Audio Decoding KILLS MacPro (2009) green factor

[nanofrog]

how does that compare to a similarly specc'd US imac? (i7, 1TB, 4GB RAM, education discount and full warranty are all the changed aspects)

The base i7 iMac is $2199.00USD ($2,377.95 AUD) before any discounts.

tariff on IT coming into australia sucks. the ISP panel would be a good $1k i reckon. i donno where all the rest of it comes from. CPU/RAM/HD = ~$600 all up. GPU would be a good $500 (its the mobility GPU i think).

The closest I can think of is the NEC 30" IPS unit (LCD3090WIXi), and it's $1900USD (best price I located in the US, and they ship to AUS).

Assuming you really don't need that much monitor, you could save yourself some serious money. Or better yet, forget IPS, and go with TN or S-PVA. More units available, and much better pricing. A 27" TN can be had for under $300USD IIRC.

An exact part breakdown might be a good idea (but send it email - keep it off the thread, as it's going sideways).

HA! payment methods for helping others in our uni generally revolve around beer, or cartons of it.

A University that pays in beer (not students paying students for favors - that's common here too)?!?!?

(That would be a major lawsuit, and maybe jail time here for underage drinking - US legal drinking age = 21).

 

[DoFoT9]

nano, email sent

no further news on audio problems thus far

 

[smacman]

For what it's worth, I have used SmcFanControl to change my BOOSTA fan idle speed to 1300rpm. The machine is still very quiet and the max temp noted while playing audio is now a more respectable 51C. Still annoying, but at least I don't feel like I'm cooking it.

NOW IF APPLE WOULD JUST FIX THIS!!

 

[nanofrog]

For what it's worth, I have used SmcFanControl to change my BOOSTA fan idle speed to 1300rpm. The machine is still very quiet and the max temp noted while playing audio is now a more respectable 51C. Still annoying, but at least I don't feel like I'm cooking it.

NOW IF APPLE WOULD JUST FIX THIS!!

Good to know it's not horrible on noise, but you're right. Apple needs to address this.

 

[smacman]

Now that we know the Quad i7 iMacs are NOT having this same issue, I wonder what the chances are that the build of OSX included with them is somehow different. Maybe they are running a special build that is optimized for i7 architecture.

 

[nanofrog]

Now that we know the Quad i7 iMacs are NOT having this same issue, I wonder what the chances are that the build of OSX included with them is somehow different. Maybe they are running a special build that is optimized for i7 architecture.

Links to the i7 iMacs would be nice, but ultimately, I'd like to see more than a couple of posts on each machine (as the testing may not have been done correctly).

Assuming it has, and these systems don't have the same issue as the MP's do, then it's likely related to the differences (namely the QPI on the LGA1366 vs. DMI on the LGA1156 models). And it's still likely the software, as it's not seen under Windows (if it were pure hardware, any OS run on the LGA1366 parts would be running high temps for the same tasks).

We need more data IMO.

 

[dzam]

I have same problems with my 2.66GHz Quad-Core early 2009 Mac Pro.

When I'm not listening to music or any kind of sound, temperatures are something like 35-45°C. When I start iTunes, Spotify or look some video from YouTube, temperatures are rising up in very small time.

Temperatures goes up to 65-67°C.

So, the problem is somewhere in OS X? Any information when this is going to be fixed?

 

[DoFoT9]

I have same problems with my 2.66GHz Quad-Core early 2009 Mac Pro.

When I'm not listening to music or any kind of sound, temperatures are something like 35-45°C. When I start iTunes, Spotify or look some video from YouTube, temperatures are rising up in very small time.

Temperatures goes up to 65-67°C.

So, the problem is somewhere in OS X? Any information when this is going to be fixed?

interesting. so just to clarify; the temperatures raise to 65°C-67°C even when just listening to music in itunes? and nothing else?

 

[nanofrog]

interesting. so just to clarify; the temperatures raise to 65°C-67°C even when just listening to music in itunes? and nothing else?

It matches the experience of other posts early in the thread IIRC. So it's not a new behavior.

 

[DoFoT9]

It matches the experience of other posts early in the thread IIRC. So it's not a new behavior.

yup sure does, was just making sure that he wasnt running youtube + itunes at the same time. because we all know flash on macs SUCK!

 

[Transporteur]

because we all know flash on macs SUCK!

Yes it does, but especially on the i7 MacPro.

The Safari Flash Player plugin on my MacPro shows more than 200% CPU usage whereas the plugin on my MacBook uses about 60-70%.

Both machines have 2,26GHz, Safari 4 and OS 10.6.1.
Pretty strange!

 

[DoFoT9]

Yes it does, but especially on the i7 MacPro.

The Safari Flash Player plugin on my MacPro shows more than 200% CPU usage whereas the plugin on my MacBook uses about 60-70%.

Both machines have 2,26GHz, Safari 4 and OS 10.6.1.
Pretty strange!

seems odd yes. do you have the latest flash updates? maybe that is worth a try?

 

[Transporteur]

Flash updates? Good question.
I never updated everything besides what comes with the OS X software update.

Will try that tomorrow morning.

 

[nanofrog]

Yes it does, but especially on the i7 MacPro.

The Safari Flash Player plugin on my MacPro shows more than 200% CPU usage whereas the plugin on my MacBook uses about 60-70%.

Both machines have 2,26GHz, Safari 4 and OS 10.6.1.
Pretty strange!

That's high on the MacBook, but horrible on the MP.

DoFoT9: It seems that even adding other applications to the mix of the sort you listed don't affect the temps that much. We need more information to narrow it down though. But ultimately, Apple's the one that needs to nail down the causality, and fix it. If the iMacs aren't going to have this behavior, we may not see it anytime soon, as 65 - 67C is still within Intel's spec. It's a potential excuse for them to ignore it.

 

[DoFoT9]

Flash updates? Good question.
I never updated everything besides what comes with the OS X software update.

Will try that tomorrow morning.

oh wow ok. well if you recall reading back when SL came out, there was a BIG hole in the flash that came with 10.6! so i hope you are not running that version, otherwise people can "hack" your computer. not really, but there is a pretty big security breach. so get onto it asap please! let us know if there are any performance increases

DoFoT9: It seems that even adding other applications to the mix of the sort you listed don't affect the temps that much. We need more information to narrow it down though. But ultimately, Apple's the one that needs to nail down the causality, and fix it. If the iMacs aren't going to have this behavior, we may not see it anytime soon, as 65 - 67C is still within Intel's spec. It's a potential excuse for them to ignore it.

true. we need more input! in a controlled environment, but we are not really going to get it the imacs seem fine from what i can tell, had input from about 6 users so far. going to apple store today to see if they have the quad cores to try and replicate the problem

 

[dzam]

yup sure does, was just making sure that he wasnt running youtube + itunes at the same time. because we all know flash on macs SUCK!

Yeah. Only iTunes. Or only Spotify. But however, temperatures goes up and down and up and down and up many times in a day. And it isn't good for components.

I have sent bug report to Apple. Do the same thing, if you hasn't yet.

 

[DoFoT9]

Yeah. Only iTunes. Or only Spotify. But however, temperatures goes up and down and up and down and up many times in a day. And it isn't good for components.

I have sent bug report to Apple. Do the same thing, if you hasn't yet.

ahh ok thanks for clarifying! nope not good for components at all, they prefer to be at a constant temperature i guess (i.e. a cold one).

 

[nanofrog]

true. we need more input! in a controlled environment, but we are not really going to get it the imacs seem fine from what i can tell, had input from about 6 users so far. going to apple store today to see if they have the quad cores to try and replicate the problem

We could get close enough to a controlled environment though, by setting up a testing proceedure, and application set. The issue is, how well are the proceedures written (to generate data accurately/properly), and how they're carried out (variances due to human error). So it can be a little difficult. But it's the best a bunch of scattered people can do.

Oooh. A whopping sample size of 6!

Hopefully, you can reproduce it in the store tomorrow. Sample size (n) is growing.

 

[DoFoT9]

We could get close enough to a controlled environment though, by setting up a testing proceedure, and application set. The issue is, how well are the proceedures written (to generate data accurately/properly), and how they're carried out (variances due to human error). So it can be a little difficult. But it's the best a bunch of scattered people can do.

well im sure me and you could come up with a fool proof procedure to follow and give to users? you write it and ill test it - im the dumbest person at following procedures so by having me test it, it will be fool proof!!

Oooh. A whopping sample size of 6!

shut up lol! how many do you have!?!?!

Hopefully, you can reproduce it in the store tomorrow. Sample size (n) is growing.

it might get to 7!

p.s. im going today, in about 6 hrs.

 

[nanofrog]

Yeah. Only iTunes. Or only Spotify. But however, temperatures goes up and down and up and down and up many times in a day. And it isn't good for components.

I have sent bug report to Apple. Do the same thing, if you hasn't yet.

What kind of variances are you getting?
Duration?
Frequency?

The processors can take quite a bit, so you may be safer than you think in terms of safe range for the high temps. Thermal cycling is what's more dangerous, which your concerns are valid (can actually damage the transistors physically - fracturing, or if the temps get really hot, diffusion of p and n type doped sections (aka drift)).

ahh ok thanks for clarifying! nope not good for components at all, they prefer to be at a constant temperature i guess (i.e. a cold one).

Ideally, yes. The cycle rate can be critical. A high usage period followed by a lull can be from a few milliseconds to days. The longer the time period for each, the better, as it's stable over some unit time. It's the rapid cycling that can cause damage to the crystaline structures, especially if the temps reach Tjmax.

 

[DoFoT9]

Ideally, yes. The cycle rate can be critical. A high usage period followed by a lull can be from a few milliseconds to days. The longer the time period for each, the better, as it's stable over some unit time. It's the rapid cycling that can cause damage to the crystaline structures, especially if the temps reach Tjmax.

interesting. so basically, hotter for longer is better then short bursts of hot/cold. that does make sense though as there is more pressure/stress on components etcetc.

 

[nanofrog]

interesting. so basically, hotter for longer is better then short bursts of hot/cold. that does make sense though as there is more pressure/stress on components etcetc.

So long as the extended period of heat isn't too high. I.e. Intel's rating of 67C on the LGA1366 parts isn't anything to panic over for long term. You'd be safe to run that temp 24/7. Now if it's near Tjmax (100 - 108C), then NO. Then you run the risk of drift occuring, and the part starts to malfunction permanently (partial logic failure), or flat out dies (won't even post properly). Either way, you end up replacing it.

Thermal cycling is a problem with components, not just the internal silicon in chips, but any part (capacitor, resistor,...). It has a habit of causing parts to "break" loose of the board (solder joints crack, and the part lifts or even fall off ). This has become much more of an issue with the lead free solders used now, as these solders aren't truly eutectic (solidifies in layers due to solidification temps of each metal used, rather than a uniform alloy at one temp). It can even lift the PCB traces off the board material itself (so does bad solder/desolder work).

 

[DoFoT9]

So long as the extended period of heat isn't too high. I.e. Intel's rating of 67C on the LGA1366 parts isn't anything to panic over for long term. You'd be safe to run that temp 24/7. Now if it's near Tjmax, then NO.

Thermal cycling is a problem with components, not just the internal silicon in chips, but any part (capacitor, resistor,...). It has a habit of causing parts to "break" loose of the board (solder joints crack, and the part lifts or even fall off ). This has become much more of an issue with the lead free solders used now, as these solders aren't truly eutectic (solidifies in layers due to solidification temps of each metal used, rather than a uniform alloy at one temp). It can even lift the PCB traces off the board material itself (so does bad solder/desolder work).

heat also causes capacitors to blow out dont i know it! haha. stupid time capsule!

does intel specify the Tjmax temps for the i7/i5 in the imacs? judging by what people have experienced with their imacs the temperature seems to hit a roof at about 65°C. seems pretty reasonable, but you wouldnt want that there 24/7 thats for sure

 

[nanofrog]

heat also causes capacitors to blow out dont i know it! haha. stupid time capsule!

does intel specify the Tjmax temps for the i7/i5 in the imacs? judging by what people have experienced with their imacs the temperature seems to hit a roof at about 65°C. seems pretty reasonable, but you wouldnt want that there 24/7 thats for sure

Temps affect electronics, period. Their behavior changes over temps as well (increased temp = less capacitance; the capacitance values listed on the part are at ambient = 25C). In the case of the caps you mention, that was PSU related (temp rating wasn't good enough for the specific situation).

In the case of the iMac (both the i5-750 or the i7-860), venting is going to be tougher due to the casing used. The temp is actually higher at max power (Tcase = 72.7C @ 95W). Source.

 

[DoFoT9]

Temps affect electronics, period. Their behavior changes over temps as well (increased temp = less capacitance; the capacitance values listed on the part are at ambient = 25C). In the case of the caps you mention, that was PSU related (temp rating wasn't good enough for the specific situation).

In the case of the iMac (both the i5-750 or the i7-860), venting is going to be tougher due to the casing used. The temp is actually higher at max power (Tcase = 72.7C @ 95W). Source.

that is a pretty decent temperature rating and judging by the early feedback from users the current imacs dont get that hot - thank god.

does anyone know the command in terminal to start those processes that use up the threads? ill try that at the apple store later today and record the temps