Well im getting annoyed with it now, i've had it replaced once already for a serious issue (which they didn't actually fix!)
Calling Applecare atm
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Audio Decoding KILLS MacPro (2009) green factor
- Thread starter handheldgames
- Start date
- Sort by reaction score
Well im getting annoyed with it now, i've had it replaced once already for a serious issue (which they didn't actually fix!)
Calling Applecare atm
Well im getting annoyed with it now, i've had it replaced once already for a serious issue (which they didn't actually fix!)
Calling Applecare atm
totally understandable. apple needs an official way to make complaints for these sorts of things. when you call you should be able to ask the assistant for top level help that is urgent. argh
Probably the best option, I don't think it will work though.
I just play the waiting game. Got three years of warranty on my Pro and I listen to music all the time. If the CPUs break, I might get a new machine (2010/2011).
Probably the best option, I don't think it will work though.
I just play the waiting game. Got three years of warranty on my Pro and I listen to music all the time. If the CPUs break, I might get a new machine (2010/2011).
Well I think we should all do the following:
1. Call Applecare, ask to be sent up to top tier support (who are allowed to receive emails!)
2. Send them evidence (in the form of screenshots) of the performance drop, temperature rise and power usage (if you have a power metre that would be better) and include a link to this thread.
3. Include a summary in this email, IE:
Problem: Performance drop, power usage rise, temperature increase
Caused by: Playing Audio, FW800 hard disks connected, Audio output devices over USB/FW
4. Ask that when they get a response from engineering that they give you a call.
Anyway, on a side point, is it confirmed that this issue is NOT in Windows?
I was thinking of making a spreadsheet showing what causes the issue and what doesn't.
CR.
Agreed. I've actually been through steps 1-4 three times now with the same result every time. Apple claims the system is working "within limits". Whatever that means..?? I totally agree that we have to keep at it though. Everyone should just keep hounding Apple about this until they do the right thing and fix it.
I can confirm first hand that this issue does NOT exist in Windows 7. Playing audio or transferring files does not cause the temperature rise and power usage stays consistent unless the processor cores are working hard.
Can I ask about your username? Concorde Rules. Referring to the fast jet?
Yes, it does refer to the supersonic passenger jet
Been lucky enough to work/volunteer on one so thats where the name comes from.I have been given the idea of sending a recorded letter to Apple HQ with the evidence of the problem, what I want in return (An upgrade to a working machine!) and see what happens.
However im no letter writer (being an in training engineer
) so I think we should stick at the problem for a while longer!In a month the 6 core version will come out, lets see if the bug is still there eh...
Edit: Maybe a little database would be a good idea:
Username: Concorde Rules
Purchased Date: 8th April 2009
CPU(s): Single 2.66Ghz Quad W3520
Memory: 6Gb (3x2Gb)
GPU: Apple 4870
Bugs: Audio playback causes 25% performance loss, power usage increase, temperature increase.
If we are able to tabulate all this, along with each persons evidence and send this off to apple (Maybe to s.jobs@apple.com) the overwhelming evidence would kick apple into submission!
Hey - I'd like to join in with this, though I may need guidance as to exactly what evidence to compile. I'm the type of guy who just wants to get on and make some music with my Mac, but I want Apple to do acknowledge the issue and do something.
The thing is, a lot of people don't seem to see this as a real problem. I started a thread in the Mac Music section of the Sound on Sound forum. I hoped I would get a number of people with the issue together there to join any group action, but not one person who actually has a Nehalem Mac Pro there seems to see it as a problem, though there has been plenty of discussion.
This is a forum where people are HARDCORE into their audio, so I thought there might be more of a reaction.
In case you want to see, here is the thread:
http://www.soundonsound.com/forum/showflat.php?Cat=&Number=797280&page=0&view=collapsed&sb=5&o=365&fpart=1#797280
The thing is, a lot of people don't seem to see this as a real problem. I started a thread in the Mac Music section of the Sound on Sound forum. I hoped I would get a number of people with the issue together there to join any group action, but not one person who actually has a Nehalem Mac Pro there seems to see it as a problem, though there has been plenty of discussion.
This is a forum where people are HARDCORE into their audio, so I thought there might be more of a reaction.
In case you want to see, here is the thread:
http://www.soundonsound.com/forum/showflat.php?Cat=&Number=797280&page=0&view=collapsed&sb=5&o=365&fpart=1#797280
Get them to run XBench or something that uses alot of CPU time.
It should show a CPU usage reduction.
Actually, FACT, it WILL show a reduction in performance when playing audio.
I think the 25% drop in CPU performance will have something to show for itself...
Give the details of the temp increases,... by including screen shots of before and after each test (gives idle and ambient pre and post for a good comparison). It would be enough to forward to the engineers, so they can reproduce the issue on their test machines, and go from there.Yes, it does refer to the supersonic passenger jet
I have been given the idea of sending a recorded letter to Apple HQ with the evidence of the problem, what I want in return (An upgrade to a working machine!) and see what happens.
However im no letter writer (being an in training engineer
In a month the 6 core version will come out, lets see if the bug is still there eh...
Edit: Maybe a little database would be a good idea:
Username: Concorde Rules
Purchased Date: 8th April 2009
CPU(s): Single 2.66Ghz Quad W3520
Memory: 6Gb (3x2Gb)
GPU: Apple 4870
Bugs: Audio playback causes 25% performance loss, power usage increase, temperature increase.
If we are able to tabulate all this, along with each persons evidence and send this off to apple (Maybe to s.jobs@apple.com) the overwhelming evidence would kick apple into submission!
Multiply that type of email by as many as possible, and maybe it will attract some attention.
Hmmmmm. How about false advertising? Read the marketing blurbs about the product. Someone mentioned green peace. Aren't those the blokes who always give apple a problem about their non green value?
Yes, it does refer to the supersonic passenger jet
I have been given the idea of sending a recorded letter to Apple HQ with the evidence of the problem, what I want in return (An upgrade to a working machine!) and see what happens.
However im no letter writer (being an in training engineer
In a month the 6 core version will come out, lets see if the bug is still there eh...
Edit: Maybe a little database would be a good idea:
Username: Concorde Rules
Purchased Date: 8th April 2009
CPU(s): Single 2.66Ghz Quad W3520
Memory: 6Gb (3x2Gb)
GPU: Apple 4870
Bugs: Audio playback causes 25% performance loss, power usage increase, temperature increase.
If we are able to tabulate all this, along with each persons evidence and send this off to apple (Maybe to s.jobs@apple.com) the overwhelming evidence would kick apple into submission!
Nice to see another aviation guy in here! I drive Boeing 777s for a living.
Back to the topic at hand... Apple are masters at playing dumb on the phone. In fact, I have three different cases about this issue open. Every time I have called applecare, they've claimed this is the first they have heard of it. I would not be surprised if they have an internal document somewhere strategizing how to deflect responsibility for these type of issues. First make the customer feel alone, then waste copious amounts of their time by making the customer remove non-factory hardware, reinstall the entire OS, etc. Finally, simply claim that the defect is "within limits". It's a painful process and every time I have had to embark down that road over the years, it has always been a struggle.
One of the other interesting phenomenons that I have observed, is the "Apple can do no wrong" mentality. Apple customers often allow Apple to get away with things that normal customers simply would not accept. I have met so many people over the years that have been willing to accept hardware problems, software issues, etc. that the average consumer would not accept. I know about 15 people that had logic board failures on 1st gen Intel iMacs when they were just out of warranty. Even with regard to this issue, I am absolutely amazed at how many people are content with losing 25% performance, drawing 100% power, and cooking their CPUs to simply play audio files or transfer files to an external disk! For some reason, Apple customers tolerate more, and Apple not only created this distortion field, they now rely on it.
I like your idea of sending an email to Apple HQ. Maybe an old fashioned letter in writing will be harder for Apple to ignore. In any case, we all need to keep on Apple. Most people buy Mac Pros to have a powerful workstation that can officially run OSX. This is the kind of issue I would expect with a Hackintosh, and frankly these machines are just too expensive for me to live with that.
I think that would be tough as Apple tends to keep the marketing info quite vague for their products. The only angle we might have, is the fact that Apple advertises this machine as Energy Star 5.0 compliant. If we can somehow find something in these specifications that Apple is not actually complying with as a result of this issue, then we might have a case.
It really shouldn't be that difficult though. We have machines that are not working properly. These machines are under warranty. Apple has an obligation to fix the problem. Maybe they are hoping to deflect and deny until these machines are out of warranty...
Give the details of the temp increases,... by including screen shots of before and after each test (gives idle and ambient pre and post for a good comparison). It would be enough to forward to the engineers, so they can reproduce the issue on their test machines, and go from there.
Multiply that type of email by as many as possible, and maybe it will attract some attention.
I personally have already done just that. After three weeks, the response I got was that it is all "within limits" and there is no problem. I wish I knew what these limits were. What are the limits for performance loss when doing simple tasks? What are the limits for excess electrical load due to OSX not managing the hardware properly? It's all a sad joke..
Take a look at the Thermal Profile in the 5500 series CPU data sheet (prt. I IIRC). (here)
It's in one of them, but I'm fairly sure it's in the first volume. But the profile shows 67C @ 100% core usage.
The Tcase vs Tcore stuff can get really confusing for me. In any case, my heatsink measured temp does approach / exceed 67C if I don't use smcFanControl to increase the fan idle speed.
The way I look at this, the CPU temp increasing is just a symptom of the real problem, which is excessive power consumption when doing simple tasks. I wonder if the CPU performance hit could be the CPU utilizing it's throttling feature..?
Tcase is the temp on the surface of the IHS (copper surface that's been laser engraved). Tcore = actual core temps (aka Tjunction).
In the hottest profile, the Tcase can hit 81C at 95W models (100% CPU utilization btw). That's rather high, and it's only acceptable for a short duration (there's a definition of how much time over a year is allowable under these conditions in the doc).
The acceptable temp for it is 67C max at 100% (24/7 duration).
It seems to me, that what's being reported, is going to exceed the short term limits as well as the normal operational profile.
Thermal probe (i.e. K type thermocouple in a copper block) mounted between the CPU and cooler.
You could try to get it off the edge of the IHS, but I don't know if there's enough room to get a reading, let alone a good one. Worth a shot though, and you could try the edge method with an IR thermometer as well. Again, accuracy could be off a bit (might end up with a board or cooler reading, not the Tcase value).
Tcase is not something I would pay much attention to. Core temps (Tjunction), in particular, the distance from TjMAX is most important. When the core temps hit TjMAX the CPU will start throttling itself to avoid damage.
In the overclocking community, the most accurate temperature reporting software is considered Real Temp... http://www.techpowerup.com/realtemp/
It would be interesting for someone to run this in Windows and observe the idle and load temps (under a stress test) to determine what we could consider as the "trusted actual core temps" at idle/load.
Then the same person could compare those to the idle/load temps reported under Hardware Monitor on OSX, to determine the offset, if any from the "trusted actual core temps" above. BTW, iSTAT only reports Tcase.
In the overclocking community, the most accurate temperature reporting software is considered Real Temp... http://www.techpowerup.com/realtemp/
It would be interesting for someone to run this in Windows and observe the idle and load temps (under a stress test) to determine what we could consider as the "trusted actual core temps" at idle/load.
Then the same person could compare those to the idle/load temps reported under Hardware Monitor on OSX, to determine the offset, if any from the "trusted actual core temps" above. BTW, iSTAT only reports Tcase.
I understand what you're saying, but you use Tcase to determine the calibration for that piece of software. Otherwise, you don't know the actual core temp.
The differential (rise) is still valid, even if not calibrated though, as the system is the same, so variances end up out of the value (simple subtraction).
It's a calibrated value (real, as it's measured).
Although I never actually formally recorded these values, I have done a similar experiment. At idle, I see Core Temps in the low-mid 40s in both OSX (Temperature Monitor) and Windows 7 (Real Temp). The difference is, in Windows 7, playing audio or transferring files to an external drive has almost no effect on the temps. They may rise by a few degrees when I initially open an audio app such as iTunes, but even with audio playing, they quickly retreat to idle levels. In OSX on the other hand, audio playback, flash playback, or transferring files to an external firewire disk cause all 4 core temps to rapidly increase to 85-90C and stay there. A 4 core stress test (CPU Test) causes temps to increase to 95-100C. What is scary, is a stress test with audio playing as well, causes temps to reach near 100C, followed shortly thereafter by a kernel panic. I decided to stop torturing my machine at this point.
Seriously smart move, and if that's the screen shots you sent to Apple, I don't see how they can claim that's within specifications.
That's electromigration level temps, and damages the CPU. Had you continued, you could have outright killed it.
Seriously smart move, and if that's the screen shots you sent to Apple, I don't see how they can claim that's within specifications.
That's electromigration level temps, and damages the CPU. Had you continued, you could have outright killed it.
The worst part is, throughout the whole experiment, the only time the CPU fan actually throttled up, was just prior to the kernel panic. And only up to 1200rpm. Totally inadequate cooling logic.
Again though, I still prefer to look at this whole temperature thing as a symptom of the real problem. Which is grossly high power draw when playing audio, copying files to firewire, etc. I have tried and failed to present this to Apple as a high temp issue.
Nice insights... If you have access to Windows still, can you run a stress test on Windows to see what the core temps hit there?
I thought of doing that but I chickened out. I have really lost faith in Apple's SMC based cooling logic (fans barely speed up) and feel that I may literally cook this thing if I keep pushing it. I will reconsider though...