YouTuber DIY Project Shrinks M1 Mac Mini By 78%, Without Sacrificing Performance (9to5mac.com) 43
In a 15-minute-long video, YouTuber Quinn Nelson from Snazzy Labs explains how he managed to shrink the current M1 Mac Mini by 78% without harming performance. 9to5Mac reports: In conclusion, by rearranging the internals and swapping out the power supply, Nelson was able to reduce the size of the Mac mini enclosure by 78%. He organized all the parts inside a 3D-printed body with a mini Mac Pro motif.
The reason that theoretical space savings are so huge is because when Apple released the first round of Apple Silicon computers, they did not change the hardware industrial design at all. So the current Mac Mini enclosure is designed to fit an Intel CPU and circuit board, including having to accommodate the large fans and heat sinks the Intel chip required.
But with the power efficiency of the M1, Apple has the headroom to do something much more drastic. Indeed, a lot of the M1 Mac mini internals is just empty space. The Snazzy Labs video gives a glimpse at what is possible if Apple is more ambitious with the next-generation Mac mini design, and tries to create something truly mini. The CAD files and schematics can be viewed here.
The reason that theoretical space savings are so huge is because when Apple released the first round of Apple Silicon computers, they did not change the hardware industrial design at all. So the current Mac Mini enclosure is designed to fit an Intel CPU and circuit board, including having to accommodate the large fans and heat sinks the Intel chip required.
But with the power efficiency of the M1, Apple has the headroom to do something much more drastic. Indeed, a lot of the M1 Mac mini internals is just empty space. The Snazzy Labs video gives a glimpse at what is possible if Apple is more ambitious with the next-generation Mac mini design, and tries to create something truly mini. The CAD files and schematics can be viewed here.
Without Sacrificing Performance? Sure. (Score:5, Interesting)
Without Sacrificing Performance
So the current Mac Mini enclosure is designed to fit an Intel CPU and circuit board, including having to accommodate the large fans and heat sinks the Intel chip required.
This mod appears to incorporate no active cooling at all.
Just because the M1 doesn't require active cooling doesn't mean it doesn't benefit from it, and just because it doesn't sacrifice performance with light use does not mean it wouldn't sacrifice performance under sustained load. That much is clear from the performance of the M1 Air vs the 13" Pro and Mac Mini at otherwise identical specs.
I'm sure Apple could trim down the size of the case, but I highly doubt this mod doesn't come with some significant sacrifices under sustained load.
Re: (Score:2)
Without Sacrificing Performance
So the current Mac Mini enclosure is designed to fit an Intel CPU and circuit board, including having to accommodate the large fans and heat sinks the Intel chip required.
This mod appears to incorporate no active cooling at all.
Just because the M1 doesn't require active cooling doesn't mean it doesn't benefit from it, and just because it doesn't sacrifice performance with light use does not mean it wouldn't sacrifice performance under sustained load. That much is clear from the performance of the M1 Air vs the 13" Pro and Mac Mini at otherwise identical specs.
I'm sure Apple could trim down the size of the case, but I highly doubt this mod doesn't come with some significant sacrifices under sustained load.
But it is obvious that he could simply increase the dims of the case by just a little, and can plop this 20 cfm 36 dB 60 x 10 mm 12 V fan right on top of the heatsink.
https://www.amazon.com/Evercoo... [amazon.com]
More than enough extra cooling!
Re: (Score:2)
What about the Apple Engineers that Improve on their old designs?
While I fully expect that to be a sarcastic comment, on today internet we can never know.
Re: (Score:2)
What about the Apple Engineers that Improve on their old designs?
While I fully expect that to be a sarcastic comment, on today internet we can never know.
Of course it's sarcastic (and idiotic); he posted AC.
Check my Username for my opinion on (most) ACs..
Re: (Score:2)
I'm sure Apple could trim down the size of the case
Or they could add some expansion capability. The M1 can't take more RAM, but it could take more SSD or HDD storage, as well as PCIe cards.
Re: (Score:2)
I'm sure Apple could trim down the size of the case, but I highly doubt this mod doesn't come with some significant sacrifices under sustained load.
If you watched the video you'd see that the current M1 cooling and power solutions are identical to the the previous Mac Mini's despite using only a fraction of the power and having a fraction of the thermal load.
No you're not sacrificing performance, unless the previous Mac mini was woefully undercooled and power constrained ... which it wasn't. Apple sacrificed size all to prevent any major retooling or part redesign.
Re: (Score:2)
While Apple sure could afford it ten times over without even registering a "dent" in their funds, re-tooling costs quite a lot of money for the rest of us.
However, I expect it more to be a style choice. Apple products have a certain look/style. They did their due diligence coming up with this look/style and invested in this.
The new M1 devices do not need a re-think of the look/style, the new hardware fits inside the "old" case and the amount of material used in the case acts as heatsink, making the new har
Re: (Score:2)
There's a difference between not affording something and giving a shit. The Mac Mini is an absolute after-thought of a product in Apple's line up. I'm honestly surprised they haven't just dropped it altogether.
Re: (Score:2)
I immediately thought of cooling, too.
Apple likes no fans, but that requires careful design of the case for vent and airflow.
And then there's the 3D print plastic itself. How does it handle heating and outgassing from the material itself?
Re: (Score:2)
I can take my Desktop PC, take all the internals out of the case and make a smaller computer from it.
I don't think when Apple designed the Mac Mini they were trying to make it as small as possible, just small enough to not get in the way. Have room for cooling and strong enough to deal with general shock and abuse that small devices tend to get. Such as taking your Mac mini, and tossing it in a satchel, purse or backback. As well I am sure they wanted to keep it a perfect square shape, because it is just m
Re: (Score:1)
Cool, though swapping out the power supply (Score:2)
Re: (Score:2)
What worries me more is that with no battery to fall back on, if you bump the cord it's lights-out. There aren't even any substantial capacitors inside there to even give it a smidge of power to tide it over while the contacts are dragging. I don't know if they had any space left over inside the case, but I'd be doing my best to smuggle in the biggest electrolytic I could find a place for.
Re: (Score:2)
Re: (Score:2)
On the other hand, now that nearly all of them have SSDs, the last moving part in the chassis is the optional fan. Power supplies tend to be a less reliable part of a modern personal computer, modulo the bathtub curve [weibull.com]. So if you move the power supply out of the case, you can keep a spare on hand [thedailywtf.com] and just swap them out, rather than having to reopen the chassis. Probably not such a big deal for something Mac Mini-sized, though.
Over the past two decades, I've had zero failures of Apple power supplies, unless you count short-circuited cords on laptop power supplies (which were the bane of my existence prior to the USB-C transition, when the cords became independently replaceable). If you go just past the twenty-year mark, I did have to replace capacitors on the logic board of a graphite Airport base station around the turn of the century. I wouldn't call that part of the power supply (it used a wall wart), but it is at least powe
Re: (Score:2)
Re: (Score:2)
I find it a cheat overall.
Sure, they made a smaller Mac Mini... by removing the power supply. You didn't make it smaller, you just removed a component and brought it outside. I'm sure I could make a smaller Mac Mini by taking the motherboard out of the case and calling it good.
Granted, the original PowerPC Mac Minis had external power supplies, but the Intel ones had built-in power supplies.
So it's less making it smaller and more "we just broke it in two parts, and one part is smaller if you don't include t
Re: (Score:2)
Small version of a decade old system? (Score:3)
This is like saying you can fit a Vax into a PC desktop case.
The new m1 mac mini could only run my test software at double the speed of the decade old Intel 2011 mac mini. Considering the new one costs twice as much, it has been a decade and constant raving reviews, I would have expected much better performance. My old slow Ryzen is much faster. The code I was running was compiled for 64bit arm as well so it isn't an emulation problem. An easy test to run is use 'openssl speed' of something like MD4 which can't make use of dedicated encryption hardware and gives a reasonable integer speed test without building test programs.
Re: (Score:2)
Re: (Score:2)
The new m1 mac mini could only run my test software at double the speed of the decade old Intel 2011 mac mini.
Sounds like your workload doesn't benefit from multiple cores... at all.
I could easily believe that you're only able to get two-thirds the expected 3.4x speedup on a single-core app. If you're only getting a 2x speedup on a multicore task, you're doing something wrong, because the 2011 Mac Mini was a 2-core dog, versus ostensibly up to 8 cores on the M1, de
Re: (Score:2)
I find safari with ad blocking on the new machine much slower than safari on the older one. Chrome and firefox are both much faster. A vast majority industry specific software is single threaded integer and much of the multi threaded only runs a small number of threads occasionally.
As far as benchmarking, I prefer throwing real software at it like openssl's speed test. It has the advantage of being on everything and easy to run even and can test multi-core performance.
Back when we got an Intel hypercube u
Re: (Score:2)
I find safari with ad blocking on the new machine much slower than safari on the older one. Chrome and firefox are both much faster. A vast majority industry specific software is single threaded integer and much of the multi threaded only runs a small number of threads occasionally.
Integer performance does indeed trump floating-point and other functionality when it comes to responsiveness. We saw that in how much snappier Mac OS X v10.4 was on Intel hardware than on PowerPC. However, I'm not finding Safari to be slower on my M1 Max. I find it to be considerably more responsive than on a 2019 MacBook Pro with a 2.4 GHz i9. Perhaps the problem is something specific to your ad blocker? (Or maybe you're running Safari emulated?)
That said, all those apps that only occasionally use mul
Re: (Score:1)
As far as benchmarking, I prefer throwing real software at it like openssl's speed test.
OpenSSL uses up ~0.01% of CPU time on desktops and benchmarks essentially just say the M1 doesn’t have AES instructions enabled. https://robservatory.com/my-im... [robservatory.com]
If you are using a Mac directly as a web server then OpenSSL scores are stupidly important, otherwise they are just stupid.
Re: (Score:2)
(But of course, the most important reason to upgrade is that it doesn't have a touchbar. :-D )
Obviously!
I would not have minded if it had a touchbar and ordinary function keys.
But it was plain for me I never would buy a computer that has no function keys.
Re: (Score:2)
(But of course, the most important reason to upgrade is that it doesn't have a touchbar. :-D )
Obviously!
I would not have minded if it had a touchbar and ordinary function keys. But it was plain for me I never would buy a computer that has no function keys.
Oh, absolutely. The lack of tactile access to critical stuff like screen brightness was a big loss, though not nearly as much as the painful rate of false triggering when typing numbers. For example, in Chrome, typing certain numbers in the search/address bar can cause your finger to graze the touchbar button that moves your cursor to the search/address bar and selects its current content, resulting in your search string getting wiped on the next keystroke. And in Finale, various chorded keystroke combin
Re: (Score:2)
I never tried one with a touch bar.
But your summary is more or less what I expected.
My function keys are set to produce the ordinary F1 - F12 codes, and to use volume or brightness, you have t hit the FN-key.
Worst thing with the original touch bar was the missing ESC key (they later had models that had an physical ESC key again) I mean seriously: the vi versus EMACS war is long over, and vi won. How should I work without an ESC key? Adding an external keyboard for just one key? I actually was googeling arou
Re: (Score:3)
Aside from the fact that most workloads see a huge increase in speed on the M1 (not sure why you are only seeing double the speed), what's your point? Most PC loads haven't been CPU constrained for a decade anyway, so while your strangely inefficient workload may not be suitable for this, it would none the less meet the needs of *most* people.
Re: (Score:2)
Not sure how you get just twice the performance of the 2011 Mac Mini - I replaced an old Mac Mini with the M1 a bit after it came out and it was 3-4x faster in single-thread tasks. It is actually over 2x faster in many things compare do the 2015 i7 Macbook Pro - I even did a blog post on perl.com [perl.org] with the relative speed in various Perl benchmarks.. It was significantly faster than my Ryzen 3500U Thinkpad (Zen+), so I upgraded to a Zen 2 Thinkpad (the X13 with a Ryzen 4650U, a beautiful machine), which is st
Re: (Score:2)
There could be some code that doesn't really improve in speed on newer processors. ...
But I assume it is waiting on some external resources - so it waits for half the time, and runs four times faster the other half of the time.
External resources: "remote" file systems, Git repositories, writing test reports somewhere else,
Re: (Score:2)
That is most likely a single threaded "test", that runs only one one core on each machine, and says you nothing about the raw speed of the machine.
High power test (Score:2)
Re: (Score:2)
Given the size of the MacBook Air enclosure, one suspects that this one is also fine.
How small does a desktop computer need to be? (Score:2)
Ok...fine, cool project. But honestly....how small does a desktop computer need to be?
My Mini already takes up practically no space on my bookshelf. If it took up a quarter of the space it takes up now, it would change absolutely nothing for me.
Re: (Score:2)
Better than smaller would be... (Score:2)
I agree. It's a desktop - it's likely to have all sorts of cables attaching it to various peripherals, and it needs to be both large enough for all those connections to be on the back, and heavy enough that they don't drag it around.
And on the aesthetic front you almost certainly want a footprint big enough to stack neatly with a matching external optical-disc drive, for those users who want one. It's also a convenient footprint for external hard drive bays, with plenty of room either a 3.5" or two 2.5" dr
Still a pig in any clothes (Score:2, Troll)
Whats apple doing about it? (Score:2)
Still too big (Score:2)
There are android tv boxes half the size. The ARM cpu is not as powerful, but still the same architecture.