comorin
Posts: 3857
Joined: Sun May 29, 2005 5:52 am

Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 12:17 am

This has been a running discussion in my firm, and I'm hoping that the brightest minds in the world (a.netters, of course) can put this matter to bed once and for all:

Do PCs run hotter when they are 'working hard' i.e. running calc-intensive programs? Specifically, does the processor chip heat up when it does more 'calcs'? Assume that the processor is not a step-down chip.

I'm arguing not, because the the heat generated by a chip is proportional to the voltage applied, and has nothing to do with internal operations.

Please help!
 
Cadet57
Posts: 7174
Joined: Mon Jul 25, 2005 2:02 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 12:25 am

It does, We run applications and cards that have digi temp gauges in my pc repair lab and the more 0's and 1's and the more difficult the load the higher the temp that the processor runs.
Doors open, right hand side, next stop is Springfield.
 
gordonsmall
Posts: 2106
Joined: Sun Jul 01, 2001 1:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 12:30 am

They do, hence why processors/graphic cards that have cooling issues tend to show them up when running calc-intensive processes such as 3D games etc rather than internet browsing or word processing which barely task a processor.
Statistically, people who have had the most birthdays tend to live the longest.
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 1:08 am

It depends on the construction principles used in the chip.

A fully-clocked and fully dynamic CPU would draw the same amount of power regardless of what it's doing.

But modern CPUs are mostly static designs which don't need to have all their components clocked all the time, so they simply stop those components which aren't needed . In addition there are multiple ways to reduce clock frequency or to even stop execution completely until execution needs to be resumed (this can be used by the operating system when the computer is generally idling and just waiting for input, so it's only woken up periodically to animate a blinking cursor or the like).

What's drawing current is actually any change of internal signals; most current sinks in a modern high-density chip are related to capacitive loads, not to resistive ones.

When the CPU's internal clock frequency is reduced, its supply voltage can be reduced as well, so power consumption will go down even further.

Even if all power-saving features are disabled you'll probably find some difference in power consumption between idle and full load situations since the CPU would usually still optimize its power consumption to a degree.

Nowadays most CPUs do in fact use various methods for reduction of power consumption, so you'd have a hard time finding one that doesn't. At most you could disable the major reconfiguration mechanisms (such as frequency switching or idle stopping) in the operating system. But making intensive use of the floating-point unit or leaving it idling most of the time can still have a measurable impact on power consumption.

And if the operating system is working properly, the difference between full load and idle will be substantial, with noise level and chip longevity being affected as well.
 
eilennaei
Posts: 1003
Joined: Tue Nov 23, 2004 8:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 1:50 am

Yes, the modern ones do. CPU power management has grown into a art. If you have a recent generation AMD CPU, you can even see the management in action. Download a tool here: http://www.amd.com/us-en/Processors/...sources/0,,30_182_871_9706,00.html
 
j_hallgren
Posts: 1427
Joined: Sun Jun 04, 2000 11:48 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 2:54 am

Definitely! My laptop temp goes up about 20F min when CPU gets really busy. Sometimes even as much as 40 degrees diff from idle...
There are numerous monitor pgms avail on Net...the one I used on old PC was MBM (motherboard monitor), which I think is still floating around but not updated recently.
COBOL - Not a dead language yet!
 
comorin
Posts: 3857
Joined: Sun May 29, 2005 5:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:13 am

Thanks everyone for the insight! I learned a lot today.

Suppose we we have Klaus's fully clocked, dynamic chip (no power management) in steady state, will it get hotter when it's 'working' because the capacitors are changing state more frequently?
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:24 am

Not necessarily... a fully dynamic CPU has all its main components clocked all the time anyway. The difference would not be that great if the data in the pipelines isn't changing in that case.

And that is one of the reasons why dynamic designs have mostly been replaced by static ones which can be clocked down or even stopped completely without losing data as the older ones would have had.

I don't think you can buy a modern PC any more with an old-style dynamic CPU nowadays... Just ones whose power management features might not be supported properly by the operating system.
 
comorin
Posts: 3857
Joined: Sun May 29, 2005 5:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:31 am

Quoting Klaus (Reply 7):

Thanks Klaus! Apart form contributing to my understanding of chips, you've helped me win an argument with my team. I had spent some time in Hamburg many years ago: it must have made me smarter than I realized... Wink
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:36 am

The bill is in the mail...! Big grin

You're welcome.  Smile
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:37 am

Quoting Klaus (Reply 7):
Not necessarily... a fully dynamic CPU has all its main components clocked all the time anyway. The difference would not be that great if the data in the pipelines isn't changing in that case.

And that is one of the reasons why dynamic designs have mostly been replaced by static ones which can be clocked down or even stopped completely without losing data as the older ones would have had.

I don't think you can buy a modern PC any more with an old-style dynamic CPU nowadays... Just ones whose power management features might not be supported properly by the operating system.

Basically, what this means is: Macs are rubbish, PCs are better.
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:44 am

It's somewhat reckless drawing that conclusion, as Macs and PCs share the same CPU models nowadays...  mischievous 
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:45 am

Quoting Klaus (Reply 11):
It's somewhat reckless drawing that conclusion, as Macs and PCs share the same CPU models nowadays... mischievous

I'm just cutting through the computing bullcrap to get straight to the point. CPU, Intel, AMD, OSX, Vista yada yada yada...Macs are rubbish. You could just say that instead of all the jargon!
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:57 am

I see you're still traumatized from our earlier exchanges; But I'm sure its nothing a few years of intensive therapy couldn't fix...!  cool 
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 10:59 am

Quoting Klaus (Reply 13):
I see you're still traumatized from our earlier exchanges; But I'm sure its nothing a few years of intensive therapy couldn't fix...! cool

Luckily, I have a PC!
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:04 am

That probably explains the desperation speaking from your initial post...!  mischievous 
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:10 am

Quoting Klaus (Reply 15):
That probably explains the desperation speaking from your initial post...! mischievous

My PC never gets hot. In fact, it generates electricity spontaneously, and feeds it back to the grid through the plug.

It's currently working out a solution to the problems in the Middle East (will be done in an hour) and is about to make me a half-caff latte with cinnamon. Yum!
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:12 am

Well, that sounds truly marvellous, indeed!

Now if it only didn't drive you to insanity before that, it would be almost perfect, wouldn't it? Big grin
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:15 am

Quoting Klaus (Reply 17):
Well, that sounds truly marvellous, indeed!

It's just everyday life for a PC user!

Quoting Klaus (Reply 17):
Now if it only didn't drive you to insanity before that,

You're only a simple Mac user. I wouldn't expect you to understand.  Sad
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:22 am

Quoting 777236ER (Reply 18):
It's just everyday life for a PC user!

Too bad most PC users live outside your blissful parallel universe and actually have to cope with Windows...!

Quoting 777236ER (Reply 18):
You're only a simple Mac user. I wouldn't expect you to understand.

Oh, the Reality Distortion Field™ is actually an Apple invention, too - but it seems the PC world has finally caught on to that as well, at last... Big grin
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:34 am

Quoting Klaus (Reply 19):
Too bad most PC users live outside your blissful parallel universe and actually have to cope with Windows...!

It's a Window to the world, dontcha know?

The only good Mac is a Big one.
Your bone's got a little machine
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:39 am

Quoting 777236ER (Reply 20):
It's a Window to the world, dontcha know?

...to a world of horrors and frustration, as anybody can see who's following all the "rescue me!" threads in here...  mischievous 

Quoting 777236ER (Reply 18):
The only good Mac is a Big one.

Well, the smaller ones - the iMac and the Mac mini - are pretty good, too. But a multi-CPU professional Mac is better still, of course!  bigthumbsup 
 
airlinelover
Posts: 5287
Joined: Thu Jun 07, 2001 8:03 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 11:52 am

As has been stated, the answer is DEFINATELY yes!

Best way to see how it is running is to download a motherboard monitoring program (make sure it's compatible) and watch it when the computer isn't "doing" anything, then run a crapload of inense stuff and watch  Smile

Chris
Lets do some sexy math. We add you, subtract your clothes, divide your legs and multiply
 
Cruiser
Posts: 920
Joined: Fri Apr 15, 2005 2:08 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Sun Feb 26, 2006 12:25 pm

My CPU runs hot all the time...that is because I run a Distributed Computing Program which keeps my processor really busy. I have no clue what my idle temperature is, but I really do not care  Big grin

James
Leahy on Per Seat Costs: "Have you seen the B-2 fly-by at almost US$1bn a copy? It has only 2 seats!"
 
comorin
Posts: 3857
Joined: Sun May 29, 2005 5:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 12:03 am

I didn't want to start another PC thread, but here's a question for this elite group of a.netters:

Are PCs with 2 hard drives much faster than 1? At first glance, you'd think that with Queuing Theory (Single queue, two servers) it would be (like 2 elevators instead of one). In the practical world , do you end up with a much faster system if you put XP and paging files on one disk, and your apps on another?

Thanks again!
 
FlyPIJets
Posts: 832
Joined: Sat Oct 11, 2003 3:32 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 12:10 am

Quoting Comorin (Reply 24):
Are PCs with 2 hard drives much faster than 1? At first glance, you'd think that with Queuing Theory (Single queue, two servers) it would be (like 2 elevators instead of one). In the practical world , do you end up with a much faster system if you put XP and paging files on one disk, and your apps on another?

Thanks again!

If you stripe the drives its faster. (Google RAID for more info).

Macs have this built in. Not sure on the Windows OS.

[Edited 2006-02-26 16:14:02]
Rex Kramer: Get that finger out of your ear! You don't know where that finger's been!
 
gordonsmall
Posts: 2106
Joined: Sun Jul 01, 2001 1:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 12:23 am

Quoting Comorin (Reply 24):
Are PCs with 2 hard drives much faster than 1? At first glance, you'd think that with Queuing Theory (Single queue, two servers) it would be (like 2 elevators instead of one). In the practical world , do you end up with a much faster system if you put XP and paging files on one disk, and your apps on another?

There isn't a yes/no answer to this question I'm afraid - it's very much situation dependent.

It is possible to make a computer run faster by utilising two hard disks, either by dedicating the swap file/partition to it's own drive or by using raid to spread the disk access across two disks along with a variety of other methods and technologies that can be used or even combined to hurry things along.

On the other hand, it's also quite easy to bring a computer to it's knees by incorrectly configuring a multiple-disk setup. Whilst it's not as much of an issue with modern SATA interfaces, I still remember the many headaches I had to solve, caused by people putting multiple IDE disks into their machines and screwing up the jumper settings and cable positions or mixing optical drives and Hard drives on the same IDE channel - and I won't bore you with the many fine details of mixing SCSI disks, SCSI disk controllers and their various 'standards' - not to mention the headaches that are SAS drives and fibre channel disk arrays.  Smile

A much simpler (though slightly expensive) way of boosting the speed of your computer is to rid yourself of the swap partition altogether, by loading your machine with an ample quantity of good quality, high speed, low latency RAM. Whilst you need large amounts (8Gb+) even for modest setups, the performance gains can be quite impressive - especially on highly loaded DB servers.

To be honest though I question whether the performance gains achieved by any of these techniques are worth the extra investment required in the context of you average home computer user.

Regards,
Gordon.
Statistically, people who have had the most birthdays tend to live the longest.
 
Klaus
Posts: 20622
Joined: Wed Jul 11, 2001 7:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 12:38 am

Striping would be one possibility (at actually reduced reliability overall, though!).

If you have too little RAM and your system is swapping a lot, a second physical disk for the system and for swap space can make a major difference for applications which at the same time access the file system, but still a lot less than an actually sufficient amount of RAM.

For applications which just cause swapping due to temporary internal memory requirements but don't read from or write to disk at the same time, the improvement would be much less, except for the likely fact that a single disk might be relatively full and fragmented, so the swap file might get fragmented as well and thus slower again.

A separate swapping partition is not necessarily a good idea; In the worst case it can lead to a slow-down again due to more frequent and longer disk head movements being required when swapping occurs during simultaneous file system activity.

So in short, it depends...! Big grin

(But actually having sufficient amounts of RAM is by far the most effective way to go.)
 
FlyPIJets
Posts: 832
Joined: Sat Oct 11, 2003 3:32 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 12:40 am

Just wondering, 'cause I'm not a physics major.  Smile

The OP title question, do PC's run hotter on cal-intensive programs?

Is there any system that produces work and doesn't have heat as a by-product?

Maybe I am over simplifying.
Rex Kramer: Get that finger out of your ear! You don't know where that finger's been!
 
FlyPIJets
Posts: 832
Joined: Sat Oct 11, 2003 3:32 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 1:02 am

Quoting Klaus (Reply 27):
Striping would be one possibility (at actually reduced reliability overall, though!).

True, the idea is, that if one drive in a striped system fails the whole RAID system fails. So every time you add a drive to a striped RAID drive, you increase the probability of failure.

Most applications of striping don't care about drive failure. Think video rendering. During the rendering process, which uses lots of RAM and drive access, a striped drive will speed that up, and you don't care if it fails, you're only after the output, then you care.
Rex Kramer: Get that finger out of your ear! You don't know where that finger's been!
 
777236ER
Posts: 12213
Joined: Sat Aug 18, 2001 7:10 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 1:05 am

Quoting FlyPIJets (Reply 28):
Is there any system that produces work and doesn't have heat as a by-product?

Not if it obeys the second law.
Your bone's got a little machine
 
comorin
Posts: 3857
Joined: Sun May 29, 2005 5:52 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 2:43 am

Quoting 777236ER (Reply 30):
Quoting FlyPIJets (Reply 28):
Is there any system that produces work and doesn't have heat as a by-product?

Not if it obeys the second law.

I think the heat generated based on reduction of Entropy by organizing information would be too small to measure. Chips don't really do any mechanical work while computing, and as Klaus points out, when you clock the CPU, that's all the 'work' that really happens. The milliwatts that come out of a chip equals V*I , so as long as I doesn't change when working, the heat generated should be constant.
 
eilennaei
Posts: 1003
Joined: Tue Nov 23, 2004 8:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 3:38 am

Quoting FlyPIJets (Reply 28):
Is there any system that produces work and doesn't have heat as a by-product?

No, unfortunately not, every form of energy ultimately degrades into heat.

&&&&&&&&&&

Quoting FlyPIJets (Reply 25):
If you stripe the drives its faster. (Google RAID for more info).

Macs have this built in. Not sure on the Windows OS.

RAIDs are being implemented now even on cheap-grade matherboards, at least to the RAID levels 1 and 0 (simple mirroring and striping, respectively; striping IS faster than 1 drive only), level 5 is often seen as well. The necessary drivers are provided for all modern OSs on the installation CDs.
No-nonsense explanation: http://www.twincom.com/raid.html

RAID has nothing to do with swapping, which is an independent occurrence. There's a "knee" in memory requirements after which the maintenance of virtual memory (a.k.a. "swapping", which is in fact an incorrect term) will begin to toll the performance heavily.
 
eilennaei
Posts: 1003
Joined: Tue Nov 23, 2004 8:41 am

RE: Do PCs Run Hotter On Calc-Intense Programs?

Mon Feb 27, 2006 4:18 am

Quoting Comorin (Reply 31):
. The milliwatts that come out of a chip equals V*I , so as long as I doesn't change when working, the heat generated should be constant.

In reality this effect is substantial, as systems are now aware if they are being clocked into doing "nothing". It has been this way since the early 1990s and the requirements of the EPA of the US.
The reason why desktop computers can't be clocked totally down to zero lies in the nature of the main memory (RAM). This is presently DRAM, D standing for "dynamic". DRAM needs to have a certain minimum refreshment rate and hence a minimum clock and a power consumption. The situation will change when the next-generation magnetic memories are introduced.
There's nothing new actually, I'm old enough to have seen a computer whose main memory was made of tiny magnetic rings that were magnetized in either of the two directions. The system could be transported hundreds of kilometres and the execution would start from where it left off as if nothing had happened!
Modern desktops and portables solve the disappearing main memory problem by having the status of the system written to a recovery file on the hard disk and read back when the "hibernation" status is exited.

[Edited 2006-02-26 20:20:42]

Who is online

Users browsing this forum: LittleFokker, LOT767301ER and 9 guests