Assembling the worlds fastest Mac Mini DualCore i7 2011/2012

radiogoober · Apr 27, 2012

What a colossal waste of time and money. All of these ridiculous upgrades made to a low-priced computer, only to have it perform marginally better than less than half the price stock configuration.

Congratulations, on a non-real world benchmark you can perform 0.02 seconds better than me. lol.

scottsjack · Apr 27, 2012

radiogoober said:
What a colossal waste of time and money. All of these ridiculous upgrades made to a low-priced computer, only to have it perform marginally better than less than half the price stock configuration.

Congratulations, on a non-real world benchmark you can perform 0.02 seconds better than me. lol.

I think you missed the point. It's like building American muscle cars and hot rods. As transportation vehicles they're pretty much worthless but it is fun to do. My own dream mini would be an Ivy Bridge quad-core i7 with two big, two fast SSDs and 16GB of RAM.

Maybe only a fool would build a $3500 mini but it would make a great Mac DVR!

MacSawdust · Apr 27, 2012

Real World Speed

I put a SSD in my 2011 Mini and its like a new computer.
I will never buy another computer without a SSD.
I ran the free geekbench in 32 bit and got a 8738.
http://browse.geekbench.ca/geekbench2/633051

Not sure what the score means, but the SSD makes it a pleasure to use.

Westyfield2 · Apr 27, 2012

radiogoober said:
What a colossal waste of time and money. All of these ridiculous upgrades made to a low-priced computer, only to have it perform marginally better than less than half the price stock configuration.

Congratulations, on a non-real world benchmark you can perform 0.02 seconds better than me. lol.

As with everything there is a point where the law of diminishing returns kicks in (and I personally would say that RAIDing SSDs is that point), but there's no harm in what AT0MAC is doing. Heck at £779 I wouldn't exactly call the 2.7GHz i7 Mac Mini that AT0MAC started off with "low-priced".

Why do I overclock my PC? Because I can. Why do I spend money tuning my car? Because I can. Why should we explore space? Because we can. Why did George Mallory climb Mount Everest? "Because it's there".

AT0MAC · Apr 28, 2012

MacSawdust said:
I put a SSD in my 2011 Mini and its like a new computer.
I will never buy another computer without a SSD.
I ran the free geekbench in 32 bit and got a 8738.
http://browse.geekbench.ca/geekbench2/633051

Nice numbers, but yours is the QuadCore Server edition and thats why you have such a high number. I agree on SSD, it makes a huge difference.

What the numbers mean is a benchmark, a comparison, of how well your CPU, RAM and HDD/SSD works together and in separate fields. Don't think much about it, just know that the higher the number the better performing a machine you should have (I say should because it's always better to test out real work tasks and compare on them) .

----------

Westyfield2 said:
As with everything there is a point where the law of diminishing returns kicks in (and I personally would say that RAIDing SSDs is that point), but there's no harm in what AT0MAC is doing. Heck at £779 I wouldn't exactly call the 2.7GHz i7 Mac Mini that AT0MAC started off with "low-priced".

Why do I overclock my PC? Because I can. Why do I spend money tuning my car? Because I can. Why should we explore space? Because we can. Why did George Mallory climb Mount Everest? "Because it's there".

I agree with the part of "because we/I can"

AT0MAC · Apr 28, 2012

Westyfield2 said:
Just been reading through all this, good work AT0MAC .

Wonder if there's anything to be gained in the lower voltage of the Crucial kit over the Corsairs? Sandy Bridge has the memory controller in the CPU, so I guess the CPU having to supply the memory with lower voltage would lower CPU temps?

Thank you

What you would gain is a slightly lower power usage, you use a tiny bit less watt in having your Mac turned on.

As I can find from various sources the memory controller is yes integrated in the CPU, but have nothing to do with the actual CPU. So you processor will still run the same GHz and have the same TDP.

Apples own site don't say anything about voltage so my guess is that they are perfectly good all of them. There is even a thread here discussing that RAM.

theSeb · Apr 28, 2012

AT0MAC said:
Nice numbers, but yours is the QuadCore Server edition and thats why you have such a high number. I agree on SSD, it makes a huge difference.

What the numbers mean is a benchmark, a comparison, of how well your CPU, RAM and HDD/SSD works together and in separate fields. Don't think much about it, just know that the higher the number the better performing a machine you should have (I say should because it's always better to test out real work tasks and compare on them) .

----------

I agree with the part of "because we/I can"

Geekbench only tests CPU and RAM. The hard drive speed makes no difference. You can test HDD speed with an app like xbench.

AT0MAC · Apr 28, 2012

theSeb said:
Geekbench only tests CPU and RAM. The hard drive speed makes no difference. You can test HDD speed with an app like xbench.

Are you sure about that?
The last section, Stream Performance, have given me totally different scores depending on how I setup my HDD/SSD

theSeb · Apr 28, 2012

AT0MAC said:
Are you sure about that?
The last section, Stream Performance, have given me totally different scores depending on how I setup my HDD/SSD

I am fairly sure since I've used it many times.

The 4 bottom bars are Geekbench 64 bit scores for the same mini server. The scores move only when I changed the RAM from 1333 MHz to 1600 MHz. Geekbench scores normally differ by ## points every time you run and those differences are not significant. Differences by ### are significant to consider.

The bottom bar is using the WD Scorpio Black 750 GB 7200 RPM. The third bar from the bottom is the same config in terms of memory, but using an OCZ Vertex 3 SSD.

I also ran the same test on two 7200 RPM in RAID 0 and the geekbench scores remain the same.

From the Geekbench website:

"Geekbench provides a comprehensive set of benchmarks engineered to quickly and accurately measure processor and memory performance."

AT0MAC · Apr 28, 2012

theSeb said:
I am fairly sure since I've used it many times.

The scores move only when I changed the RAM from 1333 MHz to 1600 MHz. Geekbench scores normally differ by ## points every time you run and those differences are not significant. Differences by ### are significant to consider.
...
From the Geekbench website:

"Geekbench provides a comprehensive set of benchmarks engineered to quickly and accurately measure processor and memory performance."

It's a little funny then, look at these two - it's my Mac Mini 2011 (then one in this thread) with one 240GB Intel 520 SSD and then again with two SSDs with block size 256k, same RAM but the stream performance is hugely different.

Westyfield2 · Apr 28, 2012

AT0MAC said:
Thank you

What you would gain is a slightly lower power usage, you use a tiny bit less watt in having your Mac turned on.

As I can find from various sources the memory controller is yes integrated in the CPU, but have nothing to do with the actual CPU. So you processor will still run the same GHz and have the same TDP.

Apples own site don't say anything about voltage so my guess is that they are perfectly good all of them. There is even a thread here discussing that RAM.

Just had a look through your Flickr gallery, the Hynix 2GBx2 DDR3-1333 (9-9-9) yours came with is 1.5V anyways (part number HMT325S6BFR8C-H9 N0 AA).

Were you able to work out what was up with Rember? Was it failing all RAM you tried, or just some?

I've never done anything like that on my Macs, but on my PCs for testing dodgy RAM I use a Memtest86+ bootdisc, and then for stress-testing overclocks within Windows I use IntelBurnTest (Intel Linpack) as I've found it detects instability a lot quicker than Prime95.

theSeb · Apr 28, 2012

AT0MAC said:
It's a little funny then, look at these two - it's my Mac Mini 2011 (then one in this thread) with one 240GB Intel 520 SSD and then again with two SSDs with block size 256k, same RAM but the stream performance is hugely different.

Image

The stream tests are testing your CPU performance and not your hard drive speed. Are you sure all of the variables were the same when you ran the tests? i.e. Was the computer busy doing something else? Indexing or anything like that?

AT0MAC · Apr 28, 2012

theSeb said:
The stream tests are testing your CPU performance and not your hard drive speed. Are you sure all of the variables were the same when you ran the tests? i.e. Was the computer busy doing something else? Indexing or anything like that?

I don't know what else it might have been doing, It's the first test I run after a totally fresh install of OS X. Same procedure every time. After the install first I do is to restart and run GeekBench.

----------

Westyfield2 said:
Were you able to work out what was up with Rember? Was it failing all RAM you tried, or just some?

I don't know why Rember gave me a error, it was only the Corsair Vengeance it did so with. I didn't see much performance gains in having that particular set of RAM so ended up sending them back (but not as an RMA, just money back guaranty).

theSeb · Apr 28, 2012

AT0MAC said:
I don't know what else it might have been doing, It's the first test I run after a totally fresh install of OS X. Same procedure every time. After the install first I do is to restart and run GeekBench

Immediately after an install OSX will be busy indexing your drive with spotlight so I wouldn't recommend running any benchmarks at that time since it could skew the results.

AT0MAC · Apr 28, 2012

theSeb said:
Immediately after an install OSX will be busy indexing your drive with spotlight so I wouldn't recommend running any benchmarks at that time since it could skew the results.

I did the same every time so I am surprised if you say it only tests memory and processor that their could be such a big difference. Anyway... It's too late to change the tests now, It took forever last time to do them all

gmartine · Jun 22, 2012

Heat dissipation

Hi,
Have you observed any cooling issues by any chance? I am building my own HTPC. I have a Mac Mini 2011 i5 running OSX 10.7.4

I am really tempted to get the Kingston PNP 1866Mhz.

I can get it for just $79.99 free shipping

http://www.newegg.com/Product/Product.aspx?Item=N82E16820104259

I am only hesitating because the impact of the overclocking with additional heat dissipation. Could this create any damage?

Thanks

AT0MAC · Jun 22, 2012

gmartine said:
Hi,
Have you observed any cooling issues by any chance?

Not really, im not sure if you only upgrade the RAM though, as spinning harddrives get a lot hotter than the SSDs Im using.

Ice Dragon · Jun 24, 2012

gmartine said:
Hi,
Have you observed any cooling issues by any chance? I am building my own HTPC. I have a Mac Mini 2011 i5 running OSX 10.7.4

I am really tempted to get the Kingston PNP 1866Mhz.

I can get it for just $79.99 free shipping

http://www.newegg.com/Product/Product.aspx?Item=N82E16820104259

I am only hesitating because the impact of the overclocking with additional heat dissipation. Could this create any damage?

Thanks

I would buy some but only a limit of 100 per customer. Dealbreaker for me!

: P

gabrielscindian · Nov 2, 2012

My 2012 Mac Mini has a geekbench score of 11655. I edit and playback my RED 4k files in Smoke 2013 beta release 5 with ease. check out my images on twitter -gabrielscindian

AT0MAC said:
Inspired by a guy known to the Internets as MKBHD I have started to build the worlds apparently fastest Mac Mini computer.

This is my 3rd Mac Mini I'm upgrading to the max so am no novice in that department.

If interested to see what "the opponent" have done, check out his videos:
Worlds Fastest Mac Mini Part1

What he have done to now claim to have the fastest Mac Mini is to get the build-to-order Intel Core i7 2.7GHz Mac Mini 2011, upgrade it with 2x8 GB OWC DDR1333 RAM and throwing in a OCZ Vertex 3 MAX IOPS Edition SSD drive.

Is that enough? Not if you ask me....

SPOILER ALERT:
There will not be any modding or actual building, but more like a step by step plan to chose the right components and how it goes after assembling the thing. Sprinkled with a bunch of benchmarks.

----------

[/COLOR]I will benchmark it and compare to both my former Mac but also compare it to the bench MKBHD have put on his Google+ and YouTube.

The machine I'm upgrading from is a Mac Mini 2010 server unibody (model ID 4.1) with specs as follow:

2.66 GHz Intel Core2Duo
nVidia GeForce 320M 256MB graphics
2x 4 GB Corsair DDR1066 RAM
2x Seagata Momentus XT 7.200RPM 500GB HDD / 4GB SSD in RAID0
OS X 10.6 Snow Leopard Server

The opponent have:

Mac Mini 5.2 2011
2.7 GHz Intel DualCore i7
AMD Radeon HD6630m 256MB graphics
2x 8 GB OWC DDR1333 RAM
OCZ Vertex 3 MAX IOPS Edition SSD (I think it's 240GB)
OS X 10.7 Lion

I have just received the new Mac Mini, in standart it is:

Mac Mini 5.2 2011
2.7 GHz Intel DualCore i7
AMD Radeon HD6630m 256MB graphics
2x 2 GB DDR1333 RAM
5.400 RPM 500GB HDD
OS X 10.7 Lion

But the main event is the build:

Mac Mini 5.2 2011
2.7 GHz Intel DualCore i7
AMD Radeon HD6630m 256MB graphics
2x 8 GB Corsair Vengeance DDR1600 auto-overclocking RAM
2x Intel Series 520 240 GB SSD in RAID0
OS X 10.7 Lion

----------

2011 Mac Mini 8GB or 16GB RAM?

SSDs in RAID0 - what disc to choose?

Screw Types /Screwdirvers needed for a MacMini 2011

Thermal paste upgrade any difference?

Selling my Mac Mini

Also run this whole thread @Bit-Tech link
And on 123macmini

----------

Info:
Image

Geekbench:
Image

DiskSpeed:
Image

NovaBench:
Image

XBench:
Image

CineBench OpenGL:
Image

CineBench CPU:
Image

----------

The GeekBench test scores on the old Mac is not that impressive, mostly because I tested it in Snow Leopard Server and only testet the 32bit edition of GeekBench.

I did however for a while run the machine with Lion and got some pretty amazing results:

4612 points 64bit
and
4316 points 32bit

To compare the new build to the machine I now sold is not the biggest deal, more to see how much the upgrading in this machine change.

----------

This is the stock version build-to-order DualCore i7 2.7GHz with the slow 5.400RPM drive and silly 4GB RAM.

Info:
Image

Geekbench 32bit:
Image

Geekbench 64bit:
Image

DiskSpeed:
Image

NovaBench:
Image

XBench:

Something strange happened, I waited forever but could not get it to make the Thread test, so here it is without and were it got stuck.
Later on I learned that a SandyBridge CPU can't run the Thread test, so that mystery is now solved

Image

Image

CineBench will probably not change so much before/after upgrading, but am testing it anyway.

CineBench OpenGL:
Image

CineBench CPU:
Image

----------

Saturday when I saw that my preferred SSDs had gone a few $ up in price I decided to order them.

My guess is that because of the newly released iPad there will be a bit higher demand for good NAND the next few weeks/month and therefor prices atm will not fall as much as they used to.

2 pcs. Intel 520 Series 240GB top-of-the-line SSDs on the way to me

Image

8)

Link to test

----------

Over at 123macmini another user pointed me to this site:

Here is a guy there with the most evil mini I have yet seen, shortcut3d, he have a 2011 Mac mini server
2.0GHz Quad-Core Intel i7
8GB 1866MHz Kingston Hyper X
Dual (2) 240GB OCZ Vertex 3 MAX IOPS!!!

That gives him a staggering 9730 points in GeekBench!

I think he have the score to beat...
The Server mini have slower graphics but a Quad-Core i7 instead (wich in raw MHz terms is also slower), he have faster RAM, but only half the size as me...

Hmm wondering if I can beat that...

gabrielscindian · Nov 2, 2012

gabrielscindian said:
My 2012 Mac Mini has a geekbench score of 11655. I edit and playback my RED 4k files in Smoke 2013 beta release 5 with ease. check out my images on twitter -gabrielscindian

Here are some images.

Assembling the worlds fastest Mac Mini DualCore i7 2011/2012

macrumors 6502a

macrumors 68000

macrumors member

macrumors 6502a

macrumors regular

macrumors regular

macrumors 604

macrumors regular

macrumors 604

macrumors regular

macrumors 6502a

macrumors 604

macrumors regular

macrumors 604

macrumors regular

macrumors newbie

macrumors regular

macrumors 6502a

macrumors newbie

macrumors newbie

Attachments

Our Staff