difficult than one would think.

After all, the logic-board is quite small, so the heat concentrates around a small area...

Install and patch W8 somewhere else, import it into Fusion and create linked clones of it...
That should set it.
There is a heat transfer plate on the cpu that pushes the heat through the base plate
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.