80646 Views Previous 1 2 3 4 5 … Next 600 Replies Latest reply: Mar 27, 2010 12:59 AM by svkrzn Go to original post
So you're saying cooler is better. No argument here. But your'e ignoring the fact that the Hatter gave:
Intel warrants Xeon for running 24/7 at more than your 60-70*C. And it is Intel that designed the processor and cooling.
The anomaly isn't that computers run hot and that shortens their lives, as it does any mechanical device. It's that apps you would think cause little extra heat at all, create quite a bit.
I'd like to see that fixed too. But I don't know where the fault lies, it seems Nehalem is the cause from one angle, and Apples OS from another. Marcel Bresink, I think, did an excellent job of pinning it on the OS, but he didn't know then that it was only happening on Nehalems. That it happens only on Nehalems, would pin it on Intel.
Well, this is where the threads on this usually end, exasperation. We simply don't have enough info and we don't know if it's being looked into at all. But we do know that Nehalems are new and being worked on to make them run better as we write. I'd say if in another six months we don't see this addressed, or the next gen MPs don't have the issue, then that's the time to get hopping mad. As it is, you are pretty far out of the range of this causing any lasting damage.
Blame it on using out of date compilers!
I know somewhere in my travels there was a good discussion on MacRumors as to max operating and extreme, when a core will shut down, and when the processor will throttle back to 50% even (110*C?) to reduce the heat induced stress.
Microsoft is having their own "WWDC" (Programming Developers Conference) currently.
Running on a treadmill, burning calories, but going nowhere... keeping hydrated and with a fan to whisk away the sweat...
Seems that most of the programs that try to read the thermal sensors are off, sometimes by a mile. All you have to do is start trying to nail down the TJ Max and it is like Quantum Physics all over: you can know spin, but not orbit, you can know the energy value, but not spin.
XtremeSystems Forums: RealTemp
There are four specifications which Intel states are required for reliability: mechanical, electrical, thermal, and stability (they call it "signal quality", or something, iirc).
Intel does not state anywhere that 86c will shorten the lifespan of an i7. They do state that if any of the four specifications are not met, reliability can not be expected.
Not what I wanted, so I'll update if I find something more.
I thought TJ Max was some kind of retail chain. The most I've gotten out of Quantum Physics is that our own observations affect the phenomena we're observing. Hence, reality is fleeting. Of course, that could be totally wrong, last time I looked it seemed different.
Man, do PC heads have a jargon and intensity all their own... I wouldn't last long in their world.
You remind of what Bresink said. These MPs don't have PPC chips, which seemed to go beyond the critical point and keep climbing. Intels just shut down. That's reassuring right there.
Edit: Hatter, what's the deal with I think, SMC Fan Control? Good or bad thing.
Message was edited by: Samsara
"what the **** does a discount clothing store have to do with computers?"
Here is another useful thread.
intel has decided for reasons unknown to NOT give users the true tjmax temps, so makers of temperature monitoring programs are forced to guess. most programs guess that tjmax is 95c for a q6600. When the temperature monitoring programs display a load temp of, say, 65c, they're actually reading a "30c to tjmax" temp from the thermal diode on the cpu.
But you are right, Intel has its own self-preservation, and they seem to feel strongly about keeping somethings to themselves.
"what the ** does a discount clothing store have to do with computers?"
Is that what they are? Well they could have expanded... No, I read your links. It all sounds like horrible stuff. Though maybe when Hardware Monitor reports a core temp of 90c, it's real temp is 60c.? Yeah, horrible stuff, beyond me and it's staying that way. I think whining and complaining is much better... I'm never getting into one these temp topics again though.
So what are your thoughts on SMC Fan Control? And can I run an "autoexec"?
I don't think the question was directed at me but regarding smcFanControl, it works fine on all Mac Pros. Since I noticed the audio/temp issue, I changed my BOOSTA (CPU) fan idle speed to 1300. The system is still very quiet, and temps now stay below 50c when listening to music. Not a proper fix, but better than nothing for now...
Most any response is appreciated.
Right now, I'm encouraged that the Flash-CPU issue has been addressed. In a more limited way for Macs it seems, but a very substantial improvement. That's from installing the Adobe 10.1 beta.
So I'm going to hold off on worrying about Audio temps for awhile and see how that's addressed in the future, without any modifications.
This is a the first thread on this where I've seen that someone's Mac had actually shut down. To me, perhaps ignorantly, I find that encouraging. These MP's aren't going to die from this. These aren't G5's. I had too many of those to go through that again.
I hope the poster whose Mac shut down writes back with what Apple did to help him.
Whether or not these heat increases will cause problems down the line, I really don't know, I don't like it either and I'd like someone to look into it as soon as possible. It's worrying though, while the Flash issue was picked up right away when discovered by the Mac press, I can't remember the iTunes heat issue being talked about at all.
Thanks for your answer.
I wrote in a few days ago about my 2.93 Quad 2009 Mac Pro crashing. It was sent in to Apple for repair/testing and apparently passed all tests. Back home it crashed within 40 minutes. Just running iTunes the temperature went from 34c to 60c in 2 minutes, to 75c in 8 minutes and 85c in 35 minutes. I then opened Lightroom and started an export, temperature goes to 94c in 1 minute, followed by a kernel crash and corruption of my Lightroom database.
With further testing if I run both applications together the temperature reaches 94c within 15 minutes followed by the kernel crash. This is clearly unacceptable.
These tests were run on a stock machine (Apples own ram and single HD etc). I am using Marcel Bresinks Temperature Monitor 4.8, temperatures quoted are from the CPU core. Ambient temperature 26c.
My 2.66 quad dual suffers the audio roasting problem. Often seen well above 80C. Found Apple support unbelievable unhelpful and patronizing. They repeat that the Nehalem runs hot and that this is normal (which I emphaticaly do not believe - especially on <1% CPU load). So I disabled SMCfan, and stopped watching Temperature Monitor etc. And now my power supply has blown.
Whilst it is in repair (apparently it takes AppleCare UK up to 3 weeks to change a power supply - which I find exasperating and utterly incompetent, I'm 50 miles from London) I phone AppleCare about the audio/thermal issue. After the usual run around with Indian operators who have incomprehensible accents and are so not with-it, I ask to speak to a 2nd level operator - they won't do that unless I have a machine to run tests on - duh.
I should have just lied, I know this thermal problem backwards, as does everyone who has struggled with it (ie. anyone like us outside Apple).
The lasting impression is that Apple has its hands clasped over its ears, is looking at its feet and humming "I am perfect so there is no need to listen", at least the bozos who guard the inner sanctum where competent people work.
Sorry, just very annoyed at AppleCare treatment, even to the point of reconsidering using Apple after more than 20 years, and regretting getting a loaded Mac Pro.
Apologies all for the rant above.
The issue in the sound/thermal problem that Apple guys don't seem to be able to grasp is that (a) the roasting is not connected with the CPU load, and (b) no sound has to be playing - that is, run a Youtube clip in a browser, leave that browser page open after the clip has finished playing - the temperature stays high. Close page, temp drops to normal. Same with other sound playing apps. Video only does not cause the problem.
My hunch is that it is the MacOS not handling turbo boost mode appropriately, I assume audio is so little load one thread only is used (just a wild guess at a direction).
The mains power load climbs too, so where is the power being drawn when no work is being done?
Last point, how is one to live with a cult like Apple (we must be cult members really) when there is no feedback. We wail at the outer walls, but the high priests far inside are deaf. Sometimes like the Rosetta logout/blue-screen bug, they get around to fixing it, but we are left in the dark with no information as to whether they are on the case or not, for customers/users it is just prayer and waiting (and trying to tolerate the 1st level support people). It would make a world of difference if they would let us know that they have heard/understood, and whether they intend to fix a bug like this (if so, with what urgency).
Drat, I've been ranting again (horrendous hardware support drives me batty).
Eh, if you've had Macs for awhile you're allowed a rant once in awhile.
I was surprised at the YouTube thing, I hadn't noticed it before. While this morning playing a song in iTunes I went from 4 cores reading about 53C to one at 82 and the other three 77-79, with my other other cores then showing a 20 degree heat rise from the 40s into the high sixties.
When I went to check out Youtube, my highest temps for any core went into the mid 70's and the rest were substantially lower. Not good, but overall better than iTunes.
The Youtube page also released the heat a lot quicker when I closed it than with iTunes. Seconds compared to minutes. It should be noted that I installed the Flash Player 10.1 beta, http://labs.adobe.com/technologies/flashplayer10/
, which has dramatically reduced CPU usage for Flash in Safari.
I'm just hoping that's not why I now notice that flash sites are heating up the cores more than I remember. But I don't remember so take that only as something to check, before and after, if one wants to install it.
The Flash issue made the Mac Press big time awhile back, maybe that's why is was addressed so quickly. But I haven't seen anything on the iTunes bug anywhere. Unfortunately, if more machines break down like yours we will. And then Apple may act.
Someone running Windows 7 Pro on 2009 Mac Pro care to verify that none of this happens?
Flash 10.1 beta was not released to address anything going on with Mac OS. There is a "move afoot" to bring the power of the GPU to web browser and help render pages faster and allow for rich media. And give OpenCL something to play with.
Flash 10.1 beta was not released to address anything going on with Mac OS
It doesn't have everything for Macs that it has for Windows but it does have the CPU fix and can be downloaded for that alone. This page that I linked to before, http://www.sheridanandco.com/
used to show the CPU percentage in Activity Monitor into the 450-500% range. Now, it rarely goes over 50% but spikes towards 100 every so often. Still, a big, big step from where it was without the update.
Edit: Happy Thanksgiving.
Message was edited by: Samsara