Mac mini 2018 Review: Apple's mightiest mini yet

123578

Comments

  • Reply 81 of 151
    Mike WuertheleMike Wuerthele Posts: 3,608administrator
    entropys said:
    entropys said:
    you know, all apple had to do was have both the RAM and the SSD in slots, and I think that would be a worthy Mac Mini. RAM only is only half way there.  Apple should be ashamed of what they charge for SSD upgrades. Ashamed.

    So, if I was to pick a configuration, I would probably go for an i5, 8GB RAM, 128GB SSD, then add an extra 8GB RAM myself and have a decent sized thunderbolt 3 external drive to boot off cabled under the desk.
    No external drive would match the internal NVMe speed, even on Thunderbolt 3. So it is understandable that Apple reasonably fix their prices according to “competition” from external devices. Show me an external TB3 drive faster than the internal SSD then I’ll agree with you.
    It’s a Mac mini. I’ll repeat that. It’s a Mac mini.  I don’t care if the Mac Mini’s TB3 connected SSD is marginally slower than the soldered internal drive.  It will drop from ludicrously fast speeds to just astonishingly fast speeds. See Mike’s comment above where he is getting 40 gigabit/s throughput. TB and NVMe mean this ain’t SATA speeds anymore. The only negative will be still thinking of the prices Apple charges for SSD in their cheapest computer every time I boot it up.
    Ask him also how much does it cost. This is an Intel Optane SSD 905P (960 GB). Thunderbolt box not included.
    That particular config, the box is included, and it is not cheap. The drive as pictured is $1300, with the enclosure if you want it separately for $199.
  • Reply 82 of 151
    MichalPfeilMichalPfeil Posts: 5unconfirmed, member
    Looks like they got rid of the infrared port. Is this correct? Does this mean one can't use the apple remote with it for home theater use like controlling Plex or what not? 
  • Reply 83 of 151
    entropys said:
    entropys said:
    you know, all apple had to do was have both the RAM and the SSD in slots, and I think that would be a worthy Mac Mini. RAM only is only half way there.  Apple should be ashamed of what they charge for SSD upgrades. Ashamed.

    So, if I was to pick a configuration, I would probably go for an i5, 8GB RAM, 128GB SSD, then add an extra 8GB RAM myself and have a decent sized thunderbolt 3 external drive to boot off cabled under the desk.
    No external drive would match the internal NVMe speed, even on Thunderbolt 3. So it is understandable that Apple reasonably fix their prices according to “competition” from external devices. Show me an external TB3 drive faster than the internal SSD then I’ll agree with you.
    It’s a Mac mini. I’ll repeat that. It’s a Mac mini.  I don’t care if the Mac Mini’s TB3 connected SSD is marginally slower than the soldered internal drive.  It will drop from ludicrously fast speeds to just astonishingly fast speeds. See Mike’s comment above where he is getting 40 gigabit/s throughput. TB and NVMe mean this ain’t SATA speeds anymore. The only negative will be still thinking of the prices Apple charges for SSD in their cheapest computer every time I boot it up.
    Ask him also how much does it cost. This is an Intel Optane SSD 905P (960 GB). Thunderbolt box not included.
    That particular config, the box is included, and it is not cheap. The drive as pictured is $1300, with the enclosure if you want it separately for $199.
    Twice the price of basic Mac Mini, (almost, @ $1300). Apple’s 1 TB internal SSD upgrade is $800...
    edited November 6 williamlondon
  • Reply 84 of 151
    Mike WuertheleMike Wuerthele Posts: 3,608administrator
    Looks like they got rid of the infrared port. Is this correct? Does this mean one can't use the apple remote with it for home theater use like controlling Plex or what not? 
    That is correct.
  • Reply 85 of 151
    anomeanome Posts: 1,158member
    lewchenko said:

    While I like this product marginally better than the revised MacBook Air, I feel that Apple is still missing the mark with its price point on the Mini. An existing Windows user should be able to unplug their existing box and plug in a Mac mini at a price that doesn't break the bank. You can get a powerful PC desktop for this price.
    When Apple produced a very-cheap mini for switchers, the complainers said it wasn't powerful enough. Now they've gone more powerful, and the complainers say it costs too much. Do you see the catch-22 here? Anyway, I'd argue the mini is no longer a cheap intro product designed for switchers -- that job was fulfilled by iPods, iPads, and iPhones. This is a small Mac.
    It doesn’t have to be a catch 22. 
    Apple decides the price and spec and could have easily accommated an affordable specification in the line up. 
    But that’s not who Apple are. Ultimately they don’t care about low end switchers. By not having one forces people to pay more or go elsewhere.
    The price of these things buys quite a powerful PC in comparison. People have to decide whether the Mac OS is worth the premium over and above a PC (far more expandable no doubt)

    to sum up apple these days - “for people with boat loads of money, or those willing to consider older tech for slightly less money - eg old Air, older 7 iPhone, older mini” 

    Except it really is a Catch-22. Apple are always going to be judged on the cheapest/lowest spec model. So, let's say they put out a US$500 Mac mini as part of this announcement. Maybe with a dual core i5 and 2 TB3 ports to keep costs down. Then the complaint would be that this model only exists so they can claim an artificial price point of US$500, and no-one is going to buy it. Plus, it costs more than US$200 to upgrade to the actual model people will buy. Some will say that's worth it, others will say it's not, and then we get into Apple overcharging for upgrades, as we are now. They charge too much for SSD, and RAM. (I agree on RAM, certainly, SSD I'm less convinced, but they are expensive.)

    This isn't entirely speculation, it's extrapolation from previous Apple product releases. They announce a new product with a base price at some landmark (US$1000, US$500, whatever) and the pundits all agree that particular model is only there so they can claim a "Starting at..." price tag. And you only have to look above to see the complaints about how much they charge for upgrades.

    jcs2305 said:
    Fatman said:
    Fatman said:
    I really don't get why Apple refuses to change the external design of their products for years - the internals are completely redesigned on this mini, but it looks identical to models of past. The same can be said for the iMac, Macbook Air. Maybe they need to hire a new enclosure designer to freshen things up?
    Ive has already spoken to this many times, you just weren't paying attention. 

    “It starts with the determination not to fall into the trap of just making things different. Because when a product has been highly regarded there is often a desire from people to see it redesigned. I think one of the most important things is that you change something not to make it different but to make it better.”

    https://www.independent.co.uk/life-style/gadgets-and-tech/features/jony-ive-interview-apple-ipad-new-macbook-air-mac-tim-cook-event-a8614421.html
    Yes, I agree, redesign to make better - not change for change sake ... they've had YEARS to work it out. Ive also made an awful AppleTV remote (which side is up? Lets add a white circle) - complete fail. A poor first Apple pencil, maybe they got it right on take 2? An iphone with a notch, and rounded screens to cut off content. An iphone case with a hump. An Apple mouse with the lighting port on the bottom. Do I need to go on. Get rid of this guy, his time is up. How about balancing ergonomics with his pretty designs - that's where real genius comes in. One more thing  ... the bad tables at the Apple stores - try sitting at a corner. For those that know my posts - I am extremely invested in Apple, and own an 'irresponsible' number of shares in the company, but bad decisions need to be challenged.
    Is this a joke?  I can't tell anymore..who is joking and who is trolling?
    Is "Sack Jony Ive" the new "Sack Tim Cook"?
    netmagewilliamlondon
  • Reply 86 of 151
    What about the level of fan noise?
    williamlondon
  • Reply 87 of 151
    entropysentropys Posts: 1,346member
    entropys said:
    entropys said:
    you know, all apple had to do was have both the RAM and the SSD in slots, and I think that would be a worthy Mac Mini. RAM only is only half way there.  Apple should be ashamed of what they charge for SSD upgrades. Ashamed.

    So, if I was to pick a configuration, I would probably go for an i5, 8GB RAM, 128GB SSD, then add an extra 8GB RAM myself and have a decent sized thunderbolt 3 external drive to boot off cabled under the desk.
    No external drive would match the internal NVMe speed, even on Thunderbolt 3. So it is understandable that Apple reasonably fix their prices according to “competition” from external devices. Show me an external TB3 drive faster than the internal SSD then I’ll agree with you.
    It’s a Mac mini. I’ll repeat that. It’s a Mac mini.  I don’t care if the Mac Mini’s TB3 connected SSD is marginally slower than the soldered internal drive.  It will drop from ludicrously fast speeds to just astonishingly fast speeds. See Mike’s comment above where he is getting 40 gigabit/s throughput. TB and NVMe mean this ain’t SATA speeds anymore. The only negative will be still thinking of the prices Apple charges for SSD in their cheapest computer every time I boot it up.
    Ask him also how much does it cost. This is an Intel Optane SSD 905P (960 GB). Thunderbolt box not included.
    That particular config, the box is included, and it is not cheap. The drive as pictured is $1300, with the enclosure if you want it separately for $199.
    Twice the price of basic Mac Mini, (almost, @ $1300). Apple’s 1 TB internal SSD upgrade is $800...
    If you shop around you could do it for half that with a basic TB3 enclosure. Let’s spend one minute on Amazon for fun.

    Here is a TB3 enclosure $160

    Here is is a 1TB Samsung 970 EVO NVMe SSD. $228.

    With more investigation a likely lad could probably get an external setup with even higher performance than these for half apple’s price, and the kicker is it would be future upgradable.

    stompylorin schultz
  • Reply 88 of 151
    chasmchasm Posts: 963member
    This machine appears — at least for typical tasks — to be comparable in performance to the current Mac Pro in terms of CPU performance. Think about that for a minute — this is a quad-core i3, and it’s in the neighbourhood of an aging but still-produced Xeon. For around 1/4th the price of a Mac Pro, and half the price of a current MacBook Pro, and cheaper than the NUC-type PC mini (in some cases by up to 50 percent!).

    There are obviously some areas where the mini will come up short compared to those machines, but for mainstream computer users this is actually a tremendous value and a justified price for this level of upgrade, IMO.
    edited November 6
  • Reply 89 of 151
    wizard69wizard69 Posts: 12,627member
    sirozha said:

    sirozha said:
    I believe that the quad-core i3 CPU in this Mac Mini has only four threads instead of eight threads; that is there’s only one thread per core. Please confirm that.

    If this is the case, the multi-threading performance of this i3 CPU should be inferior to the multithreading performance of the 2012 Mac Mini’s quad-core i7, which has eight threads (two threads per core).  


    While it does have four threads -- one per core -- your performance assumption is not. The multi-core performance, which encompasses multiple threads, is better on the new mini.

    This shouldn't be much of a surprise, given that the processor is five years younger.
    What about four threads vs eight threads for virtualization purposes?
    SMT isn’t always the big win many think it is, successful SMT behavior Is application dependent.    Beyond that virtualization adds more complexity to the discussion.   Look at it this way some installations will turn off SMT due to performance regressions.  
    netmage
  • Reply 90 of 151

    When you do the bench marking, really want to see boot times of the PCIe internal storage and external Thunderbolt against iMac pro SSD.  At this point Apple is using off the shelf Intel processors.   Maybe for the money they customized the controller to max out the PCIe lanes to the processor. This in conjunction with the 10 gigabit Ethernet would open some interesting doors for Mac software developers.  With Intel pushing Optane technology, maybe Apple baked in a surprise of their own.  I suspect this is Apple test driving architecture enhancements for the Mac Pro.

    For perspective, when the redesigned 2010 "flat mini" series was introduced it started at $699 ($805 in 2018 money) for a base Core 2 Duo 2.4, only went to $499 after the 2nd refresh in 2014.  This being the "recycled" :) series maybe a future refresh will bring a low cost mac mini.


    stompy
  • Reply 91 of 151
    Rayz2016 said:
    Fatman said:
    Yes, I agree, redesign to make better - not change for change sake ... they've had YEARS to work it out. Ive also made an awful AppleTV remote (which side is up? Lets add a white circle) - complete fail. A poor first Apple pencil, maybe they got it right on take 2? An iphone with a notch, and rounded screens to cut off content. An iphone case with a hump. An Apple mouse with the lighting port on the bottom. Do I need to go on. Get rid of this guy, his time is up. How about balancing ergonomics with his pretty designs - that's where real genius comes in. One more thing  ... the bad tables at the Apple stores - try sitting at a corner. For those that know my posts - I am extremely invested in Apple, and own an 'irresponsible' number of shares in the company, but bad decisions need to be challenged.

    The Apple TV remote. Right, you're the second person on this forum who seems to be foxed by this, so here's a quick cheatsheet.

    If you can feel the smooth bit at the top, you have it the right way up. 
    If you feel the glossy bit at top, you haven't
    If you can't feel the buttons at all, then you have it face down.
    If you can't feel the remote at all, then you've dropped it.
    If the remote feels bent, then you're holding a banana.

    Hope that helps.

    The first Apple Pencil was good. The second one was better. That's how product development works.

    The iPhone has a notch because Apple knew that the only people who'd be bothered by it are fake design expert wannabes.

    The iPhone case had a hump because no one has come up with a battery with Tardis functionality

    The Apple Mouse has the lightning port on the bottom because a mouse with a hole you can see would look like crap (the inconsistency in your thinking is beyond astonishing).

    Oh, and as for the tables at the Apple Store? Don't try sitting at the corner; try sitting at one of the flat sides where you're supposed to sit.

    Christ on a bicycle … 
    You’ve made my day with the Siri Remote instructions. 
    ntwrkd
  • Reply 92 of 151
    wizard69wizard69 Posts: 12,627member
    I’m actually wondering why Apple does offer a diskarray enclosure to vowing the Mini.   I’m not bothered by a soldered in main SSD if it is large enough.  What is needed is expandability off the Mac main housing and frankly all Mac products needs such. It would be simple to take the Mini housing an install a mother board with say four SSD slots.  Such a storage array would need on TB port and a power connection.  

    In any any event this new Mini has me stoked.  It is a much better upgrade than I expected.  Hopefully this means Apple has gotten serious about Mac engineering again.  MBA kinda undermines that hope  so who knows for sure.  
    Now I need a mac-mini-shaped hard drive enclosure to put a 6tb hard drive with my media files in there, and stack it under the actual mac mini. My life will be complete. 
    entropys
  • Reply 93 of 151
    hodar said:
    Graphics wise, I understand it's not the most powerful GPU on the planet; but how well will it perform? Given the speed of Thunderbolt 3/USB C; having an internal drive that is huge isn't that great of an issue, when you can just add a large HDD or SSD to a USB 3, or a USB C Drive Hub.
    For consideration:
    https://gpu.userbenchmark.com/SpeedTest/356797/IntelR-UHD-Graphics-630
    In context a 2010 iMac still seems to rate better...
    edited November 6 williamlondon
  • Reply 94 of 151
    wizard69wizard69 Posts: 12,627member
    mike54 said:
    It took Apple 4 years to upgrade the internals, no design work done on the case or design. Apple did the min amount of work they could to update this. Ok the internals are new, but why the crappy graphics? It should have at least have the Intel Iris Pro graphics. I didn't expect this behaviour from Apple. How can people say this is a premium product when Apple does stuff like this.
    The Mini case is almost an industry standard an is very popular in embedded applications.  At least it was, Apple dragging feet for years didn’t help its position in the modern world.  

    In any event, except for the low end this comes as close as I’ve ever seen of Apple producing the ideal Mini.   The last update Apple stressed a GPU update over main processor functionality and it was a complete flop.  With the new Mini Apple now has a machine that can handle real midrange desktop usage.  It is rather impressive that you can now buy a Mini with performance that Eclipses most Macs.  
  • Reply 95 of 151
    wizard69wizard69 Posts: 12,627member
    Rayz2016 said:
    Fatman said:
    I really don't get why Apple refuses to change the external design of their products for years - the internals are completely redesigned on this mini, but it looks identical to models of past. The same can be said for the iMac, Macbook Air. Maybe they need to hire a new enclosure designer to freshen things up?

    Because, as shown in the keynote, there are customers using thousands of these in rack mounts. Changing the design of the case would cause serious headaches.
    It isn’t just rack mounts, Minis show up in all sorts of embedded applications.  It is basically a standardized module now.   It would be foolish of Apple to build a new enclosure that screws over abit part of the market for these computers. It makes about as much sense as changing the design of a wall outlet.   
    netmage
  • Reply 96 of 151
    dewmedewme Posts: 1,763member
    blastdoor said:
    dewme said:
    blastdoor said:
    dewme said:
    blastdoor said:
    dewme said:
    This is a worthy upgrade that predictably builds on Apple's reputation of delivering premium products at a premium price. It is now safe to say the Mini is  not going to be Apple's budget or entry level product that some folks were hoping for. It's one thing for us to say "the ecosystem could use a less expensive one, at a $499 or $599 price-point" but quite another to identify what we're willing to give up on the lowest end configuration to hit that price point. By the way, Apple accepting a lower margin is not a viable answer and is not going to happen. This is still a premium engineered, amazingly industrialized, and thoughtfully sourced product that Apple put a tremendous amount of resources into to preserve what Apple values in a product it is proud to present to customers. They are not simply slapping a bunch of off-the-shelf components into an ugly white box that's stuffed with whatever they can get cheapest this week.

    I'm still hopeful that Apple can find a way to justify delivering a true entry level machine. To do this they will need to approach the problem from a different angle. For example, if Apple decided to significantly ramp up its investment in keeping newbie programmers on the professional development path they could consider something like a Mac "Coding Machine" that consisted of a what is basically a scaled down and spec'd down Mac Mini, perhaps in a poly-carbonate case, that is priced say at $299. Perhaps it only has i3 and fewer ports. I'm thinking something that is in the same sphere of focus as the Raspberry Pi but a full Mac OS machine (with XCode) that is more akin to what professional programmers are using compared to an iPad with Swift Playgrounds, which I as more like Lego Mindstorm for early exposure to programming. Apple needs to create an easy and natural path from Playgrounds to XCode, so having an educational focus coupled with an entry level product relieves Apple from having to have the "premium" wick turned up quite so high as what we see with the new Mini. Apple can do it, but needs to redefine the rationale that's driving their engineering and product development machine to make it happen. Education, STEM, robotics, and "everyone can code" may be exactly the rationale they need.

    Of course an entry level Mini won't please the current crowd here at AI, but Apple can't please everyone all of the time and they still have a business to run. We CAN pay for premium, even though we like to complain about it from our thousand dollar phones.

    The "easy" path to an entry level Mac at a lower price point is to dump Intel. Intel's gross margins are over 60%; TSMC's gross margins are just under 50%. Every time Apple ships a product with a processor fabbed by Intel rather than TSMC they are paying an "Intel tax" of sorts. An A12-based Mini would be a very strong entry level machine at a lower price point. An A12X could be offered as a BTO option for higher performance. 
    I could not agree less that the "easy" path for a high performance low-cost Mini involves a transition to an A-series (ARM) architecture, unless of course you want the Mini to be a headless version of the iPad Pro. There's a high level of Benchmark Fallacy going on right now around Apple's A-series chips. They are amazing components in their own right but don't fall into the trap of believing that just because an A12X can run benchmarks that run circles around the majority of consumer grade Intel based systems that they are a drop-in replacement for Intel x86 chips. Not so.

    Where these benchmark fueled fantasies and marketing "stretches" blow up miserably is in the domain of Software Reality. Moving a very large number of beefy line-of-business applications, not to mention core OS and machine bound services, and third part apps from x86 to ARM is very non-trivial, ridiculously expensive, and very time consuming - like they would have already had to have started 5 years ago to have a shot at hitting dates before the mid-2020s. Perhaps Apple actually started the train rolling 5 years ago, but we don't know. What about partners in the ecosystem? Do you want to rely on warmed-up iOS versions of critical business apps until the ecosystem finally catches up?

    Software developments costs can and often do dwarf hardware development costs for systems. There is no software TSMC fab-for-hire equivalent for Apple to lean on and throw bucketfuls of cash at to expedite software development progress and unburden its in-house development resources from having to do everything themselves. While there have been gains in software development maturity over the past decade, software development in general is still one of the easiest ways to annihilate vast quantities of time and money and have little to show for it. Not to beat a dead horse, but where there are significant software development and integration risks involved things like gross margins between chip/chipset suppliers and Intel taxes are secondary considerations that only come into play once the costliest contributor to overall system costs are brought under control and you have a system that can actually be built before you burn through all of your R&D resources.

    Benchmarks are meaningless until you solve the software problem.


    Apple was a tiny fraction of its current self when it made the transition from PPC to Intel, and tinier still when it made the transition from 68k to PPC. And yet they did it, quite successfully. 

    Today, changing processors would be far easier than in the past. The code base is so much more modular and Apple has vastly more resources. Everything in tech is "hard" in the sense that it costs money, takes time, and requires qualified people. But a transition to ARM is not the herculean task you make it out to be. 
    Are saying you are willing to fund this transition with your money?
    I'm not sure what point you're trying to make other than the fact that Apple's cost are ultimately paid by Apple's customers. if that's all you mean, then I have already funded Apple's transition from 68k to PPC (when I bought my PowerMac 7500) and the transition from PPC to Intel, when I bought a MacBook Pro in 2006, a Mac Pro in 2009, an iMac in 2010, and another iMac in 2014. And I guess you could say I've funded Apple's development of ARM chips over the years with my many iOS device purchases. 

    But this is silly. The costs here just aren't that big, and the benefits are likely to be bigger than the costs. Did Macs suddenly get a lot more expensive when apple switched from PPC to Intel? No -- yet Apple was a smaller company then, selling fewer Macs, the software challenges of that transition were much greater, AND Intel chips were more expensive than PPC chips. 

    Let's suppose that the cost of transitioning to ARM is $1 billion. I think that's actually an overestimate, because so much of the cost is already covered by development efforts for the A chips. But whatever, let's say $1 billion. There are about 20 million Macs sold every year. Spreading $1 billion over two years of Mac sales means about $25 per Mac. The cost savings from paying TSMC to fab those chips rather than buying from Intel are very likely to be larger than that. 


    Good points. Some of the costs have been paid to flesh out iOS functionality over the past decade. The only point of disagreement is the magnitude of costs, not just for Apple but for the MacOS app ecosystem. I believe the x86 to ARM port will be much harder and costlier than the PPC to x86 port. PPC and x86 were both targeting full desktop environments while ARM has some obvious gaps that must be filled with brand new code rather than ported code.

    I also think the porting/migration cost recovery amortization for MacOS app developers is going to be much harder for non-Apple contributors to the ecosystem. Some of this will depend on whether an ARM version of MacOS supports x86 virtualization/emulation out of the gate. If I were one of the impacted software product vendors I would be much more likely to focus now on moving my apps to iOS and trying to tap into the larger market there instead of porting my existing x86 apps to an ARM version of MacOS. This is a big investment but there is a big potential payback.  What's the benefit for third parties to natively support ARM version of MacOS - other than to avoid being left behind? 
  • Reply 97 of 151
    That particular config, the box is included, and it is not cheap. The drive as pictured is $1300, with the enclosure if you want it separately for $199.
    Twice the price of basic Mac Mini, (almost, @ $1300). Apple’s 1 TB internal SSD upgrade is $800...
    Twice the price NOW.  6 months down the road, it will probably drop to $800.  1 year down the road, probably gonna cost you $400.  2 years?  Do the math.  Meanwhile, the Apple internal NVME remains. . . .wait for it .. . . THE SAME!
    williamlondon
  • Reply 98 of 151
    wizard69wizard69 Posts: 12,627member
    dewme said:
    blastdoor said:
    dewme said:
    This is a worthy upgrade that predictably builds on Apple's reputation of delivering premium products at a premium price. It is now safe to say the Mini is  not going to be Apple's budget or entry level product that some folks were hoping for. It's one thing for us to say "the ecosystem could use a less expensive one, at a $499 or $599 price-point" but quite another to identify what we're willing to give up on the lowest end configuration to hit that price point. By the way, Apple accepting a lower margin is not a viable answer and is not going to happen. This is still a premium engineered, amazingly industrialized, and thoughtfully sourced product that Apple put a tremendous amount of resources into to preserve what Apple values in a product it is proud to present to customers. They are not simply slapping a bunch of off-the-shelf components into an ugly white box that's stuffed with whatever they can get cheapest this week.

    I'm still hopeful that Apple can find a way to justify delivering a true entry level machine. To do this they will need to approach the problem from a different angle. For example, if Apple decided to significantly ramp up its investment in keeping newbie programmers on the professional development path they could consider something like a Mac "Coding Machine" that consisted of a what is basically a scaled down and spec'd down Mac Mini, perhaps in a poly-carbonate case, that is priced say at $299. Perhaps it only has i3 and fewer ports. I'm thinking something that is in the same sphere of focus as the Raspberry Pi but a full Mac OS machine (with XCode) that is more akin to what professional programmers are using compared to an iPad with Swift Playgrounds, which I as more like Lego Mindstorm for early exposure to programming. Apple needs to create an easy and natural path from Playgrounds to XCode, so having an educational focus coupled with an entry level product relieves Apple from having to have the "premium" wick turned up quite so high as what we see with the new Mini. Apple can do it, but needs to redefine the rationale that's driving their engineering and product development machine to make it happen. Education, STEM, robotics, and "everyone can code" may be exactly the rationale they need.

    Of course an entry level Mini won't please the current crowd here at AI, but Apple can't please everyone all of the time and they still have a business to run. We CAN pay for premium, even though we like to complain about it from our thousand dollar phones.

    The "easy" path to an entry level Mac at a lower price point is to dump Intel. Intel's gross margins are over 60%; TSMC's gross margins are just under 50%. Every time Apple ships a product with a processor fabbed by Intel rather than TSMC they are paying an "Intel tax" of sorts. An A12-based Mini would be a very strong entry level machine at a lower price point. An A12X could be offered as a BTO option for higher performance. 
    I could not agree less that the "easy" path for a high performance low-cost Mini involves a transition to an A-series (ARM) architecture, unless of course you want the Mini to be a headless version of the iPad Pro. There's a high level of Benchmark Fallacy going on right now around Apple's A-series chips. They are amazing components in their own right but don't fall into the trap of believing that just because an A12X can run benchmarks that run circles around the majority of consumer grade Intel based systems that they are a drop-in replacement for Intel x86 chips. Not so.

    Where these benchmark fueled fantasies and marketing "stretches" blow up miserably is in the domain of Software Reality. Moving a very large number of beefy line-of-business applications, not to mention core OS and machine bound services, and third part apps from x86 to ARM is very non-trivial, ridiculously expensive, and very time consuming - like they would have already had to have started 5 years ago to have a shot at hitting dates before the mid-2020s. Perhaps Apple actually started the train rolling 5 years ago, but we don't know. What about partners in the ecosystem? Do you want to rely on warmed-up iOS versions of critical business apps until the ecosystem finally catches up?

    Software developments costs can and often do dwarf hardware development costs for systems. There is no software TSMC fab-for-hire equivalent for Apple to lean on and throw bucketfuls of cash at to expedite software development progress and unburden its in-house development resources from having to do everything themselves. While there have been gains in software development maturity over the past decade, software development in general is still one of the easiest ways to annihilate vast quantities of time and money and have little to show for it. Not to beat a dead horse, but where there are significant software development and integration risks involved things like gross margins between chip/chipset suppliers and Intel taxes are secondary considerations that only come into play once the costliest contributor to overall system costs are brought under control and you have a system that can actually be built before you burn through all of your R&D resources.

    Benchmarks are meaningless until you solve the software problem.


    Utter BS!   Much of the software that runs on Mac OS has already been ported to ARM.  The software that hasn’t been can be quickly cross compiled if developers followed Apples guidelines.   Seriously consider the body of software on the IPad for one.  Second consider the huge volume of Linux or open source software running on ARM right now, often the  very same software running on the Mac. 

    This software issue has been blown way out of proportion considering the vast majority of Mac users will run Apple supplied software and a couple of third party apps. 
    blastdoornetmage
  • Reply 99 of 151
    wizard69wizard69 Posts: 12,627member
    dewme said:
    blastdoor said:
    dewme said:
    blastdoor said:
    dewme said:
    This is a worthy upgrade that predictably builds on Apple's reputation of delivering premium products at a premium price. It is now safe to say the Mini is  not going to be Apple's budget or entry level product that some folks were hoping for. It's one thing for us to say "the ecosystem could use a less expensive one, at a $499 or $599 price-point" but quite another to identify what we're willing to give up on the lowest end configuration to hit that price point. By the way, Apple accepting a lower margin is not a viable answer and is not going to happen. This is still a premium engineered, amazingly industrialized, and thoughtfully sourced product that Apple put a tremendous amount of resources into to preserve what Apple values in a product it is proud to present to customers. They are not simply slapping a bunch of off-the-shelf components into an ugly white box that's stuffed with whatever they can get cheapest this week.

    I'm still hopeful that Apple can find a way to justify delivering a true entry level machine. To do this they will need to approach the problem from a different angle. For example, if Apple decided to significantly ramp up its investment in keeping newbie programmers on the professional development path they could consider something like a Mac "Coding Machine" that consisted of a what is basically a scaled down and spec'd down Mac Mini, perhaps in a poly-carbonate case, that is priced say at $299. Perhaps it only has i3 and fewer ports. I'm thinking something that is in the same sphere of focus as the Raspberry Pi but a full Mac OS machine (with XCode) that is more akin to what professional programmers are using compared to an iPad with Swift Playgrounds, which I as more like Lego Mindstorm for early exposure to programming. Apple needs to create an easy and natural path from Playgrounds to XCode, so having an educational focus coupled with an entry level product relieves Apple from having to have the "premium" wick turned up quite so high as what we see with the new Mini. Apple can do it, but needs to redefine the rationale that's driving their engineering and product development machine to make it happen. Education, STEM, robotics, and "everyone can code" may be exactly the rationale they need.

    Of course an entry level Mini won't please the current crowd here at AI, but Apple can't please everyone all of the time and they still have a business to run. We CAN pay for premium, even though we like to complain about it from our thousand dollar phones.

    The "easy" path to an entry level Mac at a lower price point is to dump Intel. Intel's gross margins are over 60%; TSMC's gross margins are just under 50%. Every time Apple ships a product with a processor fabbed by Intel rather than TSMC they are paying an "Intel tax" of sorts. An A12-based Mini would be a very strong entry level machine at a lower price point. An A12X could be offered as a BTO option for higher performance. 
    I could not agree less that the "easy" path for a high performance low-cost Mini involves a transition to an A-series (ARM) architecture, unless of course you want the Mini to be a headless version of the iPad Pro. There's a high level of Benchmark Fallacy going on right now around Apple's A-series chips. They are amazing components in their own right but don't fall into the trap of believing that just because an A12X can run benchmarks that run circles around the majority of consumer grade Intel based systems that they are a drop-in replacement for Intel x86 chips. Not so.

    Where these benchmark fueled fantasies and marketing "stretches" blow up miserably is in the domain of Software Reality. Moving a very large number of beefy line-of-business applications, not to mention core OS and machine bound services, and third part apps from x86 to ARM is very non-trivial, ridiculously expensive, and very time consuming - like they would have already had to have started 5 years ago to have a shot at hitting dates before the mid-2020s. Perhaps Apple actually started the train rolling 5 years ago, but we don't know. What about partners in the ecosystem? Do you want to rely on warmed-up iOS versions of critical business apps until the ecosystem finally catches up?

    Software developments costs can and often do dwarf hardware development costs for systems. There is no software TSMC fab-for-hire equivalent for Apple to lean on and throw bucketfuls of cash at to expedite software development progress and unburden its in-house development resources from having to do everything themselves. While there have been gains in software development maturity over the past decade, software development in general is still one of the easiest ways to annihilate vast quantities of time and money and have little to show for it. Not to beat a dead horse, but where there are significant software development and integration risks involved things like gross margins between chip/chipset suppliers and Intel taxes are secondary considerations that only come into play once the costliest contributor to overall system costs are brought under control and you have a system that can actually be built before you burn through all of your R&D resources.

    Benchmarks are meaningless until you solve the software problem.


    Apple was a tiny fraction of its current self when it made the transition from PPC to Intel, and tinier still when it made the transition from 68k to PPC. And yet they did it, quite successfully. 

    Today, changing processors would be far easier than in the past. The code base is so much more modular and Apple has vastly more resources. Everything in tech is "hard" in the sense that it costs money, takes time, and requires qualified people. But a transition to ARM is not the herculean task you make it out to be. 
    Are saying you are willing to fund this transition with your money?
    I suspect it is more about saying some here are out of touch with reality.   Apple has spent over a decade pushing developers to stick with APIs and avoid hardware specific tricks.  Thus the code based for Mac hardware are much cleaner than other platforms.   Combine this with infrastructure Apple has been building into its app stores it becomes clear that they are making a possible transition toward new processors as easy as possible.  

    This isnt to imply a perfect world or that all developers are well behaved.  Rather it says that the vast majority of Mac software would be quickly running on ARM.  Frankly they wouldn’t even need emulation and simply leave the poor developers to suffer.  If nothing else A12 & A12X are good examples of why developers need to pay attention.  
    blastdoornetmage
  • Reply 100 of 151
    kharvel said:
    That particular config, the box is included, and it is not cheap. The drive as pictured is $1300, with the enclosure if you want it separately for $199.
    Twice the price of basic Mac Mini, (almost, @ $1300). Apple’s 1 TB internal SSD upgrade is $800...
    Twice the price NOW.  6 months down the road, it will probably drop to $800.  1 year down the road, probably gonna cost you $400.  2 years?  Do the math.  Meanwhile, the Apple internal NVME remains. . . .wait for it .. . . THE SAME!
    So wait for 6 months and buy Intel Optane @ $800, what’s your point? This is Intel, not Samsung, you cannot buy it @ $800 even in two years...
Sign In or Register to comment.