My apologies for the delay on this post. This concept had been floating around in my brain for a few weeks and I still feel as though I haven't done it justice, but I also know I can't keep it around not posting it for any longer. I'll likely revisit this topic entirely in the indeterminately distant future.
One of the things I see on a pretty regular basis is that there are essentially two types of Mac Pro owner. One type is the type of person who is regularly using its CPU horsepower or is using its expansion to do things that the models lower in the product stack can't do. The other type is the group that isn't really using most of that expansion, and isn't even using a lot of the computing power of the machine, but feels as though for the past fifteen to twenty years, Apple has been on a path of betrayal to its "core users."
The "core users" bit is always particularly funny to me, because although the most vocal Apple fans will tell you that the only models worth paying attention to at any given point in history are the really big, highly expandable professional tower, it's typically the inexpensive consumer model Apple sells a bunch of that's keeping the platform afloat. These types of users like to throw around dollar amounts for just how supportive they have been to the platform over the years, and talk about how many referrals they've made. I always gently try to remind them that while it's nice their Aunt Tillie bought an iPad or a MacBook Air, Apple sold literally hundreds of thousands of them this quarter alone. It's not to devalue the influence they have over their family and friends as much as to remind them that we're no longer in a time where adding twelve new users to the platform is significant.
In 1984, when the original Macintosh shipped, the guiding idea behind the entire platform was that it was "the computer for the rest of us" and part of that was that it was essentially an appliance for productivity purposes. I'd argue that Apple has held onto that as a guiding concept for the life of the platform so far. In various ways, commensurate with the context at the time and the capabilities Apple felt they needed to provide, the Macintosh has been a computer for people who care about using a computer to get work done, but don't specifically care about the computer.
As such, these "core users" are often directly at odds with Apple's goals for the platform (while also being a drop in Apple's profit bucket), because they are in it specifically because they care about the computer itself. This isn't a bad thing, but this group tends to want more hardware choices and choices of more flexible hardware, and place a higher value on longevity (or even getting several years out of a used system) than on something being a fast or smooth experience.
While looking for information on building up an old Mac Pro, one of the members of Ye Olde Computer Forum discovered Thomas Pindelski's web site. He has an extensive guide to almost everything about the 2009 Mac Pro (except for how to let it go) and of course, he talked about the New Mac Pro. Pindelski has some Capital-O Opinions™ about the New Mac Pro, based on the fact that he is a "Mac Pro customer." I've taken a piece of the text to specifically talk about:
The driving thought behind this design – other than trying to look innovative – is the same as behind almost every Apple product. Take away user choice and lock up the box. The last Apple machine to offer user expandability in any meaningful way is the current MacPro. As Apple disregarded that machine as it morphed into a cell phone maker, the 2012 (more like 2007) MacPro is seriously overpriced and seriously dated, easily outperformed by any number of MacBook Pro laptops. Laptops! This has left a lot of graphics pros unhappy, with many migrating to high end HP and Dell workstations, trading the horrors of Windows for the speed of current silicon.
Naturally, I have some thoughts about this:
- Apple (as they so often do) had to invent or significantly refine several industrial and manufacturing processes to build this machine. We saw the same thing happen when the Retina MacBook Pro was announced, and in 2009 when the Unibody MacBook Pro was announced. You may not agree that this is a good form factor for a computer, but you can not deny the fact that this is innovation.
- What user
choice has Apple taken away? Even in regards to internal components, what could you add to a 2012 Mac Pro that you couldn't in some way add to the new model?
- The Mac Pro itself isn't even locked up, and adventurous users who do not care about Apple's warranty are free to install CPUs Apple doesn't make available, bigger RAM sticks than those that are officially supported, as well as (potentially) bigger or faster storage modules.
- Mac OS X isn't locked up either, certainly no more today than it was a decade ago, but it's just convenient if you pretend that it is, or that this aspect of the platform is changing.
- This was rectified later, but his commentary about the Mac Pro being outperformed by laoptops was apt. At the beginning of 2013, the Mac Pro was still on the Westmere micro-architecture (the tick after Nehalem, introduced in 2008), and had completely skipped Sandy Bridge. Even the new round Mac Pro is an Ivy Bridge-based machine, not a Haswell. Clock-for-clock, unless you can exceed six threads, MacBook Pros and iMacs are still definitely going to outperform the Mac Pro. But, the Mac Pro has never exclusively been about CPU performance anyway.
- This is just a personal thing and isn't even related to Macs, but Windows on a Dell or HP workstation should actually be a pretty good experience these days. It's in that market that there's almost never any junk-ware reloaded on a system, and those systems tend to get AppleCare or better levels of support, including accidental damage coverage, typically for up to five years.
Pindelski continues with some pretty interesting misinformation (and also, who is going to move tired old 1.5 and 3 gigabit disks they've had for five years into their hot new system? As a working photographer, can you really not afford a $100 disk to go with your $3000 computer?) and a few real concerns for what I imagine is a very small part of the Mac Pro market: users of expansion cards.
It's also interesting that through the site, Pindelski advocates so heavily for Hackintoshing. I have some reading to do on that front, I didn't think anybody considered it a reasonable option for professional Macintosh users at this point. Of course, his phrasing is:
It's too bad more high end users will not make the time to learn the simple process of creating a Hackintosh. For 20% of the cost of a 2013 MacPro you get an ugly box, state of the art components and exceptional cooling, and no one will laugh at you when you show them just how ugly your box is.
Of course, I literally laughed out loud when I read that. It's essentially impossible to build a system that's performance- and functionality- competitive with the Mac Pro for $600. A MacBook Pro is faster, but not that much faster, and more than $600 anyway. It's also funny to me that Pindelski (and this isn't the only place he does it) is so focused on the fact that the box he built, in an Antec Sonata case, is 1) so ugly 2) so good at cooling. The reality of it is that there's literally nothing special about the cooling on his Hackintosh, and in fact, it's merely average for about 2005.
My question is whether he places more value on "performance" he'll never get out of a single-socket desktop system or expansion he'll never use. For $600 you can get a "reasonably expandable" computer. In fact, the 7/9 series OptiPlex and Dell's media enthusiast XPS desktops are almost exactly what every MMMM advocate has been talking about for just shy of a decade. It just has to use the desktop Intel CPU, it has to have a discrete GPU sthat I can swap out, and maybe one or two other slots, it should hold two 3.5-inch hard disks, and I should be able to choose to put one or two optical drives in it.
It sounds like the only thing preventing the Dell OptiPlex 7020 from being Pindelski's (and many would-be Mac Pro buyers) ideal computer is the fact that it doesn't run Mac OS X. However, I wonder what release of Mac OS X he is running, given that he has (almost certainly) bought into the idea that the newest releases of Mac OS X are turning into iOS, by default, and with no configuration options.
Even if the OptiPlex 7020 is the author's ideal computer, he's overselling it massively. It's very, very, very far from being a Mac Pro competitor, in so many ways. You may be able to order it with a Radeon R7 240 and an i7-4790 (which will slightly outperform the quad-core Mac Pro's CPU, but primarily due to the generational IPC improvements and an additional few hundred cycles. That doesn't make it a Mac Pro. It's a reasonable competent ExcelBox, but even its ISV-certified brother, the Precision T1700, isn't really a Mac Pro competitor either. The old Mac Pro is best compared to one of today's 2-socket workstations, such as the Precision T5 series. The new Mac Pro is best compared to one of today's one-socket Xeon E5 workstations, such as the Precision T3 series.
The other thought I had was that if you're going to try to convince buyers of machines whose entire point of existing is "this is a high-end computer you can treat like an appliance" to learn how to build and use a Hackintosh, then why not just encourage them to use Windows? Windows itself is nearly as trouble-free these days, your apps will likely perform better, your data will be safer, and you'll have your choice of hardware. (This is of course after teaching the creative professionals how to determine what hardware they'll even need.) Learning Windows is almost certainly easier and faster than learning how to assemble and maintain a Hackintosh, and as a general rule, Windows on workstation-class hardware is going to be smoother sailing than a Hackintosh.
Pindelski says he's moved on from Mac Pros, and now that they're inexpensive, is using the 2009-2012 Mac Pro as his new computer. This almost makes less sense than building a Hackintosh. You have a system that has Mac Pro "features" but its several generations old at this point, and the only way to get about as much performance as the current Mac Pro is to spend thousands on the highest end CPUs and completely fill the system with expansion cards. Though, this may be fine considering he moved up to the 2009-2012 Mac Pro from an Excel-wrangling class machine.
If the point of looking at used Mac Pros is to avoid buying a system you morally disagree with, maybe you should really consider just finding a new workstation vendor. The new Mac Pro form factor is almost certainly not going away, and the CPUs, RAM, and built-in disk controllers on the old Mac Pro are never going to be faster than they are today. The old Mac Pro for all it does have also only has so much expandability, and given how old the platform is, we can't possibly be far from the point where the newest and hottest GPUs outgrow that system, requiring either faster interconnects or more CPU horsepower in order to be fully utilized. You also can't get dual GPUs into the old Mac Pro without jury rigging power delivery and giving up on almost all of the rest of the expansion. Choosing to use dual GPUs on the old Mac Pro means that you lose any room to use a faster storage controller or better or more external interconnects. (A point I always found mildly ironic given that the whole point of lobbying so hard for the big Mac Pro form factor was that it had so much room for internal expansion.) With the introduction of the GeForce 980, you have a better chance of outgunning the new Mac Pro's dual AMD FirePros in a single card, but that's something that didn't happen until 18 months after the introduction of the new system.
The biggest problem with Pindelski's 2009-2012 Mac Pro tech guides may already be obvious: what do you do with the system when a Haswell, Broadwell, or Skylake Mac Pro launches in the next few months. This system will have faster GPUs, processors with more cores, the big improvement in IPC from Haswell or newer, and most likely a boost from using DDR4 memory, and of course PCIe based storage will have gotten faster between now and then.
This crowd likes to talk about how the New Mac Pro almost isn't even a professional computer. This makes no sense though because through the relatively short history of "workstations" there have always been small workstations with relatively limited expansion capabilities. The Silicon Graphics O2 is probably the best comparison to the new Mac Pro. It wasn't the computer you would buy if you needed a whole lot of memory (in fact, it may on its own have benefitted from more memory capacity, a problem the Mac Pro doesn't have) but it was a suitable system for UNIX chores, content creation and ports of Mac applications from the time. The O2 wasn't the only UNIX workstation that had about that level of performance and expandability. Until the 2000s when the RISC UNIX workstation died out and was replaced by Windows and Linux on x86 hardware, smaller workstations were actually fairly common. I don't know if it's a matter of the computer market in general becoming more stratified, or if in order for a Windows or Mac computer to "count" s a workstation it simply needs to be huge, regardless of whether or not it actually uses "workstation" components.
This has always been the kind of issue facing the 2013 Mac Pro. Because it only has a single CPU socket, and because the base model has a CPU that's about as fast at most tasks as a top end desktop i7, people don't perceive that it's still a better computer in other ways, such as by way of its dual GPUs or the capacity to use more than 32 gigabytes of memory.
Although the Mac Pro isn't one of these either, there seems to be a lot of growth in the low end workstation market. Systems like the HP Z220 and the Precision T1700 are not all that different from machines such as the OptiPlex 9010 and 7020 or the HP EliteDesk 800G1, but they do have better power supplies, better GPUs, and are certified for use with whatever software application you're using.
Of course, and this is more general Mac Pro commentary, there's also the issue of Apple using the Mac Pro to encourage its software developers to focus on certain things. Apple has been introducing GPU-centric features to Mac OS X for years now, starting with Grand Central Dispatch and OpenCL in Mac OS X 10.6. I think the big problem with GPU computing (on both Mac and Windows systems) is that many applications (such as the Adobe Creative Suite) are programmed using nVidia's CUDA APIs. CUDA works on both Macs and PCs, but it does not work with AMD's GPUs. By putting only AMD GPUs in the Mac Pro, Apple is sacrificing some short-term compatibility with CUDA in an effort to get the developers to port code to CUDA.
Of course, it doesn't hurt that AMD almost certainly gave Apple a really great deal on the FirePpos in the current Mac Pro, keeping the cost low, or at least lower than it would have been with two Quadros. For software developers to move to OpenCL, however, It means that Apple and customers both have more hardware choice, which is important to their customers as implicitly, Apple's choice is your choice. Apple knows this and is willing to play the long game to make sure that either of those choices is effective and performant for their customers.
And this gets back to the core thing I think is the difference in the values that Apple has from its "Mac Pro customers." People who aren't really Mac users were choosing to buy the Mac Pro because it was the only Mac OS X computer that met their needs. I think it was this same (very vocal) group that was advocating for the MMMM and I believe that when this group talks about the other groups (such as the people who actually buy a Mac Pro only really intending to use it for their intended task, with an intent to replace it with a new system when the available performance increased enough.
Unfortunately, there's no "solution" for this crowd. Hackintoshing isn't realistically sustainable, and as such if you want to use Mac OS X, you either are willing to buy Apple's hardware, or you have to give up on having hardware that works as well as Apple's does for a problem-free Mac OS X. Though, given the fact that a 26-part guide about upgrading the 2009 Mac Pro exists on the same web site that has at several points in history very heavily advocated for Hackintoshing tells me that people willing to put in effort for Mac OS X on their choice of hardware definitely exist. It's neat to see it happening, but it has never been what Apple has wanted, and it's typically a much poorer experience than trying to run Windows on random hardware, simply because a legitimate industry can't be built around writing drivers for Mac OS X on non-Apple hardware.