Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Assembling the worlds fastest Mac Mini DualCore i7 2011/2012
- Thread starter AT0MAC
- Start date
- Sort by reaction score
So far I have 15 different Benchmarks to run through at least in 5 different configurations of the same machine (plus I for 7 of them also ran them on my old mm). That gives a BUNCH of benchmarks to see the progress and measure if it was all worth it...
Those I remember to take screen dumps of will end up in my Flickr
My guess is that close to Easter I will be completely done and the results will be postet here.
Those I remember to take screen dumps of will end up in my Flickr
My guess is that close to Easter I will be completely done and the results will be postet here.
the quad core is far faster then the dual core for hand brake. we did tests in the summer.
the quad core will run 8 cores (hyperthreading) at 90 percent when doing handbrake. It was almost as fast as my mac pro. a 12 min 1080p big buck bunny was done in 6 mins and change on the quad mini. and it was done in 5 minutes and some change on my custom 2010 mac pro hex 3.2 was cool seeing 12 cores at 70 percent maybe 74 percent. now at the time the mini cost 1k the modded mac pro was 3k. the i7 dual core should do big buck bunny in 8 minutes and some change.
I was running Rember memory test over night and when I came to check on it this morning I found an error that looked like this:
Ran another memory test in Tech Tool Pro-To-Go and found no error there after 15 minutes run.
I'm thinking it must have been a heating problem? The mm fan was spinning like crazy when it ran Rember...
Ran another memory test in Tech Tool Pro-To-Go and found no error there after 15 minutes run.
I'm thinking it must have been a heating problem? The mm fan was spinning like crazy when it ran Rember...
Today I have been running 10 more laps of Rember and again in test 10 out of 10 I get the same strange error as posted before...
I will bench over the weekend and go pick up my Kingston PnP 8GB kit on monday, if the 8GB kit benches very close to the 16 (or better because its faster) then I will try to return the Corsairs for refund.
Just updated my Flickr with pictures from the past two days tests.
Not actual benchmark tests, but more from when I fittet the RAM and swapped the HDD in my external G-Drive to one of the SSDs. Making it ready to be cloned off the Mac Mini HDD.
![6863408266_1638794699.jpg]()
Check out the gallery for the whole story
Not actual benchmark tests, but more from when I fittet the RAM and swapped the HDD in my external G-Drive to one of the SSDs. Making it ready to be cloned off the Mac Mini HDD.
Check out the gallery for the whole story
What geekbench score is the mark to beat on this model in os x lion?
Thought it was a question of more broad interest so gave it it's own thread
Been benching a lot today and found some interesting details I will share later on.
Thought it was a question of more broad interest so gave it it's own thread
Been benching a lot today and found some interesting details I will share later on.
I posted the result difference on my 2011 15" MacBook Pro 2.3GHz Quad-Core Intel i7. This was a very good apples to apples comparison because the previous memory was the same capacity and vendor, so the only change was the memory speed.
Post #332
https://forums.macrumors.com/threads/1133642/
4GB DDR1333 vs 16GB DDR1600
NovaBench:
![novabenchcompare4vs16pn.png]()
Faster RAM speed + more RAM = higher score.
Will have to see what the Kingston PnP 1866 gives because it's only half the size but even faster speed.
GeekBench:
![comparegeekbenchpng.png]()
Memory performance a little bit higher, mostly due to RAM speed I guess, but will see that when I have the Kingston PnP number because they are even faster.
Memory bandwidth gives a BIG difference, my guess is due to the much bigger memory size, 4GB vs 16GB. 8GB numbers should land somewhere in between.
Lloyd Photoshop benchmark:
![lloydcomparepng.png]()
Photoshop CS6 is a really memory hungry app, swallows the 16GB like candy but also gives a HUGE speed difference.
Will wait to see if it's only speed as in RAM speed or if amount also makes a difference.
Maybe I will give Corsair ValueRAM 2x8GB DDR1333 a shot also to conclude wich factor is most important.
Rumors about faster RAM gives a FPS improvement is something I have not seen. Been testing with Duke Nukem Forever and a screen recording app that shows average fps recorded, numbers were pretty close.
Also file transfer speed is not any second better, will have to wait and see what happens when I install my SSDs.
More pictures from these tests can be found @flickr
----------
A funny fact by the way, my boot times from I push the button til it auto logins and I see the desktop is with the 16GB RAM 1.5 seconds slower than the stock 4GB...
Have testet it several times and it's consistent.
On the other hand, boot to login screen is pretty much the same, average 0.2 seconds quicker.
NovaBench:
Faster RAM speed + more RAM = higher score.
Will have to see what the Kingston PnP 1866 gives because it's only half the size but even faster speed.
GeekBench:
Memory performance a little bit higher, mostly due to RAM speed I guess, but will see that when I have the Kingston PnP number because they are even faster.
Memory bandwidth gives a BIG difference, my guess is due to the much bigger memory size, 4GB vs 16GB. 8GB numbers should land somewhere in between.
Lloyd Photoshop benchmark:
Photoshop CS6 is a really memory hungry app, swallows the 16GB like candy but also gives a HUGE speed difference.
Will wait to see if it's only speed as in RAM speed or if amount also makes a difference.
Maybe I will give Corsair ValueRAM 2x8GB DDR1333 a shot also to conclude wich factor is most important.
Rumors about faster RAM gives a FPS improvement is something I have not seen. Been testing with Duke Nukem Forever and a screen recording app that shows average fps recorded, numbers were pretty close.
Also file transfer speed is not any second better, will have to wait and see what happens when I install my SSDs.
More pictures from these tests can be found @flickr
----------
A funny fact by the way, my boot times from I push the button til it auto logins and I see the desktop is with the 16GB RAM 1.5 seconds slower than the stock 4GB...
Have testet it several times and it's consistent.
On the other hand, boot to login screen is pretty much the same, average 0.2 seconds quicker.
The more hardware you add the longer the post test. Therefore, more RAM slightly increases boot time.
Just installed my new Kingston HyperX PnP 1867MHz DDR3 2x4GB kit.
Am going to Rember test them the next few hours to see if they work as expected before benching anything.
Am thinking about if I have damaged my SSDs and original Apple RAM because I have placed them next to my small desktop speakers with 4" midrange?
Can the magnetic field from the speaker do anything bad to RAM or SSDs?
Just a thought, have not tested them or anything but moved them further away just in case it was a bad place to put them...
Am going to Rember test them the next few hours to see if they work as expected before benching anything.
Am thinking about if I have damaged my SSDs and original Apple RAM because I have placed them next to my small desktop speakers with 4" midrange?
Can the magnetic field from the speaker do anything bad to RAM or SSDs?
Just a thought, have not tested them or anything but moved them further away just in case it was a bad place to put them...
After running Rember most of yesterday, failing the first few tests with same strange error as the Corsairs I finally figured out how to run the real memtest in Single User mode!
Feels like going back to the old days when I started as a computer enthusiast (ca. 1991):
![20120327307.jpg]()
The new Kingstons passed with flying colors!
A small minus regarding the Kingston compared to the Corsair is the look, they just don't look as good
not that you will ever open up and look, but just to know that they look awesome is a nice thing)
Kingston:
![20120326295.jpg]()
Corsair:
![20120322262.jpg]()
------------------
After benching GeekBench I had quite a surprise!
CPU test are very similar between the two RAM kits:
![gbcputest.png]()
But memory benchs on the 8GB Kingston kit are a whole lot better:
![gbmemtest.png]()
Gave me a 7.918 total score in 64bit!!
Sorry the pics are so small, they are also on my Flickr
Feels like going back to the old days when I started as a computer enthusiast (ca. 1991):
The new Kingstons passed with flying colors!
A small minus regarding the Kingston compared to the Corsair is the look, they just don't look as good
not that you will ever open up and look, but just to know that they look awesome is a nice thing)Kingston:
Corsair:
------------------
After benching GeekBench I had quite a surprise!
CPU test are very similar between the two RAM kits:
But memory benchs on the 8GB Kingston kit are a whole lot better:
Gave me a 7.918 total score in 64bit!!
Sorry the pics are so small, they are also on my Flickr
SSDs are in the belly
The SSDs are installed and the first test have been running on one SSD - will post more later but from now on you can at least sneak peak at my flickr
One sentence about the results so far: AWESOME to cold boot in just 18 seconds!!!!
Atm I am downloading the newest image file from App store of OS X Lion to install in RAID0 and bench some more, stay tuned...
The SSDs are installed and the first test have been running on one SSD - will post more later but from now on you can at least sneak peak at my flickr
One sentence about the results so far: AWESOME to cold boot in just 18 seconds!!!!
Atm I am downloading the newest image file from App store of OS X Lion to install in RAID0 and bench some more, stay tuned...
for a magnet to hurt a hdd or ssd it takes long exposures days and intensive strength fields. Somewhere I saw a test that even with an electric magnet of hi power the hdd did not die. With an hour or two . After 11 days the hdd had some bad sectors. I think the magnet had the power to lift 200 pounds or 100 kilos. So you are fine .
The kingston are the best for the mini . I am happy with my 1600. Which were 60 and the 1866 were 120. Now that the 1866 dropped in price you got a better deal then I did.
Super-duper about the magnet stuff, thank you
Here in Denmark the Corsairs are $186 / £117 and the Kingstons $106 / £66.
I really like the Kingston. As my results will show when I post my comparison charts the Kingstons are the best value and in many aspects also the best/fastest.
Today I send the Corsair kit back to get a refund.
Just made a OS X Lion USB install disc and about to install in RAID0
Final results have just been uploaded to Flickr.
I don't have time now to sort them out and explain them, but you can sneak peak until I return.
I don't have time now to sort them out and explain them, but you can sneak peak until I return.
The final score
The final scores have been sorted out and counted and have now ended up in easy to read graphics.
Lets start...
GeekBench 32bit:
![20120330093320am.png]()
As expected the GB score kept climbing until it reached it's max of 7.250 points wich is a lot in 32bit as far as I can tell!
GeekBench 64bit:
![20120330093329am.png]()
Same story goes for the GB 64bit scores wich tops out at 7.939 points!
An interesting fact is that the 8GB Kingston 1867MHz kit scores higher than the 16GB Corsair 1600MHz kit. Seems that GB really likes fast RAM and don't care much about the amount.
Another interesting detail is how close the SSD scores are, GB don't utilizes the read/write speeds of a RAID0 very well apparently...
For short, it's a true synthetic benchmark that only gives a faint hint of how powerful a machine really is.
The final scores have been sorted out and counted and have now ended up in easy to read graphics.
Lets start...
GeekBench 32bit:
As expected the GB score kept climbing until it reached it's max of 7.250 points wich is a lot in 32bit as far as I can tell!
GeekBench 64bit:
Same story goes for the GB 64bit scores wich tops out at 7.939 points!
An interesting fact is that the 8GB Kingston 1867MHz kit scores higher than the 16GB Corsair 1600MHz kit. Seems that GB really likes fast RAM and don't care much about the amount.
Another interesting detail is how close the SSD scores are, GB don't utilizes the read/write speeds of a RAID0 very well apparently...
For short, it's a true synthetic benchmark that only gives a faint hint of how powerful a machine really is.
DiskSpeed read/writes:
![20120330093337am.png]()
Aboslute AWESOME read/write speeds this little machine have now when it is in RAID0 = 520MB/s write!
Thats half a GB! It's like the usual amount on a a Music CD written in 1 second (if just there was a CD burner that could write that same speed...)
870MB/s read speeds!
Thats about the same read speeds as the OCZ Z-Drives have in four way RAID (whatever that is?)!
A strange thing here is that the fast Kingston 1867MHz 8GB kit scores the slowest read/write on the stock Apple 5400RPM drive.
The 16GB 1600MHz gave it a good overall speed improvement of a few MB/s more.
NovaBench:
![20120330093345am.png]()
NB looks a lot like GeekBench like a real synthetic benchmark, wich it is.
It favors a lot of RAM and a speedy disk compared to a flying speedy RAID0 system and a lot faster set of RAM.
If you find the schematics/results I screen dumped to my Flickr you see that the 16GB and 8GB kits have one mayor difference in their scores - the RAM test.
CPU test are almost similar with 459 and 467 in favor of the 8GB set, Graphics score similar 116 and 115, hardware test in favor of the speedier 8GB set with 19 compared to 21. But RAM score is 221 and 180 - only because of the amount! The speed it self the 16GB scores 8756MB/s were the 8GB scores 8886MB/s and still gets a smaller total score because the amount is smaller.
Not fair!
MKBHD scores in his YouTube video of "Fastest Mac Mini in the World" a score of 834:
RAM 219
CPU 442
Graphic 115
Hardware 58
Mine would be with the RAID0 and the 16GB 856 points:
RAM 221 (taken from 16GB RAM scores)
CPU 459 (taken from 16GB RAM scores)
Graphic 116 (taken from SSD RAID scores)
Hardware 60 (taken from SSD RAID scores)
Aboslute AWESOME read/write speeds this little machine have now when it is in RAID0 = 520MB/s write!
Thats half a GB! It's like the usual amount on a a Music CD written in 1 second (if just there was a CD burner that could write that same speed...)
870MB/s read speeds!
Thats about the same read speeds as the OCZ Z-Drives have in four way RAID (whatever that is?)!
A strange thing here is that the fast Kingston 1867MHz 8GB kit scores the slowest read/write on the stock Apple 5400RPM drive.
The 16GB 1600MHz gave it a good overall speed improvement of a few MB/s more.
NovaBench:
NB looks a lot like GeekBench like a real synthetic benchmark, wich it is.
It favors a lot of RAM and a speedy disk compared to a flying speedy RAID0 system and a lot faster set of RAM.
If you find the schematics/results I screen dumped to my Flickr you see that the 16GB and 8GB kits have one mayor difference in their scores - the RAM test.
CPU test are almost similar with 459 and 467 in favor of the 8GB set, Graphics score similar 116 and 115, hardware test in favor of the speedier 8GB set with 19 compared to 21. But RAM score is 221 and 180 - only because of the amount! The speed it self the 16GB scores 8756MB/s were the 8GB scores 8886MB/s and still gets a smaller total score because the amount is smaller.
Not fair!
MKBHD scores in his YouTube video of "Fastest Mac Mini in the World" a score of 834:
RAM 219
CPU 442
Graphic 115
Hardware 58
Mine would be with the RAID0 and the 16GB 856 points:
RAM 221 (taken from 16GB RAM scores)
CPU 459 (taken from 16GB RAM scores)
Graphic 116 (taken from SSD RAID scores)
Hardware 60 (taken from SSD RAID scores)
XBench:
![20120330093354am.png]()
XBench is also a synthetic benchmark but it looks like it's more accurate than both GeekBench and NovaBench as it favors speed above anything else.
The faster Kingston RAM gets higher score than the Corsair and also the total winner is the RAID0 system. Nice
CineBench OpenGL:
![20120330093402am.png]()
It is practically identical no matter what configuration I used, except for the old Mac Mini Server C2D machine I started with.
I had heard a rumor faster RAM should give higher FPS, don't work that way on the MM with AMD graphics - maybe on the Intel HD graphics edition were it uses a portion of the system RAM?
Decided to not test on the SSDs as it didn't seem to change anything.
CineBench CPU test:
![20120330093409am.png]()
Same story goes for CineBench CPU test, same CPU all trough so no difference what so ever.
Didn't test it on the SSD setups.
A funny thing though, the stock MM 2011 gave higher scores, maybe because I later on had SMC upgraded and a few OS X updates from Apple?
The rest of the tests are made on exactly the same system with only the changes as described.
XBench is also a synthetic benchmark but it looks like it's more accurate than both GeekBench and NovaBench as it favors speed above anything else.
The faster Kingston RAM gets higher score than the Corsair and also the total winner is the RAID0 system. Nice
CineBench OpenGL:
It is practically identical no matter what configuration I used, except for the old Mac Mini Server C2D machine I started with.
I had heard a rumor faster RAM should give higher FPS, don't work that way on the MM with AMD graphics - maybe on the Intel HD graphics edition were it uses a portion of the system RAM?
Decided to not test on the SSDs as it didn't seem to change anything.
CineBench CPU test:
Same story goes for CineBench CPU test, same CPU all trough so no difference what so ever.
Didn't test it on the SSD setups.
A funny thing though, the stock MM 2011 gave higher scores, maybe because I later on had SMC upgraded and a few OS X updates from Apple?
The rest of the tests are made on exactly the same system with only the changes as described.
Handbrake 0.9.5 video encoding test:
![20120330093418am.png]()
One of you here at Mac Rumors Forum suggested to try a Handbrake video encoding test, so I did...
It faired pretty good - quicker than a Mac Pro with dual Xeon quad-core processor (that one scored 24.3 FPS in 9:51 minutes)!
Handbrake 0.9.6 video encoding test:
![20120330093424am.png]()
But there have been released a new edition of Handbrake since that guys thread started, so decided to give it a go in the new version also.
The old version favored more RAM (Corsair 16GB kit) were the new version favors faster RAM (Kingston 8GB kit 1867MHz).
None of them however seems to care about read/write speeds on your harddrive.
One of you here at Mac Rumors Forum suggested to try a Handbrake video encoding test, so I did...
It faired pretty good - quicker than a Mac Pro with dual Xeon quad-core processor (that one scored 24.3 FPS in 9:51 minutes)!
Handbrake 0.9.6 video encoding test:
But there have been released a new edition of Handbrake since that guys thread started, so decided to give it a go in the new version also.
The old version favored more RAM (Corsair 16GB kit) were the new version favors faster RAM (Kingston 8GB kit 1867MHz).
None of them however seems to care about read/write speeds on your harddrive.