Mac Mini performance/noise/heat?

[philipma1957]

CPU throttle starts at 105C, right?

Any idea why it isn't throttling while it did a couple of days ago?
(I haven't done any further tests yet.)

I have done tests with 3 different booters.


the toshiba in the lacie it passed a 15 minute test 100 % or 99%

I did it with the samsung 830ssd in the usb3 system shut down.

finally on a longer test with the oem hdd i got throttling.

I even had the test fail twice.

this is not supposed to fail. it is a legit test for high load use.

here is a photo of my room temp.


it took about an hour to get good clean evidence of throttling. my home temp has been 77.9 to 78.7.

my guess is the hardest way to run osx is with usb3. usb2 stressed cpu more then fw800 or fw400.

my tests with really fast throttling the other day were all with usb3.

today I started with t-bolt as the booter.


t-bolt is supposed to use very little cpu so the 15 minute test was too short or it could be that I found out the best way to run handbrake is with t-bolt booters more tests are due.


the second test was usb3 i was annoyed that the first test was no throttle so I knew something was up I went back and saw I had done usb3 for the old test. well no throttle just a crash at the 10 minute mark at 216 f.


I figured let me do a longer test if you run 2 or 3 movies in a que handbrake can run for 3, 4 or more hours.

so I am running a 4 or 6 hour long test throttling started at the 30 minute mark mark a fail at 1 hour.

restarted throttling started at 1 hr 10 minutes a fail at 2 hours. still running I am at the 2 hr 30 min mark.

 

[Snowcake]

I have done tests with 3 different booters.


the toshiba in the lacie it passed a 15 minute test 100 % or 99%

I did it with the samsung 830ssd in the usb3 system shut down.

finally on a longer test with the oem hdd i got throttling.

I even had the test fail twice.

this is not supposed to fail. it is a legit test for high load use.

here is a photo of my room temp.


it took about an hour to get good clean evidence of throttling. my home temp has been 77.9 to 78.7.

my guess is the hardest way to run osx is with usb3. usb2 stressed cpu more then fw800 or fw400.

my tests with really fast throttling the other day were all with usb3.

today I started with t-bolt as the booter.


t-bolt is supposed to use very little cpu so the 15 minute test was too short or it could be that I found out the best way to run handbrake is with t-bolt booters more tests are due.


the second test was usb3 i was annoyed that the first test was no throttle so I knew something was up I went back and saw I had done usb3 for the old test. well no throttle just a crash at the 10 minute mark at 216 f.


I figured let me do a longer test if you run 2 or 3 movies in a que handbrake can run for 3, 4 or more hours.

so I am running a 4 or 6 hour long test throttling started at the 30 minute mark mark a fail at 1 hour.

restarted throttling started at 1 hr 10 minutes a fail at 2 hours. still running I am at the 2 hr 30 min mark.

These errors can also come from bad memory. I would test it atleast 10 passes (10 +-hours) with the memtest86+ bootcd. (can run on macs too)

That way you be sure it is not the cpu. Otherwise it is very bad.

If it is the cpu i am not going to buy the Mac mini. They should have indeed put better cooling in the Mac Mini. I get get the idea that Apple is doing this on purpose. Because higher temps mean shorter lifespan = Money for Apple. Bad internal design, because if you put the black bottom off you have less higher temps than with it on. I would be ashamed if i make something like this. though It looks beautiful. Your room temperature is like 78f (25C) but the Mac Mini spec page says this!:

Operating temperature: 50° to 95° F (10° to 35° C)

They are simply lying because your cpu throttles back at 78f 35f.

 

[philipma1957]

These errors can also come from bad memory. I would test it atleast 10 passes (10 +-hours) with the memtest86+ bootcd. (can run on macs too)

That way you be sure it is not the cpu. Otherwise it is very bad.

If it is the cpu i am not going to buy the Mac mini.

no i did memtest with this ram but for 3 hours not 10.

for me this machine is good ,

but i don't do a lot of handbrake.

this cpu pushed with these tests should pass. it simply is annoying.

I will do this cpu test it has a few hours left.

i will test the ram overnight but I am a bit annoyed . as I upgrade and sell a lot of minis.

the 2009 was very good to sell

the 2010 fff'g broke a lot

the 2011 was very good

I am afraid of this 2012 quad.

I think I may end up only selling 2 core 2012's with 8gb ram and 256gb ssds on the bay.

I get black screens and I managed to get the cpu test to fail twice.

what I am seeing in this test is more throttle and lower temps.

I think the osx is trying to keep it together and learning what % to run cpu at. the cpu is pulling 33 watts vs 37 watts throttle is higher but temps


are lower. well if true fresh installs would be not good. the machine would run the cpu really hot without throttling . look at current numbers

 

[Snowcake]

no i did memtest with this ram but for 3 hours not 10.

for me this machine is good ,

but i don't do a lot of handbrake.

this cpu pushed with these tests should pass. it simply is annoying.

I will do this cpu test it has a few hours left.

i will test the ram overnight but I am a bit annoyed . as I upgrade and sell a lot of minis.

the 2009 was very good to sell

the 2010 fff'g broke a lot

the 2011 was very good

I am afraid of this 2012 quad.

I think I may end up only selling 2 core 2012's with 8gb ram and 256gb ssds on the bay.

I get black screens and I managed to get the cpu test to fail twice.

Sorry but 3 hours is not enough. There were people on windows failing with similar cpu tests on Windows and they tested their ram with memtest86+ and it says ok at about 8 passes but failed at 10 passes!

They replaced the ram and no more errors with the cpu tests.

On internet the experts say to run atleast 10 passes. I personally would run 15 passes.

 

[philipma1957]

Sorry but 3 hours is not enough. There were people on windows failing with similar cpu tests on Windows and they tested their ram with memtest86+ and it says ok at about 8 passes but failed at 10 passes!

They replaced the ram and no more errors with the cpu tests.

On internet the experts say to run atleast 10 passes. I personally would run 15 passes.

not to argue you could be correct.

 

[Snowcake]

not to argue you could be correct.

Also if you have Windows i would run IntelBurnTest or LinX since they are both based on Linpack which Intel is using to test their cpu's.

I personally like IntelBurnTest.
Be sure to choose 20 runs at max memory. It kan take 5 hours. But if the results are good than i think CPUtest for mac is buggy.

Can you post the results from the memtest tomorow?
I am really curious.

 

[propower]

Speaking for myself, I do ProAudio and this is never a 100% load kind of thing. I put up a sample session and loaded up plugins till about 35% constant load on the mini. Temps stabilized ~92degC and the fan was at ~3000. Not amazing, not horrible. But I could work all day there! The mini is 3 feet from my face and the SPL is ~35dBA. My quietest place in my house is 32dBA. In many circles 40dBA is considered a silent pc!

The mini is a super capable sprinter or medium load runner to me. It will work great and pretty darned quiet for all of those uses. Now for taking a proc and hitting it at 100% load for hours..... that is what a MacPro is for! The size of good aftermarket CPU cooler is the size of the MINI !!! Physics. It just can't be done without either large thermal mass or a LOT of airflow. I don't think 1" would come close to cutting it. My MacPro can run all day at 100% and it sounds the same. Temps barely rise. But the CPU heatsink is freakin HUGE!

Interesting testing though, I encourage it go on.

That all said, I totally agree that it is wrong if the Mini actually crashes under that kind of loading!

 

[philipma1957]

Also if you have Windows i would run IntelBurnTest or LinX since they are both based on Linpack which Intel is using to test their cpu's.

I personally like IntelBurnTest.
Be sure to choose 20 runs at max memory. It kan take 5 hours. But if the results are good than i think CPUtest for mac is buggy.

Can you post the results from the memtest tomorow?
I am really curious.

I am going to run rember for 15 passes. this mini has kingston plug n play 1600 2x 4gb sticks = 8gb.

yeah memtest is a bit better but I can't find it right now.

Speaking for myself, I do ProAudio and this is never a 100% load kind of thing. I put up a sample session and loaded up plugins till about 35% constant load on the mini. Temps stabilized ~92degC and the fan was at ~3000. Not amazing, not horrible. But I could work all day there! The mini is 3 feet from my face and the SPL is ~35dBA. My quietest place in my house is 32dBA. In many circles 40dBA is considered a silent pc!

The mini is a super capable sprinter or medium load runner to me. It will work great and pretty darned quiet for all of those uses. Now for taking a proc and hitting it at 100% load for hours..... that is what a MacPro is for! The size of good aftermarket CPU cooler is the size of the MINI !!! Physics. It just can't be done without either large thermal mass or a LOT of airflow. I don't think 1" would come close to cutting it. My MacPro can run all day at 100% and it sounds the same. Temps barely rise. But the CPU heatsink is freakin HUGE!

Interesting testing though, I encourage it go on.

That all said, I totally agree that it is wrong though if the Mini actually crashes under that kind of loading!

yeah like I said i don't push my minis this hard.

but when all is said and done I can push an intel in a pc at 100% for days and keep the temp under 80c while

over clocked using an air based 50 heatsink/fan.

I don't want to push a mini at 100% but it down clocked itself and still crashed.

 

[propower]

yeah like I said i don't push my minis this hard. but when all is said and done I can push an intel in a pc at 100% for days and keep the temp under 80c while
over clocked using an air based 50 heatsink/fan.

I don't want to push a mini at 100% but it down clocked itself and still crashed.

I used to build some ahem"hacks" a while ago that were silent and cool performance monsters so I totally get you (and appreciate all the testing you do!)!

Your last sentence though has my attention - and will keep me watching this thread! Crashing should not be happening (did it do this writing to an internal drive as well??).

-Lee

 

[philipma1957]

I used to build some ahem"hacks" a while ago that were silent and cool performance monsters so I totally get you (and appreciate all the testing you do!)!

Your last sentence though has my attention - and will keep me watching this thread! Crashing should not be happening (did it do this writing to an internal drive as well??).

-Lee

yeah but I had other stuff attached..

i am going to disconnect everything I can and run the tests..

I will only have the internal oem hdd ,the keyboard, the mouse ,the monitor and wired net..


the ram test will be long 8 hrs

the cpu test 4 hour

so the results won't post til thur.


if it passes ram and fails cpu.

i will try a cooling solution.

if i cool the cpu down to 180 f and it runs at 100% its simply lack of cooling

 

[Snowcake]

I am going to run rember for 15 passes. this mini has kingston plug n play 1600 2x 4gb sticks = 8gb.

yeah memtest is a bit better but I can't find it right now.

Well you are not testing 1 or 2 gb, because that is wat Osx has taken.
Memtest uses everything because it is a boot cd.

I found it for you. Here you can download the latest iso: http://www.memtest.org/#downiso
And burning it on a cd using the build in app: Disk Utility, be sure to check: Verify burned data.

I think you are wasting your time if you don't test all of your memory. Sorry.
I also think memtest86+ is better.

Edit: If you want to boot from this CD, start up your Mac Mini and immediately press option and click on the CD.

 

[rezinous]

Mine ramps up quite loud with the fans with any gaming; CS, dota2, Diablo 3, and guild wars 2. 2.6ghz with 16gb ram.

It's on a wooden table, probably heating it up more. Might go get a stand to make it vertical.

 

[dasx]

I have done tests with 3 different booters.


the toshiba in the lacie it passed a 15 minute test 100 % or 99%

I did it with the samsung 830ssd in the usb3 system shut down.

finally on a longer test with the oem hdd i got throttling.

I even had the test fail twice.

this is not supposed to fail. it is a legit test for high load use.

here is a photo of my room temp.


it took about an hour to get good clean evidence of throttling. my home temp has been 77.9 to 78.7.

my guess is the hardest way to run osx is with usb3. usb2 stressed cpu more then fw800 or fw400.

my tests with really fast throttling the other day were all with usb3.

today I started with t-bolt as the booter.


t-bolt is supposed to use very little cpu so the 15 minute test was too short or it could be that I found out the best way to run handbrake is with t-bolt booters more tests are due.


the second test was usb3 i was annoyed that the first test was no throttle so I knew something was up I went back and saw I had done usb3 for the old test. well no throttle just a crash at the 10 minute mark at 216 f.


I figured let me do a longer test if you run 2 or 3 movies in a que handbrake can run for 3, 4 or more hours.

so I am running a 4 or 6 hour long test throttling started at the 30 minute mark mark a fail at 1 hour.

restarted throttling started at 1 hr 10 minutes a fail at 2 hours. still running I am at the 2 hr 30 min mark.

Well. The problem with the SSD there is that USB needs the CPU to function. The USB itself consumes CPU, therefore a completely loaded CPU ends up f-cking up the USB functionality.

As for the oem drive… Can't believe this is actually happening. I gotta do some more tests but I'm full of work right now so can't. Mine is the SSD though.

What i wanna try is:
1. Putting the Mini in a different location, as it seems to be around 3ºC hotter in where I have it now plus the Mini being surrounded by wood on one side and two hard drives in the other.
2. As soon as I receive the bottom cover I ordered off ebay, mod it to get free air intake in the antenna plate area. (I'll install a dust filter). Compare it to running the Mini bottomless and see.

My prediction is the Mini on the table with more ventilated atmosphere will do fine with no throttle, so my intention is to put it right where I have it now with the modded bottom cover. I assume it'll do just fine. As a last resource I can always add a USB and quite fan to move the air around a little bit more and help the Mini cool better.

----------

Be sure to choose 20 runs at max memory. It kan take 5 hours. But if the results are good than i think CPUtest for mac is buggy.

I can tell it's buggy. At least in part. When I wanna stop it, it doesn't. If I click again two things happen randomly:
1. Does nothing again.
2. Stops.

If it did nothing for the second time, a third click on "Stop" will kill the program and pop up an Apple notice saying the app exited and if you wanna ignore or inform blah blah.

So there are some bugs, that's for sure!

----------

I wanna make clear though that my mini doesn't suffer any instability. The thing is it gets too hot therefore the CPU has to throttle itself avoiding it to work at 100%.

Not a big deal as I haven't seen it at 100% except for when I run these benchmarks, lol. But still annoying.

 

[Snowcake]

Well. The problem with the SSD there is that USB needs the CPU to function. The USB itself consumes CPU, therefore a completely loaded CPU ends up f-cking up the USB functionality.

And it should not f-ck up like that. It should fuction normally. These CPU testing programs are not putting a load on the (USB) boot drive but on the CPU or/and RAM. Yeah USB uses CPU but it should definitely not give errors. Testing with USB as a bootdrive is a great way to discover faster errors on already faulty CPU's. On a good CPU it should not give errors because there is nothing wrong with the USB specification.

As for the oem drive… Can't believe this is actually happening. I gotta do some more tests but I'm full of work right now so can't. Mine is the SSD though.

What i wanna try is:
1. Putting the Mini in a different location, as it seems to be around 3ºC hotter in where I have it now plus the Mini being surrounded by wood on one side and two hard drives in the other.
2. As soon as I receive the bottom cover I ordered off ebay, mod it to get free air intake in the antenna plate area. (I'll install a dust filter). Compare it to running the Mini bottomless and see.

My prediction is the Mini on the table with more ventilated atmosphere will do fine with no throttle, so my intention is to put it right where I have it now with the modded bottom cover. I assume it'll do just fine. As a last resource I can always add a USB and quite fan to move the air around a little bit more and help the Mini cool better.

Great that you want to test your Mini Too!
If i was you, i should test the Mac Mini where you have located it right now and with the stock/standard bottom on it. Because you are gonna test the CPU right? Test it with the maximum heat, that way you know for sure it is stable and we know philipma1957 has just a bad cpu and it is not a design error. Your Mac Mini should pass because Apple design team should not be stupid enough to make a Mac Mini which fails at stock configuration/setting; you should use no mods. The maximum room temperature the Mac Mini should operate flawlessy is 35C/95F as stated on the spec page of the Mac Mini.

And if you do that, could you please also test it on osx from a usb boot device?
----------

I can tell it's buggy. At least in part. When I wanna stop it, it doesn't. If I click again two things happen randomly:
1. Does nothing again.
2. Stops.

If it did nothing for the second time, a third click on "Stop" will kill the program and pop up an Apple notice saying the app exited and if you wanna ignore or inform blah blah.

So there are some bugs, that's for sure!

That usually happens when a system is full loaded. Programs tend to not respond. Happens with several CPU testing programs on Windows too.

Edit: I actually want to buy a Mac Mini. But if people are not testing their Mac Mini or are testing and failing, i simply not gonna buy the Mac Mini. Because if one person says that his CPU is unstable at high temperatures it doesn't suprise me and i do not want to take the risk of buying a Mac Mini until other Mac Mini 2012 owners are running it stable at the same settings as Philipma1957.

I also think the high temperatures of the powersupply could make the power supply inefficient. There are testing reports that certain powersupply's are becoming greatly more inefficient at temperatures above 35c/95f. The powersupply of the tested Mac Mini above is 63c/145f!!

Well, bad CPU Vcore = unstable CPU....

Apple... Design for everything. I will praise the day that Apple give OSX in license to other PC manufacturers!

 

[dasx]

Great that you want to test your Mini Too!
If i was you, i should test the Mac Mini where you have located it right now and with the stock/standard bottom on it. Because you are gonna test the CPU right? Test it with the maximum heat, that way you know for sure it is stable and we know philipma1957 has just a bad cpu and it is not a design error. Your Mac Mini should pass because Apple design team should not be stupid enough to make a Mac Mini which fails at stock configuration/setting; you should use no mods. The maximum room temperature the Mac Mini should operate flawlessy is 35C/95F as stated on the spec page of the Mac Mini.

And if you do that, could you please also test it on osx from a usb boot device?

I've already done some tests. Talked about it in a couple of threads. Here and here.

To sum it up, I get no errors in CPUtest. What I get is throttling. The CPU gets to 100%, Core goes up to 105ºC and then the CPU throttles to around 80% and keeps going with core temp around 95ºC. That's pretty much it.

That's why I wanna try all that stuff I said, I wanna check if I'm able to run my CPU at 100% endlessly.

(I'll try to boot from a USB too when I have time. SSD recommended, I guess, right? But now that I think about it I should wait till 10.8.3, as I can't download the Mountain Lion version in the App Store, "No compatible" it says).

Edit: I actually want to buy a Mac Mini. But if people are not testing their Mac Mini or are testing and failing, i simply not gonna buy the Mac Mini. Because if one person says that his CPU is unstable at high temperatures it doesn't suprise me and i do not want to take the risk of buying a Mac Mini until other Mac Mini 2012 owners are running it stable at the same settings as Philipma1957.

I also think the high temperatures of the powersupply could make the power supply inefficient. There are testing reports that certain powersupply's are becoming greatly more inefficient at temperatures above 35c/95f. The powersupply of the tested Mac Mini above is 63c/145f!!

Well, bad CPU Vcore = unstable CPU....

Apple... Design for everything. I will praise the day that Apple give OSX in license to other PC manufacturers!

As I said, mine seems stable (can't really tell for sure, but that program has shown no errors when I've run it).
My power supply right now is at 36ºC. I'm idle but still… looks good.

My only problem is the heat. The core gets to the max and has to throttle itself. A hunch tells me this is due to where I have it located right now. Poor air circulation (well, at least worse than on a table), wood under it with a hard drive sticked to it (external), next to another external drive and touching wood on the other side. So pretty enclosed between stuff.

When I had my 2010 mini there, I could easily see it go up to 90º+. I'd swear I've seen it do a little more but wouldn't be able to say the exact number. My point being, when I put it on the table the other day, I couldn't get it over 85ºC. Also, when taking the bottom off, didn't go over 78ºC. Pretty nice.

And that was a Dual Core Core2Duo. Put a QC i7 there… lol

I'll probably do all the tests of moving the mini as soon as I get the cover and mod it.

 

[scottsjack]

I gues we can just hope the 2013 mini will have a larger case giving it room for perhaps 2 fans and with an external power adapter, a mini wich will make no noise at all regardless of how intensive the tasks are running on the mini.

Not to be too sarcastic about it but it seems that in the world of Jony the most important characteristic of the mini (or any other Apple product) is that it be really thin. The fact that as a desktop computer the mini does not need to be as thin as it is and that another 25cm of height might provide better cooling and room for a discrete GPU seems to have little effect on the thin-sickness that pervades Apple.

I still love my late 2009 mini. It chugs along as a DVR displaying EyeTV at about 50-60c all day long and hasn't missed a beat in three years.

 

[lcseds]

Not to be too sarcastic about it but it seems that in the world of Jony the most important characteristic of the mini (or any other Apple product) is that it be really thin. The fact that as a desktop computer the mini does not need to be as thin as it is and that another 25cm of height might provide better cooling and room for a discrete GPU seems to have little effect on the thin-sickness that pervades Apple.

I still love my late 2009 mini. It chugs along as a DVR displaying EyeTV at about 50-60c all day long and hasn't missed a beat in three years.

No kidding! No upgradeable memory in a desktop iMac? You're kidding! I see the reduction of size for portables, but desktops? They are going a bit too far here....

 

[Darby67]

Bottom line is that the computer, this being the third generation at this thickness, has had no issues with overheating. None. This thread, as will all of it's analogs and predecessors hasn't shown that there is an issue. If CPU Test is the program you rely on day in and day out to have fun on your computer or for work then you are using the wrong program.

If people want a thicker computer, the world is your oyster. Apple is not the only brand. But don't say because it's thin that it's flawed; it isn't.

If you are looking for a the Mini as a serious gaming computer, you know going into it that you made the wrong choice. If listening and believing one persons tinkering and emotion charged conclusions is the truth, then unfortunately you are being led astray.

For everyday computing it's a little champ. For some intensity, it's a little champ. Potentially for some heavy intensity, it's a champ. 48 hours straight, 1080p h.264 HandBrake, it ran like a champ.

 

[MatthewAMEL]

For everyday computing it's a little champ. For some intensity, it's a little champ. Potentially for some heavy intensity, it's a champ. 48 hours straight, 1080p h.264 HandBrake, it ran like a champ.

I have to agree. I ran the AppleTV3 preset @ 98c for more than 20 hours straight. It never throttled.

It's a little nosier than I was expecting under load, but quieter than my i7 iMac under load.

 

[philipma1957]

well I just got back on online.

I had 12 inches of snow in Howell NJ last night and lost power for 22 hours.

So no tests were done. I am going to start the ram test tonight. then run the other tests. I am thinking it could just be my chip.


48 hours of handbrake with 0 throttle. 20 hours of appletv3 no throttle. should be all that is needed to show a decent piece of gear. I have a ton of eyetv not compressed I could encode that for a day or 2 24 7 and see what happens. right now i am running the ram test with no issues.

 

[Darby67]

Have you thought of running multiple instances of yes command instead of CPU Test? Before suggesting the CPU or USB3 is faulty, perhaps the CPU Test app has issues.

well I just got back on online.

I had 12 inches of snow in Howell NJ last night and lost power for 22 hours.

So no tests were done. I am going to start the ram test tonight. then run the other tests. I am thinking it could just be my chip.


48 hours of handbrake with 0 throttle. 20 hours of appletv3 no throttle. should be all that is needed to show a decent piece of gear. I have a ton of eyetv not compressed I could encode that for a day or 2 24 7 and see what happens. right now i am running the ram test with no issues.

 

[philipma1957]

yeah It could be the cpu test has a bug. (i hope)

 

[Snowcake]

I've already done some tests. Talked about it in a couple of threads. Here and here.

To sum it up, I get no errors in CPUtest. What I get is throttling. The CPU gets to 100%, Core goes up to 105ºC and then the CPU throttles to around 80% and keeps going with core temp around 95ºC. That's pretty much it.

That's why I wanna try all that stuff I said, I wanna check if I'm able to run my CPU at 100% endlessly.

(I'll try to boot from a USB too when I have time. SSD recommended, I guess, right? But now that I think about it I should wait till 10.8.3, as I can't download the Mountain Lion version in the App Store, "No compatible" it says).

How long before it starts throttling?? I think that CPUTest only can test a CPU good if it is running at full speed, you should indeed put the 2012 Mac Mini on the table. (More room/breath for the Mini). SSD is good. You could install OSX by using Internet Recovery. I don't know if it works on a second harddrive/sdd.

As I said, mine seems stable (can't really tell for sure, but that program has shown no errors when I've run it).
My power supply right now is at 36ºC. I'm idle but still… looks good.

Do it throttles always when tested with CPUTest? The great thing is that your CPU is stable in throttle mode, but there is always a change it can be unstable if the cpu runs at full speed, since that is a different "power mode".

My only problem is the heat. The core gets to the max and has to throttle itself. A hunch tells me this is due to where I have it located right now. Poor air circulation (well, at least worse than on a table), wood under it with a hard drive sticked to it (external), next to another external drive and touching wood on the other side. So pretty enclosed between stuff.

I am really curious, could you please make a picture of it?

When I had my 2010 mini there, I could easily see it go up to 90º+. I'd swear I've seen it do a little more but wouldn't be able to say the exact number. My point being, when I put it on the table the other day, I couldn't get it over 85ºC. Also, when taking the bottom off, didn't go over 78ºC. Pretty nice.

And that was a Dual Core Core2Duo. Put a QC i7 there… lol

Great discovery!

I'll probably do all the tests of moving the mini as soon as I get the cover and mod it.

Well, since your 2010 Mac Mini dropped temps since you only put it on the table, I think that if you do that right now with your 2012 Mac Mini that you temps wil be below 105C and will not throttle! And that for no modding!

I found another person with a Mac Mini 2012 and he has a higher room temp than you and his quad-cpu did not throttle. I think it has to do where the Mac Mini is located. Has to be that because the Mac Mini's are the same. Well, there is one way to find out if that works!

Bottom line is that the computer, this being the third generation at this thickness, has had no issues with overheating. None. This thread, as will all of it's analogs and predecessors hasn't shown that there is an issue. If CPU Test is the program you rely on day in and day out to have fun on your computer or for work then you are using the wrong program.

No issues with overheating? Try this: http://googleitfor.me/?q=mac+mini+overheating

Temperatures of -+ 100c are considered overheating because the Intel spec page says the tjunction is 105C. Why do you think CPU's are throttling? It has a reason, it becomes simply too hot.

CPUTest is a great program to test your Mac for stability. You Mac has to be stable. If it is not stable and you will go on with it. You have to face the consequences; Because some part(s) (are)is defective and will probably degrade(s) more quickly. I also don't want bizarre errors because of instability... Dude, stability is important. I get the idea that you don't care if your pc is not CPUtest stable and getting weird issues.

If people want a thicker computer, the world is your oyster. Apple is not the only brand. But don't say because it's thin that it's flawed; it isn't.

Other brand? Can you show me a brand who delivers (supported) OSX?
In fact, the biggest problem isn't it being thin, the black bottom cover should have grills! Just like the back as the Macbooks, then the temperature would drop a lot.

If you are looking for a the Mini as a serious gaming computer, you know going into it that you made the wrong choice. If listening and believing one persons tinkering and emotion charged conclusions is the truth, then unfortunately you are being led astray.

You are right. The Mac Mini is not a serious gaming computer. Unless people have a little budget and only want a Mac. (Those people exist!(Not me) Then they have no choice. There are indeed better pc's out there for gaming but the Mac Mini should run games just as good as other computers in terms of stability and temps. Enough people who sometimes play games on their Mac Mini.

For everyday computing it's a little champ. For some intensity, it's a little champ. Potentially for some heavy intensity, it's a champ. 48 hours straight, 1080p h.264 HandBrake, it ran like a champ.

Happy to hear that your Mac Mini does it job.

well I just got back on online.

I had 12 inches of snow in Howell NJ last night and lost power for 22 hours.

So no tests were done. I am going to start the ram test tonight. then run the other tests. I am thinking it could just be my chip.

Are you sure to check it with Memtest86+? Because if you test it on OSX you will not test 2GB!(Simply because OSX is using it.) Also Memtest86+ is better. Download the latest version here:
http://www.memtest.org/#downiso
You can burning it on a cd using the build in app: Disk Utility, be sure to check: Verify burned data. If you want to boot from this CD, start up your Mac Mini and immediately press option and click on the CD.

What is the best way to test my system for maximum heat and stability?

For CPU:
Use IntelBurnTest under Windows. For best results use windows 64 bit and be sure that IntelBurnTest uses the 64-bit mode, this way it can use all the ram.


First let it run 2 passes on standard settings to see how hot your cpu get. If everything is ok, you could use the (best) settings below:

Times to run: 20 (Circa 4 hour)
Be sure to check: Output results to results.log
Set stress level to Maximum. This way it uses all the ram for better detection of CPU errors.
Set threads to 8 if you have a quadcore, or 4 if you have a dualcore. (Intel has hyperthreading)
In the program, Right click on start and select Xtreme test mode. It may look like your Mac doesn't respond anymore if you start testing with Xtreme test mode, but that is normal. You CPU is loaded to the top.
IntelBurnTest is using Linpack. Intel is using Linpack for testing their CPU's!

Good stuff. The best out there.


For OSX:
You can use CPUTest or use the yes command. Instructions in the link below:
http://macfidelity.de/2009/05/10/mac-how-to-stress-test-your-cpu-in-mac-os-x/index.html

I prefer CPUtest on OSX, but if you have somehow trouble with it you could try to duplicate this with the yes command.




For RAM:

The best way to test the RAM is to use the bootcd of Memtest86+. Unlike memtestprograms that run on OSX or Windows, the bootcd Memtest86+ is testing ALL the memory.
(You should know that Windows/OSX are using a lot of ram, -+2gb, That ram cannot be tested by memtestprograms.)

It can be useless to test with memtestprograms on Windows or Mac because if the problem is in the space that the os has taken you won't find it.


Memtest86+ should run atleast 10 passes. (Circa 10 hour)

You can download the latest iso below:
http://www.memtest.org/#downiso
You can burning it on a cd using the build in app: Disk Utility, be sure to check: Verify burned data. If you want to boot from this CD, start up your Mac Mini and immediately holding the option key and wait until you see icons and then click on the CD.

 

[dasx]

How long before it starts throttling?? I think that CPUTest only can test a CPU good if it is running at full speed, you should indeed put the 2012 Mac Mini on the table. (More room/breath for the Mini). SSD is good. You could install OSX by using Internet Recovery. I don't know if it works on a second harddrive/sdd.


Do it throttles always when tested with CPUTest? The great thing is that your CPU is stable in throttle mode, but there is always a change it can be unstable if the cpu runs at full speed, since that is a different "power mode".

OK. Did a quick test.
Put CPU in Primenet, 8 threads, unless run. Hit start. CPU immediately went to 100% CPU use.

After 1 minute CPU was at 98ºC approx, fans still at 1800rpm (lol).
After 1 minute and a half, it was at 103ºC and fans started to spin up. It got to 5100rpm and temps started to drop a little bit. (CPU still at 100%, no throttle at all).
After 5 mins, temps had droped to 96-98ºC and fan had slowed down a little bit to 4400rpm. It's been running like that until temps had gone up a little bit (97-99ºC) and fans went up again to 4800rpm. Temps then stayed in the 96-99ºC bracket.

The thing stayed like that until min15 of CPU test, no throttle or errors at all.

Gotta say room temp is a little lower right now. Maybe 20-21ºC. Also house's heating is not on yet (obviously, lol), and I got one room window slightly open so there's ought to be a better ventilation.

Pretty happy to see it do just fine. Only problem I see even if if doesn't throttle is fan noise. What I noticed with the 2010 Mini, aside from temps being lower with no bottom, is that fans needed to run much slower in order to keep temps.
So hopefully when I mod the bottom cover I'll be able to get lower fan speeds therefore less noise at 100% of CPU use.


OHH!!! And something else. This CPUtest is buggy all the way. Remember what I said when closing it? Well, I had only testes with my 2010 Mini which has just 2 threads.

When stopping it in this 2012 Mini this happened:

1. "Stop". CPU use goes down to 88%. I guess it stopped 1 thread.
2. "Stop" again. CPU use goes down to around 75%.
3. "Stop" again and app crashes. I ignore the error message and CPU stays at 75% and temps pretty high.

Checked top processes and they there still were 6 processes consuming 100% of the CPU. Saw they were are called glucas, just opened terminal and ran:

$ killall glucas

CPU went down to 0% and temps dropped to regular idle temps.

I spent a lot of years in Linux so I don't care messing around with Terminal, but hey, not everyone would've known how to stop those processes!!

Well, since your 2010 Mac Mini dropped temps since you only put it on the table, I think that if you do that right now with your 2012 Mac Mini that you temps wil be below 105C and will not throttle! And that for no modding!

Yeah, most definitely, but I wanna have it were I have it right now due to space management and real desk estate. That's why I'll try this mod. Bottom cover was like $20 so worth a try imo.

I found another person with a Mac Mini 2012 and he has a higher room temp than you and his quad-cpu did not throttle. I think it has to do where the Mac Mini is located. Has to be that because the Mac Mini's are the same. Well, there is one way to find out if that works!

Sure. When room temp is slightly high I guess where I have it now doesn't have all that good air renewal and the atmosphere of the Mini gets a little hot and the air intake gets hotter air than the overall room.

 

[philipma1957]

OK. Did a quick test.
Put CPU in Primenet, 8 threads, unless run. Hit start. CPU immediately went to 100% CPU use.

After 1 minute CPU was at 98ºC approx, fans still at 1800rpm (lol).
After 1 minute and a half, it was at 103ºC and fans started to spin up. It got to 5100rpm and temps started to drop a little bit. (CPU still at 100%, no throttle at all).
After 5 mins, temps had droped to 96-98ºC and fan had slowed down a little bit to 4400rpm. It's been running like that until temps had gone up a little bit (97-99ºC) and fans went up again to 4800rpm. Temps then stayed in the 96-99ºC bracket.

The thing stayed like that until min15 of CPU test, no throttle or errors at all.

Gotta say room temp is a little lower right now. Maybe 20-21ºC. Also house's heating is not on yet (obviously, lol), and I got one room window slightly open so there's ought to be a better ventilation.

Pretty happy to see it do just fine. Only problem I see even if if doesn't throttle is fan noise. What I noticed with the 2010 Mini, aside from temps being lower with no bottom, is that fans needed to run much slower in order to keep temps.
So hopefully when I mod the bottom cover I'll be able to get lower fan speeds therefore less noise at 100% of CPU use.


OHH!!! And something else. This CPUtest is buggy all the way. Remember what I said when closing it? Well, I had only testes with my 2010 Mini which has just 2 threads.

When stopping it in this 2012 Mini this happened:

1. "Stop". CPU use goes down to 88%. I guess it stopped 1 thread.
2. "Stop" again. CPU use goes down to around 75%.
3. "Stop" again and app crashes. I ignore the error message and CPU stays at 75% and temps pretty high.

Checked top processes and they there still were 6 processes consuming 100% of the CPU. Saw they were are called glucas, just opened terminal and ran:

$ killall glucas

CPU went down to 0% and temps dropped to regular idle temps.

I spent a lot of years in Linux so I don't care messing around with Terminal, but hey, not everyone would've known how to stop those processes!!




Yeah, most definitely, but I wanna have it were I have it right now due to space management and real desk estate. That's why I'll try this mod. Bottom cover was like $20 so worth a try imo.




Sure. When room temp is slightly high I guess where I have it now doesn't have all that good air renewal and the atmosphere of the Mini gets a little hot and the air intake gets hotter air than the overall room.

I have come to the conclusion that cpu test has many issues. my ram tests are fine.


I do have a question do you get the black screen issue?

I wonder if this has a link to the throttle or none throttle at 100C-105c.

We have posters that ran 48 hours of handbrake with no issues. I wonder if that poster has 0 black screen problems. I wonder if 0 black screens means no throttle.

Anyone that has used intel cpus for oc in pc's knows that some oc better then others the nickname is "a golden chip". It can oc higher with less power then normal.

So if you got a normal chip the mini is just short of proper cooling. And you get more black screens.

If you got a better then spec cpu you don't throttle and you don't get black screens.


Well back to testing I am going to run a big handbrake on some eyetv later today. maybe I get 100% or 99% right across the board. I will post on this sat morning.