Mac Mini performance/noise/heat?

[dasx]

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.

No black screens here. I'm using an Apple Cinema Display though, which is connected through the Thunderbolt port. No one seems to be getting the black screens when using the thunderbolt port.

 

[Snowcake]

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.

I also read that people just booted their Mac Mini and got immediately black screens, CPU is then definitely not hot. Also as dasx stated, it only happens if hdmi is involved.

What for program did you use to test your RAM?? And how long? (how many passes?)

 

[BigMa]

SPL for the stock HDD model (i7)

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!

Am I correct assuming you provided SPL data for the SSD-equipped mini?
If this is the case indeed, could you please estimate SPL for the stock HDD model (i7) under similar load, if possible?

Could you guys please take SPL measurements on your minis and get back with the results?
I am sure it will be of interest for potential byers like myself!

 

[propower]

Am I correct assuming you provided SPL data for the SSD-equipped mini?
If this is the case indeed, could you please estimate SPL for the stock HDD model (i7) under similar load, if possible?

Could you guys please take SPL measurements on your minis and get back with the results?
I am sure it will be of interest for potential byers like myself!

Yes, SPL data using American Recorder SPL-8810 with "A" weighting active. I only have the one mini so can't estimate Hard Drive noise ... but I hate the sound of spinners... and fans... and that annoying ringing that follows me around -- for reference my 2009 MacPro with 4 internal HD was around 40dBA and much more annoying


Assuming a close micing situation where the SPL of the instrument is 85dBA+ a signal to noise of >>50dB should be readily doable. Only in an extremely quiet passage would this ever be a problem. At that point I would put on an analog tape simulation effect and call it hiss

 

[BigMa]

Thank you!

Anybody else has SPL data for minis?

 

[philipma1957]

Anybody else has SPL data for minis?

how about the fan running full speed 5500 rpm?

56.6 db at about a foot a weighted with a sper scientific meter.


about 47.1 db at about 4 feet a weighted.

this is a 2012 quad 2.3 running at 100% cpu load doing a prime95 test.

put the black plate back 41.4 db and four feet and 51.6 db at 1 foot

 

[propower]

2ft, max fans, 47dBA. Way too loud for recording nearby.
2ft, 3750 fan 42dBA. Still too loud for recording
2ft 2800 fan, 37dBA, could be OK

 

[MJL]

The simple fact that the fan has been modified indicates that the 2011 models are running too hot. This is also confirmed by the higher number of logic board failures in the 2011 model (especially the mid model).

I just wonder if the 2012 fan has the same mounting layout and can be retrofitted to the 2010 and 2011 models and in that way improve the cooling performance of those models.

The Intel NUC was rumored to be released in October but have not seen it yet - I wonder if they have heat issues and what temperatures they'll be measuring. Habey has announced a fanless Ivy bridge computer.

http://www.xbitlabs.com/news/other/...Like_Fanless_Quad_Core_Personal_Computer.html

 

[philipma1957]

The simple fact that the fan has been modified indicates that the 2011 models are running too hot. This is also confirmed by the higher number of logic board failures in the 2011 model (especially the mid model).

I just wonder if the 2012 fan has the same mounting layout and can be retrofitted to the 2010 and 2011 models and in that way improve the cooling performance of those models.

The Intel NUC was rumored to be released in October but have not seen it yet - I wonder if they have heat issues and what temperatures they'll be measuring. Habey has announced a fanless Ivy bridge computer.

http://www.xbitlabs.com/news/other/...Like_Fanless_Quad_Core_Personal_Computer.html

if I get time I will test the fan switch out.

 

[motrek]

This thread is confusing me.

A lot of people talk about "throttling"--can you explain what you mean here? Do you mean the turbo boost decreases due to temperature? This seems like it would be almost imperceptible since when all four cores are loaded the amount the CPU can turbo boost is usually very small.

 

[philipma1957]

This thread is confusing me.

A lot of people talk about "throttling"--can you explain what you mean here? Do you mean the turbo boost decreases due to temperature? This seems like it would be almost imperceptible since when all four cores are loaded the amount the CPU can turbo boost is usually very small.

NO throttling is cpu use drop-age .

this test is designed to max the cpus cores at 100% .

so a quad has 8 cores and 8 x 100% is 800%.

Also pulls 37 watts of power or so.

if the cpu gets real hot the use may drop to 75-80% across the board say 640% and watts may drop to 33 or 32 watts.

when you are doing 8 core work the max is not 3.3.

when you do 1 core or 2 core work you jump to 3.3 from 2.3 .

8 core work may be 2.8 or maybe 2.9 when turbo boosted.


turbo boost. has a range 3.3 for 1 core is max

2.9 may be max for 8 cores


this whole exercise is most important for handbrake users.

as it uses all 8 cores at 100%.


so if you drop to 80% across the board to save the cpu from over heating. you are just that 80% of top speed. which means time wise a 8 hour job at 100% is slower around 10 hours.


the testing was done for handbrake users and I had never even reached the handbrake part of the test!


As my cpu is not functioning properly why bother. When the new one comes i will test again .

 

[hr10]

Everyone is saying how quiet this thing is
Can anyone of u hear hard drive loading sound
Is my hard drive alright if I hear it?

 

[fa8362]

Everyone is saying how quiet this thing is
Can anyone of u hear hard drive loading sound
Is my hard drive alright if I hear it?

Hard drives are mechanical and make noise.

 

[motrek]

NO throttling is cpu use drop-age .

this test is designed to max the cpus cores at 100% .

so a quad has 8 cores and 8 x 100% is 800%.

Also pulls 37 watts of power or so.

if the cpu gets real hot the use may drop to 75-80% across the board say 640% and watts may drop to 33 or 32 watts.
...

I suppose it's possible for the OS to change its process scheduling algorithm to eliminate the load on one or more cores depending on CPU temperature but I think it's extremely unlikely.

From your previous posts I infer that your "test" is running Handbrake, which is not designed to load the CPU to 100%, it's designed to encode video. At times it may use most (or all) of the CPU but at other times it doesn't. Certainly Handbrake doesn't use 100% of the CPU all the time on my own computer, which is definitely not thermally constrained.

If you really want to stress test your CPU I would suggest using software that's well understood to max out CPU usage, e.g., Prime95:

http://www.mersenne.org/freesoft/

 

[philipma1957]

I suppose it's possible for the OS to change its process scheduling algorithm to eliminate the load on one or more cores depending on CPU temperature but I think it's extremely unlikely.

From your previous posts I infer that your "test" is running Handbrake, which is not designed to load the CPU to 100%, it's designed to encode video. At times it may use most (or all) of the CPU but at other times it doesn't. Certainly Handbrake doesn't use 100% of the CPU all the time on my own computer, which is definitely not thermally constrained.

If you really want to stress test your CPU I would suggest using software that's well understood to max out CPU usage, e.g., Prime95:

http://www.mersenne.org/freesoft/

After I failed the cpu test I ran the mini with prime95 and I acheived the rounding errors. In fact I encourage any and all 2012 2.3 quad users to run prime95. I would like to know how many pass the test.

 

[dasx]

I suppose it's possible for the OS to change its process scheduling algorithm to eliminate the load on one or more cores depending on CPU temperature but I think it's extremely unlikely.

From your previous posts I infer that your "test" is running Handbrake, which is not designed to load the CPU to 100%, it's designed to encode video. At times it may use most (or all) of the CPU but at other times it doesn't. Certainly Handbrake doesn't use 100% of the CPU all the time on my own computer, which is definitely not thermally constrained.

If you really want to stress test your CPU I would suggest using software that's well understood to max out CPU usage, e.g., Prime95:

http://www.mersenne.org/freesoft/

It happens. I got my mini in not an ideal location right now (bad air ventilation and renewal) and when running CPUTest (kind of Prime95) I saw it go at 100% and, a couple of times, right after hitting 104ºC (219ºF) the CPU throttled to arund 80-85% and temps dropped to around 95ºC.

It just happened a couple of times, but still… I saw it.

 

[philipma1957]

It happens. I got my mini in not an ideal location right now (bad air ventilation and renewal) and when running CPUTest (kind of Prime95) I saw it go at 100% and, a couple of times, right after hitting 104ºC (219ºF) the CPU throttled to arund 80-85% and temps dropped to around 95ºC.

It just happened a couple of times, but still… I saw it.

do me a favor load the prime95 and see if the rounding errors show up. I have been doing some research the error could be in the mobo not the cpu and it is possible that many quads have the problem. the prime95 will show the error in less then an hour. you can pick 4 cores not all 8 this keeps the mini cooler. i had the errors in 3 of 8 cores and in 4 of 8 cores.

I always had the error in cores 1 and 2. which means the first cores used in the cpu. When sandy bridge came out there was a delay due to chipset errors in mobos. I am not saying all 2.3 quads will fail but in an hours test you could find out.

 

[dasx]

do me a favor load the prime95 and see if the rounding errors show up. I have been doing some research the error could be in the mobo not the cpu and it is possible that many quads have the problem. the prime95 will show the error in less then an hour. you can pick 4 cores not all 8 this keeps the mini cooler. i had the errors in 3 of 8 cores and in 4 of 8 cores.

I always had the error in cores 1 and 2. which means the first cores used in the cpu. When sandy bridge came out there was a delay due to chipset errors in mobos. I am not saying all 2.3 quads will fail but in an hours test you could find out.

Will try it tonight.

 

[motrek]

After I failed the cpu test I ran the mini with prime95 and I acheived the rounding errors. In fact I encourage any and all 2012 2.3 quad users to run prime95. I would like to know how many pass the test.

Your Mini is broken and you should return/exchange it. Computing is something that either works or it doesn't. There isn't such a thing as "sort of" working. What if you were using your Mini for something where accuracy was critical, like protein folding? It would be irresponsible of Apple to engineer a device that didn't work correctly under specified conditions (i.e., with appropriate space for ventilation and under X degrees ambient temperature) and I can't believe they've done so with the Mini, so I have to assume that if you are seeing problems it's due to a manufacturing defect.

To my knowledge it is actually almost impossible to get an Intel CPU to overheat anyway. At least with the Pentium 4, once they get close to their maximum temperature they start multiplying the duty cycle of the clock signal so the CPU will run MUCH slower (down into the low megahertz range if necesary) but it won't overheat.

So if you are seeing calculation errors I suspect a manufacturing defect e.g. bad RAM. Have you done a RAM test? Although you shouldn't have to--this should be a simple matter of taking the computer back to the Apple Store and showing them that it's failing Prime95 runs.

 

[motrek]

It happens. I got my mini in not an ideal location right now (bad air ventilation and renewal) and when running CPUTest (kind of Prime95) I saw it go at 100% and, a couple of times, right after hitting 104ºC (219ºF) the CPU throttled to arund 80-85% and temps dropped to around 95ºC.

It just happened a couple of times, but still… I saw it.

The more I think about this, the less likely I think it is that the Mini (or any Apple product) relies on the OS X process scheduler to keep it from overheating. You can run other operating systems on these computers (Windows, anyway) and certainly Windows is not designed to keep an Apple Mac Mini from overheating.

You might have seen Prime95 use less CPU after the temps went up but there must be another explanation. Are you looking at total CPU usage (i.e., green and red boxes) or just the Prime95 percentage? Because maybe the CPU was being shared with the OS or some other software.

 

[dasx]

do me a favor load the prime95 and see if the rounding errors show up.

Will start one in 10 minutes approx. Will load all 8 threads, if throttle appears due to temp I'll try just loading 4.

 

[dasx]

The more I think about this, the less likely I think it is that the Mini (or any Apple product) relies on the OS X process scheduler to keep it from overheating. You can run other operating systems on these computers (Windows, anyway) and certainly Windows is not designed to keep an Apple Mac Mini from overheating.

You might have seen Prime95 use less CPU after the temps went up but there must be another explanation. Are you looking at total CPU usage (i.e., green and red boxes) or just the Prime95 percentage? Because maybe the CPU was being shared with the OS or some other software.

I haven't actually tried Prime95 yet. This happened to me in Handbrake once and once -I think- with CPUTest.

 

[dasx]

do me a favor load the prime95 and see if the rounding errors show up. I have been doing some research the error could be in the mobo not the cpu and it is possible that many quads have the problem. the prime95 will show the error in less then an hour. you can pick 4 cores not all 8 this keeps the mini cooler. i had the errors in 3 of 8 cores and in 4 of 8 cores.

I always had the error in cores 1 and 2. which means the first cores used in the cpu. When sandy bridge came out there was a delay due to chipset errors in mobos. I am not saying all 2.3 quads will fail but in an hours test you could find out.

OK, ran it for 15 minutes. 8 threads. No problems at all.
Temp stable at around 98-100ºC (208-212ºF). No CPU throttle, all time at 100%. Fans at 5500rpm all the time. Incredibly loud, sooo noisy.

Ran Torture Test just for CPU. The way I used to when overclocking windows based machines.

 

[philipma1957]

Your Mini is broken and you should return/exchange it. Computing is something that either works or it doesn't. There isn't such a thing as "sort of" working. What if you were using your Mini for something where accuracy was critical, like protein folding? It would be irresponsible of Apple to engineer a device that didn't work correctly under specified conditions (i.e., with appropriate space for ventilation and under X degrees ambient temperature) and I can't believe they've done so with the Mini, so I have to assume that if you are seeing problems it's due to a manufacturing defect.

To my knowledge it is actually almost impossible to get an Intel CPU to overheat anyway. At least with the Pentium 4, once they get close to their maximum temperature they start multiplying the duty cycle of the clock signal so the CPU will run MUCH slower (down into the low megahertz range if necesary) but it won't overheat.

So if you are seeing calculation errors I suspect a manufacturing defect e.g. bad RAM. Have you done a RAM test? Although you shouldn't have to--this should be a simple matter of taking the computer back to the Apple Store and showing them that it's failing Prime95 runs.

Yeah the cpu throttled back watts at the max on the quad were 37 + 4 or 41 when % dropped they went to 33 + 4 = 37

ran the gold standard memtest86+ ram is good.

https://forums.macrumors.com/threads/1488392/


the blue screen is the ram test for the quad.

my guess is the mobo has a voltage problem. rounding errors can show in prime95 due to voltage not supporting the clock say 3.0 mhz need 1.1 volts and gets 1.09 volts prime95 can show a rounding error. Since i can't go to the bios the machine has been returned to jr.com they paid for ups return and will send a new quad.


Frankly I am thinking I will end up with 2 core machines and not have quads. The cooling is exactly the same for both machines. same heatsink same fan. one cpu pulls almost 2x the watts at max. quads need better cooling. just 1.25 cm or just a half inch height in these minis would have allowed for so much more quality in cooling. oh well



dasx thanks that is the test I failed so I had a bad mobo or cpu.

your temps were less then mine.

still high .

my prediction is long term life for these quad if pushed can't be good. But that is what apple care is for. I just shipped my mini back I will make due with the 2.5 dual core for now.

 

[scottsjack]

just 1.25 cm or just a half inch height in these minis would have allowed for so much more quality in cooling. oh well

I'm totally with you there. Apple's current perversion of thinness that undermines performance and durability is just something I cannot get my logic around. Would I pay another $25 for an additional 25mm of height and a larger fan on a mini? You bet.