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.
PSA for Developers: Mac Mini With A12Z Chip Cannot Be Repaired at Genius Bar or Service Provider
- Thread starter MacRumors
- Start date
- Sort by reaction score
So are you "not allowed" to run benchmarks or is the machine "not capable" of running benchmarks.
It'd be funny if you request one because you are a developer... of a benchmarking software.
It'd be funny if you request one because you are a developer... of a benchmarking software.
27" iMac has easily accessible RAM and the Mac Mini is a little complex but can be done at home. The 21.5 iMac and iMac Pro can be done but requires ungluing.
If such a developer were to apply for a kit, perhaps they might be rejected from the program. But that said, IMO there is no harm in running such software against the dev kit, but the terms would prohibit making public any results.
Re performance: Whether or not the kit itself is slow or fast is actually irrelevant because it is there to ensure a rebuilt application runs under the new architecture, where devs are expected to be using APIs that abstract away CPU behavior. There's every chance that first public Apple hardware does not use the same processor nor accompanying chip sets.
The $500 fee is probably there to cover the administrative costs of managing the program and I would not expect to get it back. If not getting $500 back is a problem, I would revisit one's business plan
For most devs, they will have to trust that Xcode will produce proper runnable universal binaries. Did Apple indicate just how many of the kits are available? Clearly not every dev organization will be able to get one, nor should that be necessary.Totally agree, they should at least not be leagues better. I once had a senior developer turn to me as the then systems and networks lead and tell me his software was ‘working as designed’ when it after an upgrade increased memory usage exponentially and crippled our ops machines. All the computers at the time were Win XP machines with I seem to recall what was a healthy 512MB RAM and of course his had 4GB (another funny story about available memory there with the same guy but I digress!). It took our MD to refuse to sign a PO for the upgrade before he agreed to optimise the code, then he of course got a junior to do it and memory usage went nicely below something silly like 8MB as he optimised it a whole lot. The senior guy didn’t stick around too long after that!
Here’s a benchmark I would like to see: Take a large project that takes ten or twenty minutes to build. Report the exact build time with Intel Xcode on Intel, with Intel Xcode on ARM, and with ARM Xcode on ARM. That would help me deciding what my next MacBook is.
[automerge]1593034192[/automerge]
There are various dual core Mac models for sale now, I can see them being replaced with something very similar to A12 at the low end. With a14, six or eight fast cores, replacing everything else up to 8 Intel cores.
Last edited:
That's a benchmark that makes sense when the new MacBook or MBP is released in its final form to devs and reviewers. What good is knowing if the dev kit is slower or faster when release hardware might not be anything like it?
But certainly, if the new equivalent of a 13" MacBook Pro is slower than the current one, there will be hell to pay.
This DTK much more compact compared with G5 enclosure with Pentium back then.
Imagine if Apple using 7,1 enclosure for DTK, people can easily peek out what is inside due upgradable nature, but it won’t practical due size of case and leaving so much empty room inside so using Mac Mini form factor seems reasonable, and Mac Mini is also semi-sealed system, prevent anyone who trying looking inside albeit it can be done with some efforts.
Imagine if Apple using 7,1 enclosure for DTK, people can easily peek out what is inside due upgradable nature, but it won’t practical due size of case and leaving so much empty room inside so using Mac Mini form factor seems reasonable, and Mac Mini is also semi-sealed system, prevent anyone who trying looking inside albeit it can be done with some efforts.
For sure. For the purposes of dev validation though, I'm surprised the DTK isn't the size of an Apple TV. Maybe needs room for the port contingent?
I think final this dev kit usually slower than final released product, remember G5 Trojan Horse are using Pentium 4 which doesn’t make it way in any actual of Intel Macs product. First Intel Mac released are using more powerful Xeon chips.
Just my assumptions though, probably this make sense for some hardware dev who needs and re-writing driver compatibility with ARM architecture, at last they need port for plugging them into host machine to interact with.
You know, get a clue. This test machine in no way represents what the final product will be. Apple doesn't want benchmarks published because the loser techies of this world will claim this test machine is the same as the released product and therefore it's a dud as far as benchmarks go. This machine is for developers to test their code on, not the finalized, optimized product that goes on sale at the end of the year. You know how this works in the fake techie universe.
This. The hardware is not the remarkable thing, by any means. The software is. Apple is just providing developers with “minimum viable product” hardware that can run the software. The reason Apple doesn’t want people benchmarking these machines is because those benchmarks will be meaningless, yet people will start treating them as meaningful. They won’t reflect the final hardware at all.
That won't stop people here from completely misunderstanding that and believing this is what the final product will be. The current OS is also far from optimized in the way the retail product will be.
100% guarantee there will be an article published here on MR with benchmarks and people will be in a tizzy over it.
Apple has no such thing as “Apple Silicone”, unless you are somehow referring to their iPhone cases. A silicone based processor would be squishy and would likely perform quite poorly. Silicon, on the other hand, is an excellent material for making high performance processors.
Totally. Apple will try to spank someone who releases the data while there will be MacRumors angst. Oh, the horror!
I'd hope that, upon return of the DTK mac, they'll at least give you $500 credit on a new Apple Silicon mac.
I'm kind of tempted to get the DTK as it would be nice to play around with the new hardware and build/test my app on it. But I'm not sure if it's worth $500 to me considering the hardware must be returned...
Yes it is. It’s an a12z. We are talking about the devkits, not the actual Macs that will be for sale.
I WAs thinking the same thing. It just looks like a mistake.
How would that help? This isn’t the machine they’re going to end up selling and it’s also running dev beta software. That would be a truly useless test.
There is ZERO guarantee that you'll get a $500 refund, credit, or new machine. Just because they send iMacs to those that dropped $1000 on an Intel development machine does not in any way mean the same will happen this time.
Unless you're prepared to be completely parted with that $500 when the new machines ship and have nothing to show for it other than some time playing around and testing your app, don't gamble the money on something that's far from a sure thing.
[automerge]1593041097[/automerge]
Exactly. It'd be a worthless test. "Here are the results that will be absolutely nothing at all like what will be available for sale."
If you want to know how it'll perform in the future, wait until retail ARM machines ship and then see real-world comparisons.
Apple is too scared that folks will benchmark based on a Windows 10 ARM build and the chips are too dog slow to play Fortnite.
[automerge]1593041322[/automerge]
Not really. Even Intel and AMD have test engineering samples for their CPUs.