MBP Overheating

[Magickiller9]

All i did was typing a report on pages and surfing hardwarezone. Using it in an aircon room on the desk and it's gotten so hot =/.

Macbook Pro 2012 mid 15" with 256GB SSD and 16GB ram. What to do, cooler fan doesn't cut it for me because i can't bring the fan to school.

 

[PetPet]

Restart your Macbook. (Most of the time it's flash plugin consuming an amazing number of CPU cycles)
Try again and see if the heat issues persist.

if it does, call up AppleCare and let them have a look.
It may be a faulty fan...

 

[Magickiller9]

i hate Apple Singapore =/. their service sucks lei.

not fan faulty for sure, open up and check liao.

 

[davidktw]

i hate Apple Singapore =/. their service sucks lei.

not fan faulty for sure, open up and check liao.

Actually it's quite normal to feel hot for MBP. Using several series of MBP for more than 7 years, I believe the reason why you feel MBP is relatively hotter than other laptops is because of the good conductor of heat by the aluminium casing.

The palm rest area will feel slightly harm, but that's alright. The vent areas will feel pretty warm, near hot, but it shouldn't be of great concern.

Previously I use application like coolbook, but since Lion no longer support it, I don't bother either and remove it. Back with the usual heat I'm getting, i don't care much either. Of course you should take note if there is any rogue processes causing high CPU.

Any case, MBP is not made of tofu. All intel chips have built in thermal shutdown mechanism, so there is no need to worry and just use as-is.

 

[IronMac]

I have a 2010 i7 MBP running 10.6.8. Right now I have eight programs open with Chrome open in about 15 windows for a total of about 100 tabs. System is running very warm but not dangerously so.

I have a couple of programs called FreeMemory and MiniUsage running to monitor my memory and CPU usage. If my memory falls low enough I can use FreeMemory to free up some and if my CPU usage goes up too much then it's time for me to look at Activity Monitor to see what is going on. Usually the problem is Flash.

 

[davidktw]

I have a 2010 i7 MBP running 10.6.8. Right now I have eight programs open with Chrome open in about 15 windows for a total of about 100 tabs. System is running very warm but not dangerously so.

I have a couple of programs called FreeMemory and MiniUsage running to monitor my memory and CPU usage. If my memory falls low enough I can use FreeMemory to free up some and if my CPU usage goes up too much then it's time for me to look at Activity Monitor to see what is going on. Usually the problem is Flash.

There is no need to run any free memory applications. They are not helpful to the performance of the system. No need to use them. I suggest you might want to read up on how memory are utilised in a unix environment

 

[MoDiE]

i use smc fan controller to force the fan to run full speed 5800 rpm when connected to power source.. when on batteries, fall back to 4200 rpm

 

[IronMac]

There is no need to run any free memory applications. They are not helpful to the performance of the system. No need to use them. I suggest you might want to read up on how memory are utilised in a unix environment

Are you talking about RAM or HDD space?

I would beg to differ as to how useful FreeMemory is, it's worked for me and for a few other users.

 

[negativzero]

Are you talking about RAM or HDD space?

I would beg to differ as to how useful FreeMemory is, it's worked for me and for a few other users.

Apps that free up memory are the most useless things though and the effect is mainly placebo and a legacy myth taken up by windows users who switch to linux.

unix OSes cache their files in ram so that your computer is fast. so if you free your ram, you're actually slowing down your system at the cost of speeding up your apps.

the only time freeing up ram is useful is when you're paging out loads, otherwise, just let the OS take care of itself.

 

[Rock-kun]

Apps that free up memory are the most useless things though and the effect is mainly placebo and a legacy myth taken up by windows users who switch to linux.

unix OSes cache their files in ram so that your computer is fast. so if you free your ram, you're actually slowing down your system at the cost of speeding up your apps.

the only time freeing up ram is useful is when you're paging out loads, otherwise, just let the OS take care of itself.

I call nonsense

Windows caches and purges files and applications in RAM much more efficiently than any Unix-based operating system when it comes to a desktop-based environment. Take a good hard look at how Windows 8 manages this.

 

[Reborns]

now i am worried,..
cuz i use my Mac on clamshell mode...

so i am not sure it will be overheating

got myself Twelve South BookArc
not sure if it will help by elevating it....

 

[davidktw]

Are you talking about RAM or HDD space?

I would beg to differ as to how useful FreeMemory is, it's worked for me and for a few other users.

If you have work with *nixes like the way I do, perhaps I would just believe you on how you beg to differ So may I enquire for your reason why FreeMemory works ? How do you tell that it is working for you ?

As I know FreeMemory is freeing up the RAM, not the hard drive space

 

[davidktw]

I call nonsense

Windows caches and purges files and applications in RAM much more efficiently than any Unix-based operating system when it comes to a desktop-based environment. Take a good hard look at how Windows 8 manages this.

If this is a debate regarding which platform manage file buffers and RAM more efficiently, it has probably digress. However Rock-kun, I'm interested to know how you conclude Windows 8 does it better ? I'm not saying it isn't, the question is how you tell it's better than any unices ?

 

[davidktw]

Apps that free up memory are the most useless things though and the effect is mainly placebo and a legacy myth taken up by windows users who switch to linux.

unix OSes cache their files in ram so that your computer is fast. so if you free your ram, you're actually slowing down your system at the cost of speeding up your apps.

the only time freeing up ram is useful is when you're paging out loads, otherwise, just let the OS take care of itself.

Freeing up the ram by 3rd party services are never useful even in situation when your system is thrashing. The reason why it thrash is due to insufficient memory to accommodate the working memory set, freeing up is just taking out the cache buffer which actually deteriorate the system ram further than it should be.

Modern OS have mechanism built into the kernel to evict file buffer as priority during memory pressure, introducing a 3rd party application to manage the memory basically mess up the algorithm of the kernel

 

[negativzero]

Freeing up the ram by 3rd party services are never useful even in situation when your system is thrashing. The reason why it thrash is due to insufficient memory to accommodate the working memory set, freeing up is just taking out the cache buffer which actually deteriorate the system ram further than it should be.

Modern OS have mechanism built into the kernel to evict file buffer as priority during memory pressure, introducing a 3rd party application to manage the memory basically mess up the algorithm of the kernel

Paging out can also occur because the kernel is not purging enough memory for running processes. Problem is not all processes are given the correct priorities and stay in ram even when they are not doing anything. So if you're paging out, sometimes freeing ram can help a little.

 

[davidktw]

Paging out can also occur because the kernel is not purging enough memory for running processes. Problem is not all processes are given the correct priorities and stay in ram even when they are not doing anything. So if you're paging out, sometimes freeing ram can help a little.

You are right, file buffers are also having priority as oppose to the applications data may it be code,text or data segments. There may be various reasons to why it is so, one is it is very recently used, the other is some other part of the system is using this portion of the file. However, if the kernel decided that file buffer should be retained in the memory and evict less recently used applications segments, that is the kernel's decision.

Introducing your own set of rules don't help the kernel unless this 3rd party utility can coordinate with the kernel, which to date, I have not come across any such tools advertising as such. I have not read about any kernel offering such information external from the kernel domain.

The kernel can choose to page out application segments or file cache segments as it deem fit, what does free memory achieve ? It merely achieve the free memory column showing more capacity, but that doesn't mean it is doing a better job than the kernel.

When you say it helps, what measurement are you based on ? For all we know, this 3rd party might evict a page segment that the kernel decided it should stay, yet the 3rd party tool remove it and when it's required by the process, this piece of information is read back into the memory again. Can the 3rd party tool be ABSOLUTELY sure such situation is not happening.

It's like having 2 cooks in the same kitchen but not talking to each other. Naturally it's chaotic. If they are not even talking, who's decision is suppose to be of priority when both are operating in their own digression ?

BTW paging out is freeing physical memory, if not, what do you think paging out is for ? So why do double work ?

 

[IronMac]

If you have work with *nixes like the way I do, perhaps I would just believe you on how you beg to differ So may I enquire for your reason why FreeMemory works ? How do you tell that it is working for you ?

Ok, you've made me do a little more digging into the FreeMemory app which I've been using for about two years now.

I'm going to turn off the app and see what happens. But, I also can't go with your solution of having a system with more memory. This MBP is already maxed out.

 

[davidktw]

Ok, you've made me do a little more digging into the FreeMemory app which I've been using for about two years now.

I'm going to turn off the app and see what happens. But, I also can't go with your solution of having a system with more memory. This MBP is already maxed out.

I suppose even if you turn off or turn on, you wouldn't feel much difference. All that you feel is what you want to see

The only way to prove that your FreeMemory works or not is using scientific methods. I must say, ever since I started computing from the age of 16, I have not seen any of such utilities made a scientific approach to prove their "enhancement" works. All they did is show you that the free memory column shows free. People that doesn't understand how virtual memory works in modern operation system will buy in that more free memory is good, which is in fact silly. You buy 4GB of memory, why let them be free in the first place ? It's like you buy a football field, but you only keep on using the goal post patch, what good does it makes ?

That's why modern operating system fill all these free memory with something as much as they can. No paging algorithm today are perfect, because they are designed by human. As much as the paging algorithm we see today, they are based on certain rules but with flexibility using weigh-age system. Once weigh-age is introduced, it becomes heuristic in nature. Feel free to read up a not so low, but also not high level introduction to paging system in linux @ Linux Knowledge Base and Tutorial. Mac OS X Darwin, a variant of BSD, is no much different in concept from System V, HP-UX, AIX, Linux and all other unices. The algorithm may differs, but the big concept remains. History has told all these unices have certain amount of kernel codes exchanged and there were years (in fact on-going) where infringement of code based lead to law suit over plagiarism claims. As such, what you experience in Mac OS X is not any different form Linux. Give a good read into how virtual memory works and you will be enlightened about why free memory is a silly understanding.

I have come across some people in the IT industry who knows nothing about these kernel stuffs and insist that system should have at least 60% free memory as their SOP. These people basically don't understand unices and its kernel and force their system to under perform when they spend X dollars procuring that amount of memory just to make them look good on the chart. I have one word for them "stupid".

The correct approach to scientifically ascertain that FreeMemory work is by performing memory load test aggressively over a period of time using certain workload. My recommendation is using web server as a good gauge. Web servers are very interesting workload because they exhibit large working memory set, heavy using of mapped memory files for serving resulting in high I/O, and also high computing requirement for dynamic pages.

When you run a lengthy sequence of request over a period of time, you iron out biased due to probability and you get 2 important result, total throughput and median response time. These 2 values basically tell the overall health of the system and its average performance.

You must understand, no individual components in the whole system is isolated. As such performance done on isolated components are highly inaccurate. I have a good example, if you run a tight for..loop code, what will you find ? You will find the processor spin at 100%. My question for you is why isn't when you run an application that you find all the time it runs at 100%. It does, but not all the time. The rest of the time, even if it's not waiting for user response, the application is interacting with the memory, the hard disk or even the network. When you combine all these together, you find delays and latencies all over the place causing certain load to be not CPU bounded. If the paging algorithm of FreeMemory is really better, then it should result in a better overall performance in the system with all other components remain the constant.

This is the only scientific approach to ascertain if FreeMemory is work as claimed. So why is FreeMemory not doing such tests ? Anything else is BS, sales and marketing gimmick. With billions of non-IT people living in earth, what does a million user means ? NOTHING...

Tell me 1 million IT personnel uses it and certify it's working, it will be a more convincing feedback versus 1 million people with only 5% of them are IT personnel and yet not necessarily equip with the right knowledge to assess FreeMemory's claim.

I hope it make sense to you

 

[negativzero]

You are right, file buffers are also having priority as oppose to the applications data may it be code,text or data segments. There may be various reasons to why it is so, one is it is very recently used, the other is some other part of the system is using this portion of the file. However, if the kernel decided that file buffer should be retained in the memory and evict less recently used applications segments, that is the kernel's decision.

Introducing your own set of rules don't help the kernel unless this 3rd party utility can coordinate with the kernel, which to date, I have not come across any such tools advertising as such. I have not read about any kernel offering such information external from the kernel domain.

The kernel can choose to page out application segments or file cache segments as it deem fit, what does free memory achieve ? It merely achieve the free memory column showing more capacity, but that doesn't mean it is doing a better job than the kernel.

When you say it helps, what measurement are you based on ? For all we know, this 3rd party might evict a page segment that the kernel decided it should stay, yet the 3rd party tool remove it and when it's required by the process, this piece of information is read back into the memory again. Can the 3rd party tool be ABSOLUTELY sure such situation is not happening.

It's like having 2 cooks in the same kitchen but not talking to each other. Naturally it's chaotic. If they are not even talking, who's decision is suppose to be of priority when both are operating in their own digression ?

BTW paging out is freeing physical memory, if not, what do you think paging out is for ? So why do double work ?

I understand where you're coming from. And the kernel is extremely good at what it does, which is managing ram. It would be ideal if we could get the kernel to dump and cache ram when we want but like you said there is no such app that does that now and the problem is the computer doesn't know what the user needs at the moment.

Freeing ram is a way of telling the computer that the processes it has now are not needed and it needs to dedicate its resources to using one app in the moment, eg photo editing.

That's why sometimes people see that "boost of speed" when they dump their ram when their ram is maxed. Because the resources the computer is dedicating to it right now is exactly what the user is doing. Whether this is consequential to slow or fast performance in future depends again on how the computer manages the ram from then on. Which is why if you are paging it helps to free ram a little. In a non paging environment, freeing ram is useless.

 

[davidktw]

I understand where you're coming from. And the kernel is extremely good at what it does, which is managing ram. It would be ideal if we could get the kernel to dump and cache ram when we want but like you said there is no such app that does that now and the problem is the computer doesn't know what the user needs at the moment.

Freeing ram is a way of telling the computer that the processes it has now are not needed and it needs to dedicate its resources to using one app in the moment, eg photo editing.

That's why sometimes people see that "boost of speed" when they dump their ram when their ram is maxed. Because the resources the computer is dedicating to it right now is exactly what the user is doing. Whether this is consequential to slow or fast performance in future depends again on how the computer manages the ram from then on. Which is why if you are paging it helps to free ram a little. In a non paging environment, freeing ram is useless.

One think since you understand, the kernel knows best how to manage the memory, so if that's the case, how does another external party help ?

You mentioned freeing memory is a way of telling the computer, are you sure it is actually telling the computer ? The 3rd party is not telling any system anything, it's working in its own rule, so where is the communication to ascertain your claim that indeed it is telling the computer ? Do you see the contradiction ? I'm not playing with words here, I'm asking for claims validation.

Besides across the whole system, multiple application share the same code base such as the glibc library, the dnsresolver library, the network modules, the graphical library of mac os x, and so forth, so where is the line drawn between this application and that application ? There is no such thing as let more to this application or that application. Please refer to my post #18 above, and at the same time, take as much time as you need to understand better at the granularity of how the kernel visualise memory allocation and deallocation. You will see even when you open an application, you don't read in all the contents of a file, which is memory mapped.

The memory boost is a placebo effect that most have experience. You see what you want to see. The only way to say freeing up memory is helping the system is to use actual scientific tests, no other way of "guts feelings" make sense. But when you are doing so, remember, when you free up the memory, you are at the same time cannibalising other parts of the system happening in the background that you did not time, nor can you see since they are not necessarily a windowed application. So if you tell me that it is useful, you better be able to measure it, or else all talking is pointless.