I am an owner of the "latest" Mac Pro, and I had the liquid cooled PowerMac G5 before that. I don't really know what to feel when it comes to the Mac Pro anymore. I used to compile WebKit a lot, and it takes forever (change one .h file and go get some lunch while you wait). When I got my Mac Pro 2 years ago it compiled it in under a minute I think. It was incredible. The best part was, back then, Xcode had the distributed builds feature, so I could go anywhere in my apartment with my Macbook Air and compiling was still crazy fast since the work was offloaded to the Mac Pro's 12 cores on my home network. Then they removed that feature from Xcode. I also had a two monitor workflow, but then they broke (for me) dual display support with Lion. The Mac Pro also doesn't support the latest and greatest graphics cards, so its not worth it in my opinion to upgrade them for Cinema 4D or games.
I don't even know if I mind really. The reality is that for the last 3 or 4 years, it feels like Apple has more and more been pushing me to not really love the computer experience anymore. When I think about my next "big" purchase, I will probably build a PC. I have never done that before in my life. I've exclusviely owned Mac. But "leaving" Mac OS X doesn't seem like a big deal to me anymore. I'll just be trading one set of frustrations (Messages, Notification Center, 5 minute long sleep wake up times, endless WiFi issues, etc) for another. At least I'll actually have fast graphics.
And Xcode will suck on my Macbook Air the same way it sucks on my Mac Pro today, so no real loss there, since I don't compile WebKit anymore. Again, its mainly just indifference now. I don't know if I miss the days when this stuff was fun, but Apple has certainly finally succeeded in making me see their computers as just tools.
>When I think about my next "big" purchase, I will probably build a PC.
I did precisely that. Instead of upgrading my MBP, I built an i5, 16GB RAM, and a gtx650ti for well under $700. I've had some teething issues with regard to my workflow but VirtualBox has been a life saver.
Years ago (but not as long as tolmasky's "exclusive"), I would have given everything to get a Mac (I was living in a place where Apple's product were neither available nor affordable).
I spend most of my time on my computer programming (or at least I did this until very recently).
What I've noticed in the last 2 years is that except maybe the Xcode environment, Apple has started to ignore every other dev environments. People have been working a lot on keeping things going (Fink, MacPort, homebrew), but by the day my feeling is that Apple doesn't care at all about this.
Basically the target audience shifted from geeks to mass audience (click to buy this app, watch a movie on this great display, etc.) and AppStore developers. Business-wise it's clear they can ignore the small segment of developers. Karma-wise I'm not convinced that's wise.
The only way I could agree with you more is if you didn't think they were catering to App Store developers. Xcode is terrible, the provisioning profiles are irritating, their requirements can be inscrutable. It seems like a bad experience compared to what few hoops Android developers have to jump through.
But I digress. With Lion and Mountain Lion is it becoming clear that they don't want any people that know how to use computers in their target demographic, they just want iPad users to have a unified experience.
I used to urge my friends and family to get OS X computers, but I have reversed that opinion and now urge them to get anything else.
My comments about Xcode were uninformed, based only on the fact that I don't run into tons of complains about it. The only time I fired Xcode to compile a project, it took me quite a bit of time to figure out where it places the compiled files. But I categorized that as "I'm a newbie. I don't know where to look"
I think it's definitely not you that is the problem :) Even the latest releases love to hog all the memory (even when idle) and mine creates zombie processes every time I run the simulator. It's the single crashiest piece of software my computer runs (multiple times daily).
Sometimes it cannot even copy files right; drag and drop, oops it failed, drag and drop again, oops it already copied the files to the folder but didn't add them to the project, now it will neither copy nor add them, you have to go delete them yourself.
I am in the same. I used to build all my PCs, then sonys, then dells, then the best *nix desktop experience and intel came along in Macs and my Mac Pros (iMacs are nice but I need more). Mac Pros and the power with that was a winner, it could continue... I think it is a bad move not to continue it. Where else can you run OSX on a beast? well a Hackintosh but that sucks. Please Apple, don't abandon the high-end, if you do let me run OSX on other hardware. I recall a time in the past when Apple got to this point (http://en.wikipedia.org/wiki/Macintosh_clone), Jobs ended it when he got back. But they turned to it because they didn't deliver to the high end well enough.
"[Apple] turned to [licensing Apple clones] because they didn't deliver to the high end well enough."
No, the inverse was true. Apple allowed licensing because they thought clone manufacturers would be satisfied building low profit margin entry level Mac clones. Apple wanted to stick to products with higher profit margins.
What ended up happening was that companies like DayStar and PowerComputing started building way faster, more extensible tower workstations than Apple's PowerMacs. Power users loved it and Apple's high end sales suffered for it.
On the low end workstation side, Apple's Power Macintosh 4400 was actually quite price competitive with the clone offerings of the day.
The 4400 was one of the first CHRP-type boxes. They used a Motorola made motherboard and usually had board headers for parallel and PS/2. The same board (or at least a variant) was used in Motorola StarMax machines, which were quite powerful for their price, except for the 40mhz system bus. I had a "PowerCity" which was an illegal clone made from leftover Motorola parts after the Jobs decision. It was a great machine, and it was nice to have a PS/2 port on mac at the time!
I remember it well. There was a lot of talk about CHRP, but in the end, Apple released no Macs that were fully based on it. The 4400 had the same Apple designed Tanzania motherboards [1] as Motorola used in some of their StarMax machines [2]. Those boards had support for commodity PC ports like PS/2 and parallel, as well as Apple's ports. The Power Macintosh 4400 did not have parallel or PS/2 connectors. It had ADB and GeoPort serial ports [3].
I considered getting a 4400, but I ended up spending a little more on a 7300, which was a way better computer.
It's sad all of the negatives you brought up started with the Lions. Is the pendulum swinging back? I'm even writing this comment on a MS Surface Pro which I'm completely loving.
Somewhat cynically, I think that Apple wants to be out of the tower business because you can't obsolete a tower every 9 months and get people to buy a new one. Towers have user-upgradable parts, can be serviced by laymen (and thus don't have the upsell opportunity when you have to bring it in to fix an issue), and are generally resilient against not being the latest-and-greatest.
In a Macbook Pro, by comparison, you are charged $200 for an extra 8GB of RAM and $500 for an extra 384 GB of SSD, not because the components or labor are worth that, but because you'll pay whatever they want you to, because you can't upgrade it later.
For better or worse, Apple is in the appliance business now. Buy your appliance, use it for a few months, then Craigslist it and buy the new one when it comes out. It's obscenely profitable and they have their consumer base trained to do it on command. Giving people something that they could keep current for a slim fraction of the cost of a new device is not in their interests.
All the things you mention that make towers great for advanced users also make them a hassle for normal users. When someone's sealed box has a defect, if you know what you put in the box, your troubleshooting and repair scenarios are vastly reduced compared to letting the consumer into the box. To a normal consumer, knowing that they could have upgraded their tower doesn't just give them more upgrade anxiety, it also convinces them that they're being ripped off while the smartass kid down the street is not. The sealed box is the ultimate equalizer.
I think all your points hold, I'm just saying there are a few somewhat less cynical explanations that may also apply.
I think you're absolutely right, but I think that's just another point in the "Apple as an appliance vendor" category. Sealed boxes are easier to service, easier to replace, and easier to get past consumer hesitation with. I think it's a rock solid business strategy (and Apple's financials would seem to agree), but I think it makes for an absolutely abysmal computer company.
I don't think that Apple wants to be a computer company anymore. A tower wouldn't fit into who Apple is today.
Agreed, and I think that's the point. Apple has shied away from being a computer company for many years now. To wit: it officially changed its name from "Apple Computer, Inc." to "Apple, Inc." in 2007, almost concurrently with the release of the iPhone.
I think this has less to do with a cynical desire not to have to service computers, and more to do with a forward-looking belief that the future of consumer devices has less and less to do with computers (at least as we currently use the term).
I mean, it's not totally inconceivable that Apple will have stopped making any computers, per se, within the next 10 years -- or, at least, that the consumer-facing distinction between "computer" and "device" will have blurred to the point of obsolescence.
I agree with the premise, but I think a larger, more capable iPad would actually be a more influential 'halo car', given today's Apple clientele.
An iPad with a 15" screen could display twice as much information and would allow for more on-screen controls. Apple could port the full versions of Pages, Numbers, and Keynote for iOS, but also the Pro Apps like Final Cut, Color, Aperture, and Logic. An iOS version of iBook Author could be a killer app. I've also long dreamed of a WYSIWYG authoring suite for making iOS apps on iOS (along the lines of Automator or Hypercard) – not for submission in the App Store, but to run on your own iOS devices and to share with friends and family.
Nice to have: add two Thunderbolt ports to the iPad Pro. Not only for connecting fast SSDs, but also so that you could link another iPad Pro. This would add a second screen, and for processor intensive tasks, the two CPUs could work together (think xGrid or GCD).
You know... I want a large format iOS device... but as a pro user the very last thing I want is a sandboxed jailhouse. Apple is abandoning pro users across the board. iOS apps are essentially toys because of the limitations of iOS and the app store itself. The promise and the potential is there for something truly incredible, and that's why it's so hard to watch.
For (some) developers, I can understand they need full access to the OS, file system, and services. However, many professionals in other lines of work don't need such an open, vulnerable environment. For them it's more important to have a safe, stable system that's easy to use and doesn't require a lot of configuration. That's why I believe the future is in iOS, not OS X.
Also, for web development, I find that the iPad works very well (for me). I have all I need with Diet Coda, OmniGraffle, OmniOutliner, DesignPad, Idea Bucket, Writing Kit, and Photoshop Touch. I just wish for a larger screen.
Task switching in iOS is pretty horrible, as with most mobile OSes. If you're switching behind that many apps you must have a high tolerance for pain.
Also I disagree with you that iOS is good for professionals. The way it attempts to hide the idea of files makes it very inconvenient to work with (coupled with the lack of a compelling alternative). I can't think of any professions where files aren't important.
"The way [iOS] attempts to hide the idea of files makes it very inconvenient to work with"
It doesn't hide documents, iOS has a different way of presenting and organizing them than you're used to. Instead of putting them all on one pile for you to sort out, iOS takes an app centric approach. You never have to search for that spreadsheet you made last week, you know it's in the file drawer in the Numbers app (and with iCloud, the spreadsheets you made on your Mac will also be there).
Of course the standard way iOS handles files doesn't work for everyone. If you have thousands of files you need to sift through, you might prefer using Dropbox or SFTP with an app like GoodReader, or use a web app like Google Docs.
The app-centric approach breaks down when you have to share that data out to colleagues who collaborate on it or gasp use it on a non-Apple device. These things can be fixed but as it stands you really have to go out of your way and adapt your own workflow because of the limitations of the platform.
I found that task switching in webOS was REALLY nice.. and am glad that Android has since adopted a similar switcher. Now the lack of a physical menu button on most android devices is a different issue.
Task switching in Windows 8 is also really nice (very similar to Web OS). Swipe in from the side to show the task switcher, then click on what you want. Snap a second app to the side of the screen if you want. Very nice balance between keeping the interface simple and offering some flexibility.
I use a lot of those same apps, and I agree that they work. But they work the same way on a laptop. The unique properties of the tablet format can't really be taken advantage of with those apps. I do have an issue with the way files are handled, though, but I think it's minor and will be addressed over time. For instance the "horizontal scrolling list of gigantic previews of documents" just doesn't scale to anything a professional can deal with. And the document syncing is all over the map - Omni told me they won't/can't support Dropbox, etc.
But again, I feel that is a relatively minor point compared to the bigger issue of using the tablet format to its potential. Apple is limiting it in the most severe and terrible ways in my view. This new format DEMANDS (caps!) experimentation, demands out of the box thinking, demands breaking stuff left and right... but the app store and DRM and content limitations are putting harsh brakes on it.
I agree and that's what is so terrible. Paper isn't pro or non-pro. You can go but expensive brushes and paint, or buy cheap stuff, or whatever you want. There are no limitations in the physical world. But in the digital world, you have severe, totalitarian definitions of pro and amateur. You are not free to paint on your typed document. Ever. Unless the developers foresaw and expressly allowed for that event.
It does seem there is a bit of willingness in Apple to build a higher-end iPad (given the 128GB model). I would expect something along the lines of a larger model designed for artists and situations where a big surface is required. A 15" screen gets you 9 x 12 in screen which would be good and still portable. I would bet it has the same connections as every other iPad.
I do wonder if when an AppleTV SDK appears, does that mean we will get an iOS version of the iMac? At that point the Mac becomes a prosumer / pro OS.
bad analogy to use a car analogy a high end iPad would just be a tuned scooter as opposed to a full on 4wd super car say the Audi R8 with the factory approved performance pack.
> "Apple should keep pushing the limits of PC performance because it’s a company that loves personal computers."
I love reading Siracusa, and want to agree with him here, but this sentence stops me.
Apple clearly doesn't 'love' personal computers anymore. How much clearer could the signs be?
They took their engineers off Mac OS X and put them on iPhone OS. Their PC OS lagged and stagnated. The 'new' Mac Pro isn't the only insult to power users; OS X 10.8 itself is, too. It is the worst update in (Mac) OS X history, not only with serious show-stopping bugs, but also with new 'features' that are ground-breaking in how dumbed-down they are[1].
Their new love is closed, power-user-hostile, quasi-disposable consumer devices. Steve Jobs himself broke up with the personal computer in his last years, and the company followed suit.
[1]: LaunchPad, newly-crippled Expose, disabling extra monitors in fullscreen mode, double clicking hidden files doesn't open them, moron open-save dialogs, by-default iCloud file storage that prevents one app from being able to open another app's docs (even to attach as mail attachment), etc etc et
Perhaps you'd find the change of company name in 2007 more convincing?
1977-2007 Apple computer
2007-2013 Apple inc
I very much doubt they did that just on a whim. As a user/developer of both platforms (iOS and Mac OS X) it's been clear for some time that the company has refocused on consumer gadgets - it's 'the next great thing', to quote Steve Jobs. That change might even accelerate under Tim Cook because that's where the money is.
Re the Mac Pro, I'd love to see them sell a stackable tower of modular units like the mini in form factor, offering as much hot pluggable storage and CPU power as you want. Then they could widen the market and sell it to consumers who think of themselves as pros too, and actually make money from it. At present there just isn't enough money in mac pros for Apple inc to bother updating them, and like the server line they have languished.
What's the last major OSX feature that wasn't bad? For a company known for its successes, you look at Launchpad, inverse scrolling, notification center... does anyone like those features? I've never read a single positive comment about any of them. The most positive thing I've read is that they are easy enough to ignore.
To be fair, there are a lot. FileVault 2 is a massive improvement over the (hackass) first version. In 10.7, Mail got dramatically better at dealing with huge volumes (hundreds of GB, dozens of accounts) without crashing. Buttons finally became roughly rectangular. The OS supports Retina displays, one of the most significant general advances in recent history. Menlo beats Monaco. Resizing windows from any edge was only 20 years too late.
Most importantly to me, (since for all my bitching, I spend a lot of my work time writing Mac software) Xcode went from 'spectacularly awful bag of suck' to just 'kinda sucks', and ARC, blocks, Objc-2, non-insane literal syntax are all forces for Good.
And yes, I love inverse ('natural') scrolling and it is enabled on all my Macs. :)
It's not that Apple does nothing at all good, it's that their focus has become even more relentlessly "systems that the average shopper at walmart can use" at the expense of those users who have a stronger grasp of general computing concepts.
When they disable all the extra screens in fullscreen mode, the chances of my dad getting confused and looking at the wrong monitor are reduced. But at the same time, the amount of data I can see at once, and my ability to manipulate it, are also reduced.
Well I love share sheets, iCloud syncing, and AirPlay, personally. Gatekeeper is a great compromise between a fully locked down system and a fully vulnerable one. And the new Safari is the best Safari ever, as the saying goes. So I count 4.5 on my end.
Anyway, an OS is more than the sum of its newest features. On the whole, OS X 10.8 is a stable, polished, powerful, fully-featured operating system. There might not have been a big feature like Time Machine or Exposé in a couple versions, but I'd much rather use 10.8 than 10.5.
Gatekeeper is an okay compromise. I'd feel better about it if they didn't include restrictive app policies. What if I want to run apps that are signed by a certificate from some third party CA? I'd feel great about it if it would also inform me what permissions are being granted and let me disable arbitrary permissions like internet access.
I think 10.7 and 10.8 bring more irritations than anything else. iCloud sync and Gatekeeper are the only things I think are worth upgrading for, but the Gatekeeper implementation is still lacking and unfree, and I use other services to sync my files (I also think the implementation of iCloud leaves something to be desired... and it's another form of lock-in). I'm still using 10.6 and shall use it for the foreseeable future; at least until my next computer, which will most likely be running Linux.
Enough about Launchpad. It's perfectly ignorable, and no sillier an interface than FrontRow, which shipped on most of the supposedly better earlier versions of OS X.
Remember when every Mac shipped with an IR port and a little remote? Laptops shipped with remotes. But apparently an optional app chooser interface (LaunchPad) is the most anti-PC they've ever shipped.
If front row still existed, I wouldn't be throwing away my 08 blackbook. It's not really powerful enough to be a server, but it would make an awesome HTPC.
I thought I would use Front Row a lot when I hooked my Mac Mini up to my TV, but it turns out that most of the video we watch is through a browser--on YouTube, Vimeo, or Hulu. And I control my iTunes library through the Remote app on my iPhone. So I don't miss Front Row.
I like inverse scrolling. Old scrolling feels weird to me now. And anyway, why is a minor, optional tweak to scrolling direction anything more than a minor, optional tweak to scrolling direction? It seems like every move Apple makes, no matter how trivial, takes on apocalyptic significance to people.
Notification Center is a list of notifications you can turn off if you don't like notifications. What's the big deal?
Launchpad is a little icon you can click to see your apps. What's the big deal?
That is clearly false. GP and others have said that they like inverse scrolling and normal scrolling feels weird now, which is nicer than "at least you can turn it off". QED. Geez.
The sentence in question is not an assertion about Apple's feelings towards PCs. It's recommended pairing of action and motivation. "Joe should marry Sue because he loves her, not because she's rich." Whether or not Joe loves Sue is another matter.
The point remains that if they however don't love personal computers, then both reasons stand. In other words, perhaps Joe should marry Sue because he loves her, but its moot because he doesn't (he instead loves her much younger sister). EDIT: In other words, you're both right.
If it cost 5¢ to dismiss an App Store Update notification, I'd be in more debt than the US government. Notification Center is the single most annoying (and unstoppable) OS feature I can remember.
To be pedantic, 7.2 or 8.0 were pretty damn bad. 7.3.5-7.5.5 were the best. I think there were some dogs around system 5, but I'd have to get my 128k out of the dump to be sure.
There was no System 7.2 or 7.3.5. They went from 7.1.2 to 7.5. You may be thinking of 7.1.2 (the first PowerPC release) or 7.5.3 (the last full version posted as a free download).
It wasn't really until 7.6.1 they worked out all the teething issues with PowerPC Macs (mainly, the whole machine halting with "error type 11" requiring a reboot)
The metaphor would work but for the fact that smaller is a harder technical challenge than bigger. They're not skirting some kind of spiritual calling to make better technology by focusing on portable instead of workstations--if anything, they are doing exactly what the author wants. To equate pushing the envelope with losing money is very superficial reasoning.
As I noted in the article, there are many more axes of innovation beyond just performance. But performance is definitely one of them!
Apple's pursuing top-end performance with gusto in its laptops and mobile devices. But its desktops have been left to languish, not even using the fastest available off-the-shelf CPUs and GPUs (let alone using custom high-performance silicon designed in-house, a la the latest iOS devices).
I'm saying there is no such relationship between halo cars and the Mac Pro. They're "leaving them to languish" while inventing amazing technology on the side of the spectrum. They've managed to make a $330 product seem aspirational.
The more I think about your argument the less sense it makes. Great article, well written and full of interesting facts, but I don't think your point holds much water.
Exactly. The real technological breakthrough is still 2048 × 1536 pixels in the iPad, together with the processing power needed to drive it small very mobile form and ten hours of work on battery. Compare that with the whole PC industry. The smallest monitor I can buy in my rich european country, which has nearly so much pixels on the shorter side is 27 inches in diagonal, having only 1440 pixels and just 110 dpi. Or with the Microsoft's Surface Pro of only 5 hours and "only" (a lot compared to the 1366x768 Surface RT, only compared to the iPad) 1920 x 1080.
I wonder about Tim Cook's personal history with computing; he's brilliant at what he does, but he doesn't strike me as a long-time computer enthusiast. Nor does Ive; his enthusiasm is wrapped up in crafting beautiful objects.
Obviously there are many other influential execs and engineers at Apple, but I don't know of any with the pure love of Woz or Akio Toyoda, who'd be able and willing to burn a billion dollars to push the limits of desktop computing for its own sake.
Simply put, I think they've already checked out of the computer industry. They're now a consumer electronics company that happens to make use of computer-grade hardware and software engineering.
Would Apple really be innovating by coming out with some smoking fast workstation? We know how to build these things: cram them full of top-of-the-line Xeon CPUs and tons of SSDs and GPUs.
The real problem is Apple's pro software is rotten, and throwing more hardware at it barely helps.
Shake was acquired and then killed.
FCP is no longer a pro application. They've added lots of nice first time user experience things, but largely ignored what "professional" users need. Which is fine. Apple wants to sell software (and iphones and ipads) to Moms and Dads and college kids.
FCP may not be a pro application for editors who use tape, but it is most definitely a pro app for editors who have gone fully digital. I doubt Apple spent millions of dollars and several years to rebuild FCP from the ground up in order to sell iPhones to Moms and Dads.
The analogy to Corvettes and Vipers is massively flawed. Automotive companies use sports cars to market their brand directly via racing and indirectly when the owners of production cars drive them down the street.
On the other hand, the days when speed was a major selling point of computers are long passed, and online, nobody cares whether I'm surfing HN with 8 Xeon cores of my desktop or one ARM core on my phone.
The only people who care about the MacPro are people who already want to buy Apple products. And an Apple logo which sits under your desk doesn't promote the brand in coffee shops and airports.
For me, the thing I lusted before switching completely to Mac's was indeed the PowerMac G5. I saw it at an agency back in the day, and I couldn't get that massive tower out of my mind.
I eventually bought a MacBook Pro, but soon after a Mac Pro followed suit. I don't care if it's 8, 12 or 16 Xeon cores, but I do care about speed. The way I can use all the applications and virtualizations I need at the same time is amazing. I initially thought of it as overkill, but I'm never going back. Unless Apple forces me.
Plus, that tower is just gorgeous. I love looking at it. I even have an empty G5 case that I cannot bear to part with. That's what a "halo car" can do to you, I guess.
Apple can compile their toolchain and OS to run on any hardware they want. And when it comes to OSX and iOS, over the long term, they would probably prefer not to have to support third party development. The analogy for iOS development is sharecropping, and it doesn't exist anymore because it's more efficient for the landowner to buy a tractor.
In practice, they don't give their own software developers non-production machines except in special circumstances. Most OSX & iOS developers in apple are using apple production hardware.
A filmmaker might mention Apple in the credits for a few seconds as people file out of the theater...but would need permission to use the logo. A musician might mention Apple in the liner notes...but iTunes doesn't deliver jewel cases. And a designer...well "Made on a Mac" is from the Geocities age.
Racing sells cars because I can buy a Corvette and hoon it on the road in front of my house. But a MacPro is still going to give me the same Facebook friends and won't improve my porn.
I have (even recently) managed codebases that took hours to compile on decent commodity hardware. I just put a requisition in for a more-core box, though I might retract it and wait to see what this Mac is.
Why? The only core components you can expect Apple to innovate on are the display and the form factor. They're not going to be shipping any CPUs that aren't already on the market.
Trying to reconcile this post with the ghosts of past Unix workstation vendors, who tried to push the boundaries of what you could do with microprocessor-based computers, where spending tens of thousands of dollars per box was common.
Fast-forward to today, all the knowledge gained in building those high-end workstations have been distilled in the smartphones we have in our pockets now.
So for Apple, R&D for its own sake or for controlling the narrative?
Most people aren't willing to spend $1700+ for a laptop, with that case and that screen. They'll pick up one of the Airs or the old style MBPs. The Retina MBP is nice to look at when browsing at the store, and subconsciously, they know they'll have it in 3 years when the technology trickles down the product line.
To Apple, it's not about raw performance. It's about the total package. Which is unfortunate, because it means they're willing to make concessions in performance.
I run a HP Z620 Workstation with a bunch of processing power as my daily.
This is HP's 'Mac Pro'! I'm guessing it serves to keep the HP boffins doing cool things. It has 16 GB memory and a XEON on a daughter card (and the same again on the main board), which can be unclicked and slid out. All custom, and all different from the sister high-end box, the Z820.
This is serious bad-assery, and more power to HP for doing it.
I sincerely hope that Apple build a great new Mac Pro, and keep the flame of high-performance, no-compromise hardware alive.
I say go the other way: Sell a version of OS X without video drivers. (ok, maybe a text console so it can cry piteously when it is broken and you can create an account or something if you have to.)
Remote desktop to manage; if you have a supported GPU it is 100% free for compute; only support chipsets and processors that Apple supports anyway. The engineering cost is minimal. Compatible hardware is the customer's problem. People that need horsepower for their render farms or whatever can build or buy the systems that make sense the month they need to buy. People that need to stick code in a hosting service can do that. Stop making me do my engineering twice; once for the native Mac OS, and once for the servers. Stop losing customers when the Pro line gets too long in the tooth. Give it a "no brainer" price. You aren't doing much engineering and it isn't canibalizing your human interaction machines, it protects them from migration.
I'm not sure about the engineering cost being minimal. My research group builds a compiler for a parallel programming language, and we have trouble getting reliable performance on OSX. Many of the things you take for granted on Linux are not there:
- Pinning the kernel to one core
- Reliable pinning of threads (instead of just suggesting that groups be co-located on the same cache bank)
Honestly, we have a fairly long list to even get to basic performance, and that's before you start hammering network stack reliability (as my distributed systems friends complain), etc. OSX on the client with Linux on the server works pretty nicely, and I'm somewhat concerned that the OSX team would have to drop client work for a whole release to focus on hardening the server experience, as well as figuring how what they need to implement to be able to support the wide variety of Linux libraries that don't quite work due to holes in the OS that performance people rely on (numactl; papi; etc.).
That said, I thought most of the Apple developers have Mac Pros under their desks. If for no other reason, to maintain a highest-end client workstation, it makes sense to keep things alive. But the server sell seems a bit of a reach, engineering-wise.
I don't think you'll pry linux developers away with this. But for people adding distributed backend workers to existing OS X apps it makes great sense, even if you don't end up at the same speed as a tweaked out Linux implementation.
Likewise, you probably won't beat a good Linux network stack for jamming bits through, but lots of people just need to be up and running, and if the code is shared with the client, that's great.
I think its a case of "showing up", they can worry about being best later, if it makes sense, but showing up at all is important. Apple just lost a premier video production house run by a friend of mine over the Mac Pro lifecycle. They'd have stocked up on render machines and waited out the front ends if they could, but that wasn't an option.
(Apple will beat my last round of 100% Intel motherboards in Linux servers which have a years old, unresolved flaw in one of their two ethernet controllers that make them reset whenever you push them too hard. They also have a good chance of beating Linux in GPU computation because the vendors actually care a bit about Macs. I've got a lot of GPU power laying around that I can't use at all in Linux.)
It is common knowledge that Apple mainly uses PC hardware in their data centers. You can get an idea of the third party products they use by searching the jobs postings at jobs.apple.com. I'd bet they would be open to using non-Apple hardware in other areas if they thought there was a good business case.
This blog post comments that "Consider Larrabee, Intel’s project to create a massively multi-core x86-based GPU. Rumor has it that Apple was working on integrating the technology into a Mac Pro. Intel eventually scuttled the project, but consider what would have happened if it had taken off, reshaping the GPU market in the process." and while Intel has not released a consumer oriented graphics card based on Larrabee, they have released a card called the Xeon Phi based on this tech. For the HPC world the card has the potential to compete well with nvidia and ati's gpu compute offerings. It's a specialized chunk of hardware for sure, but it is not a failure.
As for an updated Mac Pro, all I can do is chuckle. Apple cares about selling high profit margin consumer electronics items to millions of consumers. Apple has long long ago dropped any concern for people who were "professional mac users" (see shake, final cut pro, etc. etc. etc.)
We know that there hasn't been a significant update to the line in ages. We also know Cook has publicly stated the line is not dead and that there will be something for pros this year. That's where facts end and speculation begins. For me, the big question is "what are they waiting for?"
I don't think it is a CPU; they could easily ship something with faster intel CPUs, and I doubt (understatement) they would release a pro line with Apple-designed ARM CPUs (that would certainly explain the wait, though).
So, what is it they are waiting for? Faster RAM? Unlikely. A terabyte SSD? I doubt it. Some retina 3D monitor? Unlikely; that could only cause delays if it were built into the hardware and I would not expect pro stuff to have a built-in display. Some interconnect (optical thunderbolt? Something wireless?) that will allow them to come up with a new form factor of a pro box? That, I think, is the most likely. But hey, let them surprise us.
The Halo Car is an interesting parallel, but one that I believe to be fundamentally flawed.
Here's the problem: Apple isn't building a Corvette, or a Viper, or even an LFA. It's building the world's fastest steam locomotive.
Is it impressive? By Jove, yes. Is it fast? Ridiculously so. Is it powerful? Let me count the ways.
Does it do something people want?
...
The problem with the Mac Pro is that it's a solution to a problem almost no one has. It uses server hardware to solve consumer problems. Great, but ... why? Is there any consumer app out there that really takes advantage of 12 cores? Seriously, what applications have been parallelized to that degree? Anything that benefits from 12 cores will benefit more from CUDA and massive GPU parallelization.
I like the idea of the Mac Pro. I like going balls out just for the hell of it. I just don't think it's a computer that anyone can really use.
The problem with the Porsche 918 is that it's a solution to a problem almost no one has. It uses racing hardware to solve commuter problems. Great, but…why? Is there any public road out there that really takes advantage of 820 horsepower? Seriously, what routes have been designed for these speeds? Anything that benefits from 820 horsepower will benefit more from lighter weight and a better CD.
I like the idea of the Porsche 918. I like going balls out just for the hell of it. I just don't think it's a car that anyone can really use.
The Porsche 918 is a prestige product, though. The jet set and saudi princes can point to the 918 as a status symbol. The Mac Pro doesn't have that going for it.
All hypercars, from the Huayra, to the La Ferrari (gag), to the Veyron are ridiculously impractical. However, that's a feature, not a bug. Their very impracticality cements their position in the hypercar pantheon.
The Mac Pro just doesn't have that cachet. It doesn't give you bragging rights, except among a tiny subset of the population. Even then, they'd likely be more impressed by triple-SLI'd GPUs in a liquid-cooled tower with a ridiculously overclocked CPU.
Only a tiny subset of the population cares about fast cars. A larger portion cares about expensive cars, of which supercars are only a small subset. Many people are dazzled by wealth. Far fewer know or care anything about performance.
I run a maxed out last-gen iMac. 3.4GHz i7, 16GB RAM, 2GB VRAM. I still run into occasions where things slow down, mostly when I'm running a zillion things at once. It's not one app that needs 12 cores, it's the overall multitasking experience.
You also resolved your own statement when you mentioned massive GPU parallelization. The Mac Pro is the only Mac to use desktop grade graphics (iMacs have to use mobile chipsets due to thermal constraints), and it's the only Mac where you can throw in more than one card.
Unless someone comes out with a Thunderbolt PCI-e adapter for desktop graphics cards to plug into, there's still going to be a need for an expandable tower workstation machine.
I want to see whether Apple delivers on Tim Cook's alleged, cryptic comment. Given some of the very clever things Apple has done technically (GCD, neatly adding closures to C, Thuderbolt,...) I remain hopeful Apple has at least one nice surprise left for the Mac Pro faithful.
But don't assume it will be a new Mac Pro. Why not a new portable that can transparently offload compute-intensive operations to the cloud, and seamlessly utilize external acceleration via Thunderbolt?
I agree with the idea, but Apple is stuck with Intel's Xeon roadmap, which at this point multi-processor is lagging far behind the single processor lines.
Although Intel's Westermere-EX was last updated in late 2011, the specs are still pretty good. Those CPUs have up to 10 physical cores (20 virtual thanks to hyperthreading) and 30MB of L3 cache. The clock speed isn't amazing, but that's to be expected when you have to deal with the heat from 2.5 billion transistors.
The Ivy Bridge version is supposed to ship in Q4. It should have even better performance.
I think we should also look up the exact details of how "stuck" they are with Intel, if this new-Mac-Pro-like-thing ends up being as revolutionary as some hope...I seem to recall way back at the "switch" to x86 that there was some deal with Intel and Apple that only locked them together for a certain number of years? That number of years may be approaching, I don't remember the details.
A Mac Pro with a Xeon Phi would be a genius move--satisfying the core graphics market while not requiring the use of insanely priced general purpose Xeons, which are intended for the datacenter, not the PC. Most buyers don't want to run a database, they want to crunch pixels. Of course, they need Adobe on board. Given how shittily Adobe makes software, we could be waiting a while.
The iPad is Apple's "halo car." It's their RISC machine with in-house designed CPU running their in house designed OS. The Mac Pro is at best a assemblage of off-the-shelf PC parts in a nice case. Want to build the most powerful Mac Pro possible? Just buy the most expensive CPU/motherboard Intel is offering that day and pair it with a ton of RAM. Done before lunchtime.
How about something with a separate screen like a Mac Mini, but in a much bigger vertical profile chassis based on iMac internals, more CPU cores, liquid cooling or heat pipes, with lots of passive cooling using its increased surface area? (Like the area the screen would have taken up?)
I don't even know if I mind really. The reality is that for the last 3 or 4 years, it feels like Apple has more and more been pushing me to not really love the computer experience anymore. When I think about my next "big" purchase, I will probably build a PC. I have never done that before in my life. I've exclusviely owned Mac. But "leaving" Mac OS X doesn't seem like a big deal to me anymore. I'll just be trading one set of frustrations (Messages, Notification Center, 5 minute long sleep wake up times, endless WiFi issues, etc) for another. At least I'll actually have fast graphics.
And Xcode will suck on my Macbook Air the same way it sucks on my Mac Pro today, so no real loss there, since I don't compile WebKit anymore. Again, its mainly just indifference now. I don't know if I miss the days when this stuff was fun, but Apple has certainly finally succeeded in making me see their computers as just tools.