New Mac Pro and Pro Display XDR Offer Power for a (High) Price

I think you’re onto something. This system might cater to movie makers or high-end creatives, but in science from all I can tell Apple is done in terms of workstations. You’ll still see loads of MBPs and some people use a mini or an iMac for desktop work. But in terms of number crunching and real scientific computing, it’s usually all Linux on generic iron these days. Apple’s last stronghold there was scientific visualization and that’s dying too now that Apple has determined that GPUs are solely geared towards FCP. Many years ago Apple gave up the Xserve (of which we had several dozen here in our department alone), then they screwed up the MP with the trash can’s poor design followed by years of neglect, and now together with the continued refusal to provide for CUDA on macOS they’re making sure they will be left out while science moves on to embrace big data & ML. Eventually they’ll probably kill X11 support (uncool and too technical) or an ARM transition (hey, it works for iPad) will kill off simple porting of Linux x86 binaries and then they’ll really be kicked out.

It’s a shame that Macs in science are being relegated to basically (shiny) netbook status. They’re still popular as heck notebooks, but their use is essentially for office work and manuscript writing. Actual scientific computing happens on other boxes people connect remotely too. Code can be developed and debugged on such systems too, no need to do that on a Mac. Kind of a local small scale netbook/cloud implementation relegating the Mac to a dumb terminal conduit to where the real action happens. Lab equipment control and scientific computing on Mac used to be really strong but are becoming increasingly rare (and often it’s just legacy). A shame really. Apple had a great run in edu and science. But I guess these days that’s no longer cool. It’s been a while since top Apple execs had actual academic experience and understood what really counts in the lab.

And all of this I observe at UC Berkeley, where we’ve traditionally always been very Apple friendly.

Simon, storage is way, way different these days. Spinning platters and pokey SATA worked great… 20 years ago. It’s all in solid state memory, like we used to talk about sticks of RAM, we now talk about sticks of storage. I have 1T of 2 SSDs on a PCI card that currently can be 300-350 bucks or so that is literally 7 times faster than any of my platter drives. AND TODAY for under $200, I can get something that is 2-3 times faster than those SSDs!

I was curious about the scientific community… I used to have points of contact back when I worked for a sci-tech publisher, but am long retired. I had a vague memory that CUDA was a pretty big deal there, so your mentioning it is significant.

2 Likes

Yes, research computing has continued to move more and more into datacenters and/or cloud providers. That includes access to GPUs, including those that support CUDA, for using that hardware to perform computations. The desktop or laptop is largely a dumb terminal There is no computer Apple or any other manufacturer can provide that would reduce that trend. You’re right that scientific visualization is a particular category where there’s still some value in having some “oomph” in the computer in front of you.

Given all that, the threat to Macs is Windows in this sector. Windows Subsystem for Linux (WSL) keeps getting better and is probably a better choice than macOS for people who want tools like they find on their servers; macOS tends to have dated and/or different versions (BSD vs. GNU) of open source software while WSL lets one use software from standard Linux package management systems. Obviously, lots of people still prefer macOS to Windows for everything outside the terminal window.

I know Nvidia blames Apple for the poor (and since Mojave, non-existent) support for their cards on Macs but Nvidia is responsible for making the drivers for their hardware. I think they decided there was enough value in putting in the work to write drivers for Apple’s APIs, particular the “Metal” API. I saw a mention of on an Nvidia forum that maybe the upcoming DriverKit could make a difference but Apple says the framework is “to create drivers for USB, Serial, NIC, and HID devices,” video cards aren’t on that list.

cwilcox
Curtis Wilcox

    June 6

Simon:
Kind of a local small scale netbook/cloud implementation relegating the Mac to a dumb terminal conduit to where the real action happens.

Yes, research computing has continued to move more and more into datacenters and/or cloud providers. That includes access to GPUs, including those that support CUDA, for using that hardware to perform computations. The desktop or laptop is largely a dumb terminal There is no computer Apple or any other manufacturer can provide that would reduce that trend. You’re right that scientific visualization is a particular category where there’s still some value in having some “oomph” in the computer in front of you.

I think that this a very big reason is why Apple has been moving so aggressively into services, and why they pioneered, and continue full speed ahead, with mobile devices.

Given all that, the threat to Macs is Windows in this sector. Windows Subsystem for Linux (WSL) keeps getting better and is probably a better choice than macOS for people who want tools like they find on their servers; macOS tends to have dated and/or different versions (BSD vs. GNU) of open source software while WSL lets one use software from standard Linux package management systems. Obviously, lots of people still prefer macOS to Windows for everything outside the terminal window.

And I think this is a big reason why they developed a super duper, fully tricked out, rocket engine new MacBook Pro. A very big majority of high end creative and production pros might know everything about, and can do any kind of incendiary effect, using VFX software like Smoke, Flame and Maya (which they used for the dragon and destruction scenes in Game Of Thrones), which only runs on Macs or Linux, but will turn into quivering masses of jelly if someone accidentally kicked out the plug their computer is connected to.

BTW, a year’s subscription for one seat of Flame runs $4,205, Smoke $1,545, Maya $1545. And that’s just for the fire. You need lots of different software to build and animate the dragons, though Maya is probably used for both. And there’s a lot of other software necessary for burning people (especially if actors in live action are involved), constructing and collapsing virtual buildings (and bridges and boats are different ), etc. And different stuff is necessary to create ice and explosions. It’s probably why HBO only kept the number of dragons to three.

The new Mac Pro doesn’t look so expensive at all when you consider what it costs to run editing and VFX software. And if the new Mac Pro will enable the production company to save $$$$$ because it will get things done faster and easier, and as it is upgradable, it will be an investment that’s well worth it, and one that will pay out over years.

I think we are in adamant agreement. For those, like the original poster, who think $5,000+ is “high”, that figure pales in comparison to what it cost early adopters to own computers in the past.

Well, this kid bought a fully tricked out Amiga back in 1988 instead of the 20% downpayment on a house.

Just spent about 5k on an iMac. It’s a lot for us to spend, but it’s vital for my work and the investment will play out for years on my desk.

1 Like

As many of you have, been devouring everything I can read about this new machine. ONE piece seemed to me to stand way out ahead of everything else… I learned more stuff about what is going on in this article than anywhere else:

I have reflected on my feelings… I, and most likely almost every damn Mac fan who has asked for an expandable machine they will buy, now feel hugely let down. NOW we are hearing that we aren’t “professional” enough because we simply can’t afford (or justify) a $40K workstation. WE are the folks who kept buying their hardware through all the turmoil… we were the loyal cadres. NOT the big film production houses. I DO know someone who spent a few years at ILM… their workers loved Macs but they worked on very different hardware.

As an enthusiast (my Pro days are over) I am way curious what they have come up with… I DO want to know what is going on under it’s hood. I have no issue with them building such a machine, BUT I take very serious issue with the fact they COULD have very easily given what I dare say many thousands os Mac users have wanted 2 years ago… and release this one when they will. This “MacTaxWrite-Off for Huge Studios so We Can Pay 25 bucks to See their Work Product!” is kind of like a slap in our collective face. They did NOT make this for me… or any of hundreds of thousands of others. Making a machine for us would cost them almost nothing, it’s actually all very straight forward stuff. Their choice was to ignore us.

AND if any marketing types may read this… guess what? NONE of the purchasers of this machine are in any way candidates for any of your “services.” Think any big time film production house if going to buy 100 subs to Music or News+ along with their order of 100 machines? Think you’re going to sell that sub stuff to the folks you turned your backs on? Not this guy I can tell you that.

Are you a fan of the show “Arrested Development?” Like Gob’s suits, every time you mention the new Mac Pro, the price goes up. Please don’t exaggerate, reality is bad enough. The reality is the entry-level price of the new Mac Pro will not be in the five figures, it will be $6,000. The entry-level price of the “trash can” Mac Pro in 2013 was $3,000.

I assume there are good arguments for why the entry-level model is “worth it,” it may have comparatively higher specs than the entry-level trash can did and a longer expected lifespan (because of better internal expandability), that doesn’t change that fact that it will cost more to “ante up.” I work with a group in the market for Mac Pros; they got 4 trash cans 5 years ago in more of a mid-range configuration. This year, 4 entry-level Mac Pros will require, at minimum, $5,000 more than those did, probably more than to get better than entry-level video cards. Most likely, one of the four staff won’t get a new Mac Pro and the group will spend more time shuffling work assignments so the one just does the “easy” jobs.

Plenty of people want an expandable Mac. Fortunately, every Mac with a Thunderbolt port is expandable. For those who, quite reasonably, should get one without a built-in display, the current Mac mini is a good, performant computer. Out of all Macs right now, there is a Mac mini config that is in the top 5 for single-core performance and top 10 for multi-core, beating all the trash cans except the 12-core model. In a few years, if Apple again neglects the Mac mini product line, it won’t look like a good choice and people will, rightly, complain again.

1 Like

Thanks for the AppleInsider link, I was wondering if the longevity of the new Mac Pro would be hampered by a soldered processor. Thankfully, it’s a slotted processor. The article touches on another risk to longevity, it having PCIe 3.0 slots. They probably would have had to delay the Mac Pro release even longer to have used PCIe 4.0, hopefully there won’t be a spate of cards in the future that won’t work or will have their performance hampered by the slot’s bandwidth.

The ultimate risk to longevity is duration of macOS support and support of applications on less-than-current OSes. MacOS Catalina will not support the 2010 Mac Pro but it will have had 9+ years of official macOS support (possibly requiring a video card upgrade), not bad. We’ll see what 3rd party software vendors do.

Apple’s pro computers were always a halo product, which trailed recommendations for the rest of the product line. Tim Cook just figure this out last year it appears, hence the new Mac Pros finally. After neglecting our business for eight years he’s scrambling now: with a computer which costs more than a small car.

Apple’s pro computers were always a halo product, which trailed recommendations for the rest of the product line. Tim Cook just figure this out last year it appears, hence the new Mac Pros finally. After neglecting our business for eight years he’s scrambling now: with a computer which costs more than a small car.

What Cook did was wait until high end pro software had evolved enough so that a totally reconfigured, highly and easily customizable box that could would make a difference to the production schedules for architects, video and film editors, animators, VFX pros, retouchers, game designers, etc. I don’t know much about audio production or sound design, I’ll bet the new Mac Pro could be a valuable addition here, as well as for scientific teams.

As I mentioned before, this new box is clearly not aimed at the average TidBITS Talker. Take a look at what competitive boxes as well as stands cost at Adorama or B&H; the majority cost a lot more for a lot less.

Well, you shouldn’t be betting then because I do know much about scientific teams and know that for us this new box brings nothing to the table. It does not deliver on what science wanted, similar to the last MP. For science to large extent, the trash can was a failure and this new box does little to fix that screwup.

The new MP might be the greatest thing that has ever reached TV studios for all I know, but nobody needs to come here and conjecture about its potential in science. We moved on when Apple told us to **** off because they preferred doing hip hop radio and gold plated wrist bands.

1 Like

I know nothing of what science wants, so this isn’t a challenge to your evaluation at all, you’ve just got me curious. What is it that science wants?

1 Like

@SteveJ1, this just from the view of our department.

There’s a lot they could do to improve things on the software side and it’s simple to mention many things they could do to improve the MBP for us. But let me just summarize for the desktop. On the low-end something like the mini is nice, but what we need is real desktop-class CPUs with a case that does not require downthrottling or mobile low-wattage CPUs, one slot for a GPU (not necessarily for graphics), and support for Nvidia/CUDA. 4-8 RAM slots and easily swappable disks make sense. TB3 and external expandability is great, but it should be in addition to not in replacement of external expansion. On the higher end it’s essentially similar, but more of it all. You’d want support for Xeons, some people I work with used to demand two CPU slots, personally I don’t care as long as it supports the best CPU core/mem bus combo Intel has to offer. Several full slots for at least two, ideally more GPUs, again CUDA is most important. As many RAM slots as the MCH will support. Internal disk bays and slots for SSDs. Sure, that sounds like a boring tower (Apple would make it non-boring, eg. the PowerMac, early MP), but that’s exactly what a lot of my colleagues need. It’s awesome if it looks nice and is quiet and fancy and yada (and we’ll pay for that no question), but all of that is for naught if it doesn’t have the required oomph. The last cheese grater MP came quite close, as did usually all MPs before that. With nothing left from Apple those folks will now just get some generic Dell or HP iron, put on Linux (CentOS or Scientific Linux usually) and log on remotely from their MBPs. Every one of those boxes is a sale lost for Apple because those people generally use portable Macs and they like their Macs too. They know they’re less hassle, usually need little to no dedicated support, offer good longevity and good warranty (AC) service. None of those people would get the generic PC stuff to save $1k (of course with this new MP, it’s more like $3k for what they want), but they will get it the moment Apple tells them ‘no more Macs for you’. That has been the case almost since the demise of the cheese grater. The mini is too limited, the iMac not adequate, and the MP somewhere between out of balance and expansion challenged.

What scientific field are you in, Simon, and if you’re talking about other fields here too, what are they? Just curious…

1 Like

Physics (and chemistry).

1 Like

While I have absolutely no doubt you’ve worked successfully in high level scientific teams, I do know that there are an infinite number of types of scientific and engineering teams. I do know some people who work in some of them. My cousin who has worked with NASA for decades and is also an astrophysics professor has fallen in love from afar with the new Mac Pro, he says other people he works with feel the same. A quick search verified something I remember he told me about years ago; the Mars Rover was not only made on Macs, it had the same processor as a G3 Mac:

http://osxdaily.com/2012/08/11/mars-curiosity-rover-mac-g3/

https://appleinsider.com/articles/12/08/06/nasas_control_room_flooded_with_macs_during_mars_curiosity_landing

Some of my friends and family members involved in medical and pharmaceutical fields also prefer Macs, especially those involved in imaging and research. I called a radiologist who told me about this application that runs on the current Mac Pro, and that it’s much more cost effective and user friendly than dedicated hardware that costs tens of thousands of $ more:

https://en.m.wikipedia.org/wiki/OsiriX

While the creative market is an easy target that’s most likely to immediately shell out mega wads of cash when the new Mac Pro is released, I will still take big odds that Apple isn’t thinking about just selling Watches when it comes to healthcare. $15-25,000 bucks for the highest workstation and display is a lot more cost effective than $50-80k range, especially when tricking out specialized work groups. In fields like research, radiology and microsurgery where imaging and visualization are critical, calibration and 3d are equally critical. And in addition to power, flexibility and ease of use, Apple has privacy and security advantages. Apple Watches and iPhones are already collecting data that helps facilitates the development of apps that hospitals, doctors and researchers are already using to monitor health and fitness, ResearchKit and CareKit:

https://www.apple.com/researchkit/

Though I never heard of CUDA before it was mentioned in recent threads, I have worked in advertising and strategy with clients and ad agencies in consumer and trade electronics, imaging, pharma and healthcare accounts for decades. I have an excellent eye for marketing opportunities. There are a wealth of scientific communities that Apple can, and will, bring to the table with the new Mac Pro, monitor and stand.

Hi Paul and everyone else,
not to be condescending, but take a look and some of the software companies statements on this. Many of the producers of 3D and video editing software including Adobe, have announced they are going to be adding Metal support to their applications. Let’s not think the creatives need CUDA until the new software is available that supports Metal. Hopefully it will be available before the 2019 Mac Pro is, and people can try it on existing Macs with AMD graphics cards.

1 Like

The new Mac Pro is apparently too cheap. At these cut throat price points, manufacturing had to go back to China.

1 Like