As an Amazon Associate I earn from qualifying purchases @AMAZON
2017 iMac Pro: Usage Observations and Conclusions
Related: 2017 iMac 5K, 2017 iMac Pro, 4K and 5K display, iMac, iMac 5K, iMac Pro, Mac Pro, Macs, Other World Computing, RAID, SoftRAID, SSD, Thunderbolt, Thunderbolt 3, weather events
The data on this page and others was painstakingly obtained, made possible by the support of OWC / MacSales.com and B&H Photo. Please order through our links via these companies and/or subscribe—thank you.
Not sure which Mac to get or how to configure it? Consult with MPG.
On January 9, I switched to the iMac Pro 10-core 64GB 2TB Vega 56 for everyday work for a week or so. The about $7399 10-core iMac Pro on loan from B&H was my chance to see the iMac Pro prove itself in everyday work as a noticeable improvement over the 2017 iMac 5K. The 2017 iMac 5K has/had been my primary workstation for the prior 7 weeks.
So far the iMac Pro in daily use has at best has delivered parity with the 2017 iMac 5K. That’s not much of an endorsement for spending $2000 to $4000 more.
Be clear: Lloyd’s workflow is his own—see the tests and see for yourself whether an iMac Pro makes sense or not, taking into account how often tasks are actually done and whether any savings are significant in the context of total workflow. For example, video users needing real-time editing for 4K and/or night-long renders will find the iMac Pro a far better fit than an iMac 5K.
I should say a few things first which I cannot pin on the iMac Pro but which are troubling in general, and which I am inclined to blame on new bugs in macOS 10.13:
- In the course of two days, two different devices had their file systems corrupted. I cannot recall this ever happening before, over years of usage. Why two damaged volumes in two days? Two weeks prior I two SSDs had their partition maps toasted while plugged into the LaCie 2Big (Thunderbolt 3) on the iMac 5K. So the iMac Pro is probably not the culprit, but macOS, since one device was plugged in directly, the two SSDs were plugged in via the LaCie 2Big, and the AccelsiorPro Q stripe was in a Helios 2 enclosure hanging off the OWC Thunderbolt 3 Dock—no commonality except having to go through the Thunderbolt 3 bus with bug-ridden macOS 10.13.2.
- Strange performance problems. Mounting evidence in my own use: wild performance differences up to 4X in both Java and native code for the same task with the same software, e.g., 3.1 GB/sec versus 0.6GB/sec with minor changes in configuration or no changes at all. Along with similar weird sporadic issues makes me think there are serious new bugs in macOS.
- More crashes than ever before: my most used software, Dreamweaver, crashes about 50% of the time in the File Open dialog.
- It appears that Apple is not being truthful about the content of macOS updates any more.
The foregoing serves to say one thing to me: macOS 10.13.2 seems to have several loose screws rattling around inside. So in what follows I try to separate those types of issues from the iMac Pro.
Issues with the iMac Pro
A number of core design features are anti-features with the Mac Pro:
- The new secure enclave that requires action to boot off an external drive is extremely concerning in the even of an internal drive failure. It’s an anti-feature that has negative benefit in my view.
- I have not been successful in booting off any USB3 SSD with either macOS Extended or APFS. That is disturbing, because being able to boot off an external disk is a fundamental data safety and usability issue if something goes wrong. I am entirely dissatisfied with the idea that I can only boot off Thunderbolt 3 or the internal recovery partition. What the hell is Apple thinking here?
- The new secure enclave stuff creates an incompatibility with key software I use (SoftRAID). This came as a surprise to the SoftRAID developers . How can Apple ship a new machine and not give a sh*t about breaking software of that kind? I could see that as an isolated circumstance—but in the context of the massive quality control failures, macOS has become a shit sandwich for customers like me.
- A pro machine by my definition is upgradeable. The iMac Pro is not upgradeable without special effort and then only for memory upgrades.
- Buying a a $7K to $10K machine that lacks support for 8K video (HDMI 2.1) is quite bothersome to me: 8K displays are one the few technology things coming that are highly appealing to me.
- The fact that dust cannot be cleaned out internally is worrisome over years of use. This is also true of the iMac 5K, but the iMac Pro is a much more expensive proposition.
I’m always open to spending more money if it can save me significant time, but the iMac Pro will cost more money and not only not speed up my work, but actually slow it
The assumption that the iMac Pro is faster for some things is true, as the tests show. It is also true that small differences can’t be noticed, and many of the iMac pro wins are small. Video users will probably love it and certain other special-task users. I am not one of those.
The key things I do are slower or no faster.
- Retouching focus stacks in Zerene Stacker: retouching lags noticeably with the 10-core iMac Pro; the 2017 iMac 5K has leg lag, critical when retouching. This is a BIG DEAL because I usually spend at least 15 minutes retouching a stacked image (an interactive process) and sometimes up to two hours. There is nothing 'pro' about slower retouching. The fact that the iMac Pro can do the initial stack about 50% faster is irrelevant in the context of 10 minutes to 2 hours of retouching (20 minutes is typical, still far longer than a few minutes to stack the images, where I can do something else anyway).
- Creation of images series is a multi minute task (as much as 12 minutes) and there are days that I do 5 or even 10 of these. The 2017 iMac 5K beats out the iMac Pro for making image series—all of them.
- Adobe Camera Raw in Photoshop: with the 2017 iMac 5K side-by-side with the iMac Pro 10-core, I pulled up 20 or so raw files, doing this several times with different camera raw file formats (Nikon, Fujifilm, etc). I used the arrow keys to simultaneously go to the next or previous image; this requires a multi-second render. The two machines finished that exactly at the same time, at least within my perception trying it repeatedly. So the iMac Pro offers nothing here.
- General Photoshop work—no difference.
- General operations: my impression is that basic things of many kinds are either no faster with the iMac Pro, or a touch slower—the 2017 iMac 5K is faster.
See the problem? There is no 'win' for my work.
On the downside, the 2017 iMac 5K is much more prone to fan noise kicking in right away, whereas I have yet to hear the fans above a whisper on the iMac Pro.