Audio Decoding KILLS MacPro (2009) green factor

[Inconsequential]

I just did the same tests with my machine, even though I had several other apps running.

Average of 3 runs Geekbench without iTunes playing: 13304
Average of 3 runs Geekbench with iTunes playing: 12897

Makes a difference of 3%.
Absolutely nothing to worry about.

I don't have the time to make some real life tests now, but I will keep an eye on it as soon as possible.

Mine dropped by 5% on average in Geekbench.

I've already done a battery of real-life tests, your machine is inflicted but as you have the octo your lucky as your performance drop is less. But it is still there.

See my tests here: https://forums.macrumors.com/posts/9030203/

"Absolutely nothing to worry about"...

Doesn't help the people with a serious problem with a serious drop tho does it?

Cinebench R10 multi-CPU render w/o iTunes: 17,672 (Safari running)
Cinebench R10 multi-CPU render w/ iTunes: 17,105 (Safari running)

Difference: 3.2% slower w/ iTunes

No firewire devices were attached or used during this test.

WHAT'S UP WITH THIS? My results don't show appreciable slow down with either Geekbench or Cinebench while running iTunes in the background. Cinebench is more sensitive and shows a 3% decline in performance, but since iTunes has 2-3% CPU utilization, the results are perfectly in line with expectations. OpenGL scores take a 21% hit in Cinebench when running iTunes on my machine.

Truly bizarre results.

Look at your single core results, they show the largest drops.

and 3% cpu usage /= 3% performance drop...

 

[AZREOSpecialist]

Mine dropped by 5% on average in Geekbench.

I've already done a battery of real-life tests, your machine is inflicted but as you have the octo your lucky as your performance drop is less. But it is still there.

See my tests here: https://forums.macrumors.com/posts/9030203/

"Absolutely nothing to worry about"...

Doesn't help the people with a serious problem with a serious drop tho does it?



Look at your single core results, they show the largest drops.

and 3% cpu usage /= 3% performance drop...

A CPU an do a finite amount of work in a finite amount of time. If you consume 3% of its total capability, you only have 97% of the full CPU performance available to other applications. I say tomato, you say...

 

[AZREOSpecialist]

What raises my eyebrow is the 21% drop in OpenGL performance simply by playing a song in iTunes. Isn't OpenGL performance dependent upon your graphics card? Something about playing iTunes causes OpenGL to run like a dog.

 

[TonyK]

Results in Knoxville, as if it matters.

My MP 2008 with 2x2.8 runs fine

My wife's MP 2009 (early) with 1 quad core Xeon went from 35C to 59C when I stopped it after 3 songs.

THIS IS HORRIBLE. Glad we have AppleCare. Probably going to have to use it soon. Now back to reading the rest of this monster thread.

 

[smacman]

What raises my eyebrow is the 21% drop in OpenGL performance simply by playing a song in iTunes. Isn't OpenGL performance dependent upon your graphics card? Something about playing iTunes causes OpenGL to run like a dog.

Exactly.. Something is messed up, and instead of debating the severity of this performance penalty, we should work on pressing Apple for a fix.

 

[Inconsequential]

A CPU an do a finite amount of work in a finite amount of time. If you consume 3% of its total capability, you only have 97% of the full CPU performance available to other applications. I say tomato, you say...

You never use 100% of a CPU, hence why you never, ever hit a CPU's TDP.

Only the manufacturers can do that.

Just because the system is using 100% doesn't mean every single transistor in the CPU is firing, especially now they have integrated memory controllers.

 

[TonyK]

Installed smcFanControl on the wife's system. Set BoostA to 1800 rpm and all the other fans it found to 800. Played 6 songs and temps never got higher than 44C.

A 15C difference is pretty good. Also browsed the web, made a post in the Apple discussion forums and started iPhoto. I can live with 44C for the moment BUT Apple needs to address the real issue. Playing MP3 in iTunes, or streaming music in a browser, should not increase the temp of the CPUs that much.

Take care,

Results in Knoxville, as if it matters.

My MP 2008 with 2x2.8 runs fine

My wife's MP 2009 (early) with 1 quad core Xeon went from 35C to 59C when I stopped it after 3 songs.

THIS IS HORRIBLE. Glad we have AppleCare. Probably going to have to use it soon. Now back to reading the rest of this monster thread.

 

[Inconsequential]

Installed smcFanControl on the wife's system. Set BoostA to 1800 rpm and all the other fans it found to 800. Played 6 songs and temps never got higher than 44C.

A 15C difference is pretty good. Also browsed the web, made a post in the Apple discussion forums and started iPhoto. I can live with 44C for the moment BUT Apple needs to address the real issue. Playing MP3 in iTunes, or streaming music in a browser, should not increase the temp of the CPUs that much.

Take care,

People are on the wrong track.

Something to do with the CPU is being used, hence and foremost the performance drop, then the power usage and finally temperatures.

Fix the performance drop, fix the rest...

 

[Kallel00]

Why dont you guys submit this to performance testing hardware sites like Anandtech or TomsHardware?

 

[cl-user]

[...] MP 2009 (early) with 1 quad core Xeon went from 35C to 59C when I stopped it after 3 songs.

Dropped in to Apple Store (Regents St London) yesterday. Got on a 2x2.26 MP, downloaded Temp Monitor, ran iTunes, temps rose, but nothing too bad. Ran a stress test (1590% load) and left the machine for 15 minutes (whilst other folks puzzled at the laggy machine). Mac temp got to 61C (ie. AFAIK within operational range). Yes mad power hog and hotter than it should be on audio, but I'd be happy using that machine with audio or anything any time (ie expect no damage).

My conclusion, this audio bug hits 2.66 and 2.93 09 dualies worst (meaning bad enough to damage the CPUs). Not that any should waste power on audio.

My guess is that the 2.93 gets hotter (running core audio) than the 2.66.

BTW the Apple geniuses that I explained the problem to were almost all totally unhelpful. Usual issues: basically over their heads and arrogant/blindingly following the party line. One chirped in, (when I explained that the same box did not roast when using Windows on it (ie. it's an OS issue)) saying "then buy a PC is that makes you happy".

One "genius" tried to help, but said that there was no mechanism for them to get through to the appropriate people in Apple.

It can feel like trying to report that a priest has been playing with children's private parts - total denial of there being any possibility!

I'm getting the impression that the case for energy wastage (carbon footprint, power costs, etc) is hopeless, Apple is utterly indifferent. Only a wave of hardware failures seems likely to get them to consider the issue at all. But given that that hasn't happened yet, only time will tell whether Apple gets lucky and the failures come in after the 3 year period.

 

[akadmon]

only time will tell whether Apple gets lucky and the failures come in after the 3 year period.

I suspect historically majority of MP owners don't buy Applecare, since most problems that are likely to occur (dead drive, bad RAM) are easily dealt with and less expensive than the cost of extra warranty, or happen to a components that Applecare does not cover anyway (third party drives, RAM). In this case, however, we're talking about an increased chance of processor getting fried, so one would be foolish not to get Applecare!

 

[nanofrog]

I suspect historically majority of MP owners don't buy Applecare, since most problems that are likely to occur (dead drive, bad RAM) are easily dealt with and less expensive than the cost of extra warranty, or happen to a components that Applecare does not cover anyway (third party drives, RAM). In this case, however, we're talking about an increased chance of processor getting fried, so one would be foolish not to get Applecare!

There seems to be enough MP sytems that need PSU and/or logic board replacements that the extended warranty (Apple Care) is a really smart thing to do, as the cost of replacement parts and labor post warranty coverage is at least as much as the warranty cost (i.e. PSU part is ~$300USD, and a logic board <-06 - '08> has sat around ~$800USD last I checked). Quite pricey IMO, so the cost of the extended warranty seems well worth it.

 

[smogsy]

for increased media attention i've sent a email to Chris Pirillo (lockergnome) he has a massive geek user base & hopefully help us get the word out so apple will fix this faster.

HI Chris

there is a new issue being seen by all Mac pro 2009 users (seems like its related to the new QPI) & we need your help to get it noticed as you have great media attention & a lot people watch your videos. we have already sent emails to Steve jobs (not replied as of yet) apple stores are saying the mac pros are "within its limits". so i was hoping you could a video to show users of the problems & how serious they are,so Apple will hopefully notice your video and be want to fix the issue soon rather than say its "not a issue". I'm not looking for anything sinister here just leveraging your media attention to help the mac pro community

below is a the bug/problem


The bug which was noticed around October 08, has been noticed by a large majority of Mac Pro 09 users. they are reporting is regarding audio playback & Audio Devices with there Mac Pro 2009. Users are reporting ambient temperatures between 25c-30c however after 2-15 minutes of just playing back a song from Itunes temperatures increase to 50-90c which is very alarming for very light usage. A number of Hardware Monitoring tools are reporting all cores are 1% usage or less whilst these temperatures are hitting 90 degrees. Unfortunately this is a side affect of the most worrying problem. whilst playing any Audio, power consumption increases to 50w for playing a simple Audio file, which does not fit Steve jobs strive to be a great leader in being Green.

There are also other ways of reproducing this problem. which include: plugging in any external Audio Device, Copying Data over firewire & opening professional applications like Logic. Some People have took there Mac Pro 09 to the store due to the high temperatures & Kernel Panics & to be told "there within there limits" or they could not reproduce the problem.


your can read more on this


https://forums.macrumors.com/threads/795966/ - Mac Rumors

http://discussions.apple.com/thread.jspa?threadID=2228019&start=0&tstart=0 - Apple Forum.

Hope you can help
Smogsy

 

[akadmon]

There seems to be enough MP sytems that need PSU and/or logic board replacements that the extended warranty (Apple Care) is a really smart thing to do, as the cost of replacement parts and labor post warranty coverage is at least as much as the warranty cost (i.e. PSU part is ~$300USD, and a logic board <-06 - '08> has sat around ~$800USD last I checked). Quite pricey IMO, so the cost of the extended warranty seems well worth it.

I had bought Applecare for my 2006 MP (never got to use it, d**n!), and I would have gladly paid another $200 to Apple for two more years (since I'm planning to go on to my MP for at least that long). Unfortunately that is not an option

One great thing about Applecare is that, even though my MP is off it, I still get phone support when I'm in a pinch. I just give them the serial number of my daughter's MBP and I'm good to go. I'll be covered this way for almost 3 more years

 

[AppleWorking]

Sorry for not reading the whole thread, but I just want to be clear on this... The audio and FW temp/performance issue only appears in OSX and not in Windows, right?

 

[Inconsequential]

for increased media attention i've sent a email to Chris Pirillo (lockergnome) he has a massive geek user base & hopefully help us get the word out so apple will fix this faster.

Missed out the performance change there and tbh thats the most important thing.

If we are going to tell the media, gotta get the full and correct story out...

 

[BillB50]

I ordered the Maya PCIe sound card today. It should be here in a few days. I will post my results with before and after temperatures for this thread as soon as I get my hands on the card.

-bill

 

[AZREOSpecialist]

More "benchmarks"

I ran 8 instances of "yes" through Terminal. Wikipedia and other sources point to the "yes" command as a way to quickly max out CPU usage as a method for testing various different cooling systems for CPUs. I went ahead and ran 8 instances of the following command:

$ yes > /dev/null

All 4 of my physical cores, as well as 4 hyperthreading cores, show as maxed out in Activity Monitor. CPU utilization shows 8 x 97%, CPU graphs are all the way to the top for all cores. I also am running iTunes in the background playing a song for kicks. For the last ten minutes, CPU A Temperature Diode is at 70C and all cores functioning at 83-85C. This is with all 4 physical cores and 4 virtual cores maxed out.

After killing all 8 instances of "yes", CPU A Temp dropped quickly to 55C with cores in the 64C-66C range. After killing iTunes, temps dropped again to below 35C. I'm not at all trying to minimize the problems others are having, which appear to be quite serious, but I don't see the overheating issue on my system manifesting itself. Yes, my temps rise more than one would expect while running iTunes, but not to the point that it's a cause for concern. Running iTunes at 3% CPU utilization spikes the CPU 20C, and using the remaining 397% available to the 4 physical cores apparently causes only a 15C rise in temperature above iTunes levels. How do we know for sure that iTunes isn't solving global warming simulations in the background?

Smacman, have you tried re-seating your CPU and doing a clean application of thermal paste?

 

[BillB50]

I ran 8 instances of "yes" through Terminal. Wikipedia and other sources point to the "yes" command as a way to quickly max out CPU usage as a method for testing various different cooling systems for CPUs. I went ahead and ran 8 instances of the following command:

$ yes > /dev/null

All 4 of my physical cores, as well as 4 hyperthreading cores, show as maxed out in Activity Monitor. CPU utilization shows 8 x 97%, CPU graphs are all the way to the top for all cores. I also am running iTunes in the background playing a song for kicks. For the last ten minutes, CPU A Temperature Diode is at 70C and all cores functioning at 83-85C. This is with all 4 physical cores and 4 virtual cores maxed out.

After killing all 8 instances of "yes", CPU A Temp dropped quickly to 55C with cores in the 64C-66C range. After killing iTunes, temps dropped again to below 35C. I'm not at all trying to minimize the problems others are having, which appear to be quite serious, but I don't see the overheating issue on my system manifesting itself. Yes, my temps rise more than one would expect while running iTunes, but not to the point that it's a cause for concern. Running iTunes at 3% CPU utilization spikes the CPU 20C, and using the remaining 397% available to the 4 physical cores apparently causes only a 15C rise in temperature above iTunes levels. How do we know for sure that iTunes isn't solving global warming simulations in the background?

Smacman, have you tried re-seating your CPU and doing a clean application of thermal paste?

I question this logic - I read Anand's article on upgrading CPUs on a Nehalem MP. He found that 8 core MPs use special CPUs without the normal outer shield/heat spreader. In fact, he ruined his processor board during the tightening portion of the upgrade. I can't believe our issue boils down to improper assembly from the factory? I have been following this thread and don't recall anyone trying this and reporting success - am I incorrect.

Not intending a flame with this post, just questioning reseating. It seems that Apple has done enough research/work with Intel to order specific processors assembled without outer heat spreaders so that heat could more quickly be pulled away from the cores (here is a link to the article again:http://www.anandtech.com/mac/showdoc.aspx?i=3597&p=10 )...

 

[raziel777]

Forgive me if this has already been posted, but has anyone checked this in 10.5.8 at all? I checked today and it it appears to pull the same voltage, but the temps. did not rise as much. In Leopard my fans ran at higher RPM's than in Snow Leopard. With Snow leopard they are almost silent. This might explain the higher temps. for some people. Still takes a loss in geekbench though.
Still dissapointing either way.

 

[smacman]

I ran 8 instances of "yes" through Terminal. Wikipedia and other sources point to the "yes" command as a way to quickly max out CPU usage as a method for testing various different cooling systems for CPUs. I went ahead and ran 8 instances of the following command:

$ yes > /dev/null

All 4 of my physical cores, as well as 4 hyperthreading cores, show as maxed out in Activity Monitor. CPU utilization shows 8 x 97%, CPU graphs are all the way to the top for all cores. I also am running iTunes in the background playing a song for kicks. For the last ten minutes, CPU A Temperature Diode is at 70C and all cores functioning at 83-85C. This is with all 4 physical cores and 4 virtual cores maxed out.

After killing all 8 instances of "yes", CPU A Temp dropped quickly to 55C with cores in the 64C-66C range. After killing iTunes, temps dropped again to below 35C. I'm not at all trying to minimize the problems others are having, which appear to be quite serious, but I don't see the overheating issue on my system manifesting itself. Yes, my temps rise more than one would expect while running iTunes, but not to the point that it's a cause for concern. Running iTunes at 3% CPU utilization spikes the CPU 20C, and using the remaining 397% available to the 4 physical cores apparently causes only a 15C rise in temperature above iTunes levels. How do we know for sure that iTunes isn't solving global warming simulations in the background?

Smacman, have you tried re-seating your CPU and doing a clean application of thermal paste?

I know what your saying, but even if I could get my temps slightly lower by voiding my applecare and re-seating the CPU, I would still have a system with a CPU power management fault. The temps are only a symptom...

 

[smacman]

Can someone try deleting the AppleIntelCPUPowerManagement.kext, restarting, then report back? I am far from my Mac Pro at the moment and unable to try this myself..

 

[eyeruh]

I previously saw a ~20% drop in performance when testing my MP with Lightroom and Xbench but I didn't try geekbench until today. Geekbench only showed about a 5% performance drop BUT there was a 20% drop in memory performance.

This makes a certain amount of sense to me since Lightroom transfers a lot of data in and out of memory--I remember one of the Lightroom architects saying the FSB was a real limiting factor for them on many machines. This really does make me wonder if it's not something related to QPI . . .

 

[smogsy]

Missed out the performance change there and tbh thats the most important thing.

If we are going to tell the media, gotta get the full and correct story out...

Hey ive made a formalized letter to send to Anandtech,Wired,cnet,tech world
news,try for yahoo & for google news,twitter (lockergnome)

can you please verify ive not missed anything out?

I am writing to you Wired because I have a very interesting story for you, it’s been made aware there is a new problem that has come to light on Apple latest Mac Pro line the 2009 edition. This bug has been confirmed by many Mac Pro 2009 users unfortunately so far apple have kept quiet about if they are going to fix the problem or not. Some users have took back there Mac Pros to the apple Sore with to be told the Mac Pro “is within its Limits”.

The Issue all Mac Pro 2009 Users are facing is a very critical problem for most professional users & consumer users. The problem is related to a majority of simple tasks for any computer including the following:
Playing Audio via any software
Opening Logic
Transferring Data over Firewire
Attaching any external audio device

The above simple tasks results in the following problems:
Higher temperatures on the CPU(s) (From 25c ambient up to 90c)
Performance loss up to 25% (1-2 Cores)
Power consumption increase to 50w

The above tasks can just be run on there own to show the problems listed.
Many users have emailed Steve Jobs directly as of yet no response has been made.

A majority of the Mac Pro 2009 Community have done various tests to rule out & help understand the issues they are facing.

The following has been done:
Tested the MP on Windows – No problems
Checked HW Monitor software – Temperature increase
Checked the MP with Hardware Power consumption tools – increase in wattage
Checked Geek Bench & other various Benchmarks - Performance loss
VMware Fusion bootup tests with audio on/off Firewire on/off – decrease in performance seen.


Here are some tests done by Concorde Rules at Macrumors & Overclockers Forums on he’s Quadcore 2.66 MP

Application 1: VMware Fusion 3.0.1 and Windows 7 x64
Test 1: Booting from Bootcamp partition, timing starts when I hit enter after entering password to unmount the HD. When Windows Live Messenger opens, restart from the Start menu. Once Windows Live Messenger starts again then Quit VMWare fusion, timing stops when VMWare Fusion disappears from the dock.

Run 1: No Audio or FW800 Hard disk
Run 2: Audio + No FW800 Hard disk
Run 3: No Audio + FW800 Hard disk plugged in and copying 172Gb of videos to an internal HD 4.
Run 4: No Audio or FW800 Hard disk (Repeat of Run 1)

Results
1. 4 Minutes 55 Seconds
2. 5 Minutes 30 Seconds (+12.2%)
3. 5 Minutes 38 Seconds (+14.6%)
4. 4 Minutes 56 Seconds (0.3%)



Application 2: Lightroom 2.6
Test: Export 1535 40D JPEG images to 1080p, 100%, sharpened JPEGs. Timing starts when you click "export" Finishes when lightroom says its done

Run 1: No Audio or FW800 Hard disk
Run 2: Audio + No FW800 Hard disk
Run 3: No Audio + FW800 Hard disk plugged in and copying 172Gb of videos to an internal HD 4.
Run 4: No Audio or FW800 Hard disk (Repeat of Run 1)

Results
1. 16 Minutes 36 Seconds
2. 20 Minutes 10 Seconds (+21.5%)
3. (Can't be bothered! Takes too long!)
4. (Can't be bothered! Takes too long!)



Application 3: Quicktime 7
Test: Using 1525 1080p JPEGs, time to encode from hitting "Save" from the export menu into a 1080p, 10Mbit H264 mov file. Timing stops when it finishes.

Run 1: No Audio or FW800 Hard disk
Run 2: Audio + No FW800 Hard disk
Run 3: No Audio + FW800 Hard disk plugged in and copying 172Gb of videos to an internal HD 4.
Run 4: No Audio or FW800 Hard disk (Repeat of Run 1)

Results
1. 7 Minutes 24 Seconds
2. 8 Minutes 33 Seconds (+11.9%)
3. 8 Minutes 39 Seconds (+16.9%)
4. 7 Minutes 22 Seconds (-0.5%)



Application 4: Cinebench in OS X
Test: Cinebench scores

Run 1: No Audio or FW800 Hard disk
Run 2: Audio + No FW800 Hard disk
Run 3: No Audio + FW800 Hard disk plugged in and copying 172Gb of videos to an internal HD 4.
Run 4: No Audio or FW800 Hard disk (Repeat of Run 1)


Results - 1 Thread
1. 3575
2. 2727 (-23.7%)
3. 2716 (-24.0%)
4. 3563 (-0.3%)

Results - 8 Threads
1. 14,468
2. 13,535 (-6.4%)
3. 13,826 (-4.4%)
4. 14,446 (-0.2%)

Results - Open GL
1. 6395
2. 5161 (-19.3%)
3. 5025 (-21.4%)
4. 6230 (-2.6%)


So as we can see, major, major issues...

Image 1:
Here I am copying 172Gb of video files to one of the internal HDs...
http://www.thunder-keep.co.uk/site/m...800hd_test.jpg

Image 2:
Here is the temperatures from tests 1&2.
http://www.thunder-keep.co.uk/site/m..._tests_1&2.jpg

Image 3:
Temperatures at idle 100% and 800% load, with iTunes playing at the end.
http://www.thunder-keep.co.uk/site/m...mperatures.jpg





We hoping for you to show this story in your magazine or at least online so we can get some answer from apple & and a fix.

If you would like more Data, tests & Benchmark visit the below forums

https://forums.macrumors.com/threads/795966/ - Mac Rumors (has the most test data)

http://discussions.apple.com/thread.jspa?threadID=2228019&start=0&tstart=0 - Apple Forum.


Thanks
Samuel G

 

[dryjoy]

Thanks, Smogsy, for taking the time to write that letter. If you want, I could help you to reword it so it reads a bit better, but would have to wait until this evening (UK time - it's morning now).