Loaded Mini - Actual tests. HOT CPU

[dasx]

Hey there. I received today my Mini. i7 2.6GHz with 4GB of RAM (which I upgraded to 16GB) and the 256GB SSD.

Didn't really wanna just run Geekbench as I can go and check results on their site, and that number doesn't tell me anything, so prepared a couple of -silly- tests of stuff I do everyday to compare processing times.

Machines in comparison:

Mac Mini 2010: Core2Duo 2.4GHz, 4GB RAM, 500GB HDD.
Macbook Air 2012: i5 1.8GHz, 4GB RAM, 128GB SSD.
Mac Mini 2012: i7 2.6GHz, 16GB RAM, 256GB SSD.

Test 1: HANDBRAKE
Converting a 1 minute 50 seconds long HD file to the preset Apple TV3. Video is MKV, h264, DTS Audio and one subtitle track.

Mini 2010: 7m20s
MBA 2012: 3m49s
Mini 2012: 1m40s

Well, times sum it all I think. Amazing performance.
CPUs got damn hot. Gotta say I was using iStat Menus to monitor them and I don't know how precise these are, because a lot of readings were down to 0 (voltages and some temps) and it was only showing 2 CPUs... it looked legit, but who knows). The readings for these 2 CPUs went up to 105ºC (221ºF).

Test 2: SUBLER
Converting an 8GB HD movie. Video is, again, MKV, h264, DTS Audio and one subtitle track.

Mini 2010: 12m28s
MBA 2012: 2m06s
Mini 2012: 2m12s

I guess here the SSD is the trick. 6 seconds isn't really noticeable so I guess Subler doesn't really know how to use the extra power of the CPU.

Test 3: COMPRESSION
"Raring" a 1GB file into 10 files of 100MB each using SimplyRAR.

Setting: Fastest
Mini 2010: 2m07s
MBA 2012: 1m03s
Mini 2012: 48s

Setting: Normal
Mini 2010: 15m46s
MBA 2012: 10m44s
Mini 2012: 9m57s

Test 4: UNCOMPRESSION
"UnRaring" a 4GB file which was previously "rared" into 40 files of 100MB each.

Mini 2010: 5m32s
MBA 2012: 1m40s
Mini 2012: 1m19s

SSD rocks. Also notice that the Air uses a Toshiba one, which only achieve 270MB/s for write while the Mini's Samsung drive can do 410MB/s.

Test 5: CALCULUS (My favorite)
I'm working on a project at work in which we programmed this code that has to do 'something'. Can't really say what we want it for, so just to clarify this is a CPU starving process that has to calculate different iterations of this 'something'.

At work we usually do quadrillions of iterations, but of course we got a magical machine there that I'm in love with. LOL

Here I set it to do 30.000 billions of iterations.

Mini 2010: 24m10s
MBA 2012: 14m26s
Mini 2012: 7m23s

As I said, this program is only CPU. And It's amazing to see the difference.
CPUs got to around 98ºC (208ºF) during this test.

I know these are a bit silly tests, but this is the way I found to see how much of an improvement I'd actually notice from the upgrade: Doing regular tasks I do.

Cheers!

[flatfoot99]

thanks!

[Darby67]

I wouldn't say silly at all, if these are real world for you, then they are much better than GB or other metrics. Enjoy!

[glitch44]

did your fan kick on high too?

[martinm0]

I had purchased one of the quad i7 Mini servers last year when they came out solely to run Handbrake 24/7. I too noticed how hot the CPU would get and the fans would run full blast. I returned it and went the refurb base model iMac and its runs much cooler/quieter when doing encodes.

I was hoping the heat issue was resolved with the different Intel chipset, but it seems like quad i7 CPUs are just a bit too hot to run all the time.

[calvol]

I guess the heat problems arising from TurboBoost in Sandy Bridge were not solved in Ivy Bridge, actually maybe even worse with Ivy's die shrink. Would hope for better cooling, but only so much you can do with the Mini form factor. Maybe they went to the asymmetrical fan that's in the rMBP, although may not help with flow, only quietness. Well, this could also be a result of global warming too, as Al Gore said, it's hot out!

[dasx]

Originally Posted by glitch44

did your fan kick on high too?

Actually... pretty high. I was monitoring the speed of it too and it went up to almost 4000rpm during the handbrake test. It still had a margin till the maximum speed of 5500rpm, but the damn fan was very loud.

Originally Posted by martinm0

I had purchased one of the quad i7 Mini servers last year when they came out solely to run Handbrake 24/7. I too noticed how hot the CPU would get and the fans would run full blast. I returned it and went the refurb base model iMac and its runs much cooler/quieter when doing encodes.

I was hoping the heat issue was resolved with the different Intel chipset, but it seems like quad i7 CPUs are just a bit too hot to run all the time.

Originally Posted by calvol

I guess the heat problems arising from TurboBoost in Sandy Bridge were not solved in Ivy Bridge, actually maybe even worse with Ivy's die shrink. Would hope for better cooling, but only so much you can do with the Mini form factor. Maybe they went to the asymmetrical fan that's in the rMBP, although may not help with flow, only quietness. Well, this could also be a result of global warming too, as Al Gore said, it's hot out!

Yup. I didn't do handbrake for 24/7 except for a couple of times with my old mini, but now that I've seen how hot this gets and how loud the fans sound... I probably never will.

When I use handbrake is usually to convert a file I can't convert with Subler, which isn't very often as I try to make/get everything in h264 already so in this case it's good to have such a powerful CPU.

But for 24/7... too loud. I wouldn't say too hot, as my old mini got to 91-94ºC during Handbrake, which isn't that far away from 105ºC, problem are the fans. Way too loud.

[creativet]

I bought the mid range mini 2.3 i7 and at the moment sitting at 4GB of RAM. (Waiting for the 16) to use just for handbrake encodes specific. I ran a couple of 720p encodes and noticed that the mini did get a little warm, but nothing offensive or bothering. I can't say that I actually heard the fans disturbing me at all with being loud. (It sits right in front of me on the desk.)

Nevertheless, I decided since I was going to be using my mini as an encoder that I should look into some methods of keeping it cool. So I picked up a Notepal I 100 (http://www.coolermaster-usa.com/prod...roduct_id=3107) for like $15.00. Looks great and the mini sits right overtop the fan. This works really well even if I didn't have the Notepals' fan on, as it keeps the mini raised from the surface of the desk.

I'm now doing my third encode in a row without break of 1080p using handbrakes AppleTV 3 preset. I also installed iStat Menus to watch the CPU usage, and also the internal temp, plus fan exhaust rpm. It's taking roughly 1hr 45mins to complete the encode. It is using all 8 cores at pretty much full (System Activity monitor says handbrake using 776% - 790% CPU. The internal tempature hasn't gone above 94c and the exhaust rpm is around 3400.

So really, not that bad. I can put my hand on the mini and barely feel any heat at all. If I find it starts running way too hot, I'll just limit the CPU cores used via handbrake advance setting. My only real issue is with Handbrake, it's gobbling up every last bit of RAM it can get!

[dasx]

Originally Posted by creativet

I bought the mid range mini 2.3 i7 and at the moment sitting at 4GB of RAM. (Waiting for the 16) to use just for handbrake encodes specific. I ran a couple of 720p encodes and noticed that the mini did get a little warm, but nothing offensive or bothering. I can't say that I actually heard the fans disturbing me at all with being loud. (It sits right in front of me on the desk.)

Nevertheless, I decided since I was going to be using my mini as an encoder that I should look into some methods of keeping it cool. So I picked up a Notepal I 100 (http://www.coolermaster-usa.com/prod...roduct_id=3107) for like $15.00. Looks great and the mini sits right overtop the fan. This works really well even if I didn't have the Notepals' fan on, as it keeps the mini raised from the surface of the desk.

I'm now doing my third encode in a row without break of 1080p using handbrakes AppleTV 3 preset. I also installed iStat Menus to watch the CPU usage, and also the internal temp, plus fan exhaust rpm. It's taking roughly 1hr 45mins to complete the encode. It is using all 8 cores at pretty much full (System Activity monitor says handbrake using 776% - 790% CPU. The internal tempature hasn't gone above 94c and the exhaust rpm is around 3400.

So really, not that bad. I can put my hand on the mini and barely feel any heat at all. If I find it starts running way too hot, I'll just limit the CPU cores used via handbrake advance setting. My only real issue is with Handbrake, it's gobbling up every last bit of RAM it can get!

It'd be interesting to see actual temperature differences using the pad and not in a lets say 3h 1080p encode with Handbrake.

Now they're idle and at 38-40ºC. (100-104ºF). Pretty cool.

[dasx]

Originally Posted by Darby67

I wouldn't say silly at all, if these are real world for you, then they are much better than GB or other metrics. Enjoy!

Actually, what I do most is "Test 5" related stuff. So that CPU helps a lot there.

[blanka]

Hot does not matter. Intel monitors its CPU very well, and when "too" comes in ranch, it will adjust speed.

Please run Cinebench 11.5 for GPU performance. Nobody here did that yet

[dasx]

Originally Posted by blanka

Hot does not matter. Intel monitors its CPU very well, and when "too" comes in ranch, it will adjust speed.

Please run Cinebench 11.5 for GPU performance. Nobody here did that yet

OK. Downloading.

[The-Pro]

take off the plastic cover at the bottom of the mini and place the mini on its side, you will see a drastic reduction in temperature.

[dasx]

Originally Posted by The-Pro

take off the plastic cover at the bottom of the mini and place the mini on its side, you will see a drastic reduction in temperature.

Yeah I guess. I just don't wanna have to live with a vertical and bottom-naked mini. :/

I'll try raising it a little bit with some little rubber pads.

----------

Originally Posted by blanka

Hot does not matter. Intel monitors its CPU very well, and when "too" comes in ranch, it will adjust speed.

Please run Cinebench 11.5 for GPU performance. Nobody here did that yet

Here.

[The-Pro]

Originally Posted by dasx

Yeah I guess. I just don't wanna have to live with a vertical and bottom-naked mini. :/

I'll try raising it a little bit with some little rubber pads.[COLOR="#808080"]

Understandable. I raised my 2010 with two 7cm styrofoam blocks. One on either side right and left. When encoding stuff CPU temps went from 87°C with 4000rpm fans down to 65°C and 2200rpm fans. Pretty good.

[philipma1957]

lost cost and nice look this will lower temps!


http://www.amazon.com/Sm%C3%A5land-B...ds=ikea+trivet

[dasx]

Originally Posted by The-Pro

Understandable. I raised my 2010 with two 7cm styrofoam blocks. One on either side right and left. When encoding stuff CPU temps went from 87°C with 4000rpm fans down to 65°C and 2200rpm fans. Pretty good.

You serious? From 87ºC to 65ºC?? WOW. That's amazing. Will try to lift it tomorrow with anything and if I get that result I'll find a way to make it look nice.

EDIT: Did a quick test. Lifted it up around 2cm and didn't really see any change. :/

[The-Pro]

Originally Posted by dasx

You serious? From 87ºC to 65ºC?? WOW. That's amazing. Will try to lift it tomorrow with anything and if I get that result I'll find a way to make it look nice.

EDIT: Did a quick test. Lifted it up around 2cm and didn't really see any change. :/

Yeah man, im not kidding. Ill post some screenshots, ill do it again

Weird. Well it takes a while for it to go down. I don't know how it is with the newer minis. Mine is a 2.4 C2D. They produce less heat then the i7's etc.

Here you go:
1st screenshot. bottom panel on, 25 mins after I started maxing out the CPU's with a handbrake encode.
2nd screenshot. took off the bottom panel, stood mini on its side, kept handbrake going. Screenshot taken 15 mins after bottom panel was taken off.

[dasx]

Originally Posted by The-Pro

Yeah man, im not kidding. Ill post some screenshots, ill do it again

Weird. Well it takes a while for it to go down. I don't know how it is with the newer minis. Mine is a 2.4 C2D. They produce less heat then the i7's etc.

Here you go:
1st screenshot. bottom panel on, 25 mins after I started maxing out the CPU's with a handbrake encode.
2nd screenshot. took off the bottom panel, stood mini on its side, kept handbrake going. Screenshot taken 15 mins after bottom panel was taken off.

Ahhh. I thought you meant you got that temp drop just by raising the Mini 7cm. I can understand 20ºC difference by taking off the bottom and putting it on its side.

I'll try to do some serious tests this week. Raising it around 3-5cm and see what happens.

[thedenethor]

Originally Posted by The-Pro

take off the plastic cover at the bottom of the mini and place the mini on its side, you will see a drastic reduction in temperature.

Since lots of people recommends this I decided to give shot. And yes, my mid 2011 became noticeably silent. Now I think I can wait till Haswell for my upgrade
Noise was the only issue for me with my current Macmini and I did not want to loose dedicated graphics because of Madvr.

Now the new issue is dust since I removed the buttom cover. I guess i will build my own DIY air filter

[dasx]

Originally Posted by thedenethor

Since lots of people recommends this I decided to give shot. And yes, my mid 2011 became noticeably silent. Now I think I can wait till Haswell for my upgrade
Noise was the only issue for me with my current Macmini and I did not want to loose dedicated graphics because of Madvr.

Now the new issue is dust since I removed the buttom cover. I guess i will build my own DIY air filter

An idea that crossed my mind was getting the bottom cover (here for instance), drill a hole right were the fan is and put some kind of filtering paper or something.
The just raise the Mini and voila, a more cooled Mini without having to lose appearance and having to leave it bottom-naked.

[thedenethor]

Originally Posted by dasx

An idea that crossed my mind was getting the bottom cover (here for instance), drill a hole right were the fan is and put some kind of filtering paper or something.
The just raise the Mini and voila, a more cooled Mini without having to lose appearance and having to leave it bottom-naked.

Thanks for the suggestion. I think I will gowith this one .
For the riser I'm already started to usemy rMBP rubbers

[dasx]

Do you know btw (or have any way to check it) if the 2010 Mac Mini bottom cover is compatible with the 2012 model? I'm saying this because didn't find any specific 2012 Mini bottom cover. Just 2010.

Originally Posted by thedenethor

For the riser I'm already started to usemy rMBP rubbers

And which are the results? Any noticeable differences?

[thedenethor]

Originally Posted by dasx

Do you know btw (or have any way to check it) if the 2010 Mac Mini bottom cover is compatible with the 2012 model? I'm saying this because didn't find any specific 2012 Mini bottom cover. Just 2010.

And which are the results? Any noticeable differences?

Yes they are compatible just tried that to confirm. (I'm still keeping my 2010)

Edit: Upps sorry I can only say that 2010 and 2011 covers are compatible. I do not have 2012 to test.

For the rubbers yes they did the job. It s not %100 silent but good enough. You can't hear the fans from 3-4mt while Plex or Dvblink doing real time transcoding. Fans are at 3000 rpm. If you need 2000-2500 rpm you may need higher hills. Will do the test during night since the environment more silent. After that test I'll make my conclusion.

[LostSoul80]

That's interesting. Performance is great for an entry level product, but cooling is a decisive factor in any computer's lifespan.