diglloyd Mac Performance Guide
In Motion There is Great Potential
SSD Wishlist…
USB-C Dock for MacBook

4 USB3 ports, 1 USB-C port, SD card reader, gigabit ethernet, audio ports, HDMK 4K port!

Checking drives before putting into “production”

diglloydTools

See also Even ”Enterprise Grade” Drives Fail.

According to Tim Standing of SoftRAID, initial verification of a drive can preclude a large percentage of relatively early drive failures.

While SoftRAID has a certify command, it does not graph the performance, and so I would add this point: in my experience, drives that show aberrant behavior are also more likely to fail, even if they pass certification.

The diglloydTools DiskTester fill-volume command can test 99% of the drive and graph the behavior, as shown below, where 5 samples were tested simultaneously, and then graphed together to verify consistent performance—important for RAID setups.

Over the years I have found that aberrant performance behaviors (obvious in a graph) are often an excellent predictor of flaky drives. The test-reliability command is good too, with the major benefit of being able to operate on in-use drives—no need to take a system down for days to certify (which means having to completely wipe) the drives.

Show below is drive performance for five samples of the HGST 8TB Ultrastar He8 hard drive. All samples deliver the same pattern of declining performance as the drive fills up—exactly as expected. Drives with excessive remapped sections tend to show weird spikes in speed in the wrong places. This is bad for RAID performance as well (when drives do not perform the same in the same area).

Performance across 8TB capacity of HGST 8TB Ultrastar He8, 5 samples

Lifespan of Uninterruptible Power Supplies (UPS)

Get APC UPS at B&H Photo.

A bad smell developed in the garage, where I keep a couple of servers running on uninterruptible power supplies (UPS). Two batteries, each weighing about 75 pounds, are daisy-chained for extra “juice” off an APC SmartUPS 1000XL. I do so because I’ve had power failures up to 24 hours.

Both of the add-on batteries had leaking battery acid, or rather, signs of it having leaked such as white powder and blackened wood under the units and what looked like fluid on the plastic casing. I disconnected them and set them aside for the next hazardous waste day.

UPDATE 18 Jan: APC tells me that the add-on batteries were from 2005. I’m skeptical since that’s the year I started my business and I don’t think I bought 150 pounds of batteries way back then, but that’s the claim. If true, the batteries were way past their lifespan. But the APC SmartUPS 1000XL gave no warning or indication of a problem.

The main unit with the “smarts” to which these batteries were daisy chained does not appear to be leaking, but its battery is now over 3 years old. The unit is running hot even at an ambient temperature of 40°F, and the charge lights show only about 80%, so it is done-for. These things should never get even warm under such conditions, so that battery is surely toast.

This is not the first stinky leaking battery problem with APC; I had this happen just as I was leaving for a trip last summer. That time, the battery had bulged so much that I had difficulty extracting it. I gave up and it went to hazardous waste disposal.

Maybe this is my fault: after all, lead acid batteries have a limited lifespan. But the fact that the APC unit cannot detect such potentially serious failure situations is disappointing and calls for taking care to assess the health of any UPS on a regular basis (3 months is probably about right).

I’m probably going to pick up a new APC Smart-UPS 1500VA with LCD and Audible Alarm Disabled, which has pure sine wave power when when running on battery power. It can also take additional monster batteries which plug into the rear of the unit, for extended runtime.

Tips:

  • Check your UPS battery every few months for signs of leakage: put an appointment in your calendar to check.
  • If the UPS is running hot, there is a very good chance the battery is bad. It could leak or rupture and cause damage. A fresh UPS does not run hot (once fully charged), since it is just bypassing current to devices.
  • Use a marker to write a date onto the battery itself, as shown.
Interior of APC SmartUPS 1000XL

Jon L writes:

I have used replacement batteries from this source of several years now. My backup demands are not as demanding as yours, but I have used several replacements from this firm that have served me well.

Below is the link to the APC RBC7 battery mentioned in Don H's reply to your post.

http://www.batteryspec.com/cgi-bin/cart.cgi?action=link&product=182

The listing of replacements for all APC UPS Systems:

http://www.batteryspec.com/html/apc-ups-replacement-batteries.html

This company has been in business for > 30 years so they are not a flash in the pan. Substantially less expensive than APC replacements as well. I’m sure you know more about the technology behind these batteries than I do.

You might take a look. No relationship with the firm: just a satisfied customer.

MPG: I’ve been leery of 3rd party batteries, but the prices above are far lower. But this site has way more brains to their approach: a listing showing UPS models and their batteries. It is WAY better than the APC site. Also, it's about 30 miles from me. I’m going to see if I can just drive down their and get things swapped out.

One thing APC *does* include is free return postage for disposing of the old one. I know little about the issue; I’m just a customer who doesn’t want an exploding or leaking battery. And since 3 (actually 4) APC batteries have leaked or bulged in the past year, I’m open to credible alternatives at a lower price.

Don H writes:

In the 1990s I worked for a growing company and had to spec out a UPS large enough to support five servers for a six-hour outage. At the time APC had the best modular system, so we bought that and it performed well. This was back when Cisco’s product line only had three basic models of routers while Apple had a sprawling product matrix with far too much overlap among machines. (We called that phenomenon ‘Spindlerization’, due to Mike Spindler’s penchant for trying to match the PC world model-for-model.)

Over the years I have bought and used APC UPSs for personal use, but noticed that they too expanded their product line to the point that one can no longer navigate it. (Meanwhile, Apple famously managed to rein in their own product line.) One consequence of this lack of cohesion is that many UPSs became orphaned because the batteries were no longer available, or became so expensive that it was not economically feasible to buy a replacement battery for almost the cost of an entirely new unit. And this is what irritates me about a lot of companies: the changes made to the parts are all just different enough to thwart cross-product or backward compatibility, yet the changes are not significant enough to provide any functional benefit.

Sony did this with their consumer electronics power supplies (it seemed like every new product changed the charging connector and power brick.) Wristwatches (remember those?) have a battery variety that fills fat books. Samsung’s entire business is a pathological example of change for the sake of change. (I have no idea how their service department handles it all, partly because I will never buy a Samsung product.)

I have discarded otherwise functional APC UPSs because it was too frustrating to chase down replacement batteries. The last time I looked, their web site was a mess*. Because of these practices I have written them off - APC is dead to me. Other UPS companies have similar, but not quite as acute disorganization, but I haven’t had the bad experience (yet?) with other brands that I have with APC.

Well, crap. I just looked at Tripp Lite’s product line. That’s 279 different models, each with one insignificant feature different from a dozen others.

I guess it’s better to have an embarrassment of riches rather than nothing at all, but I am personally getting really tired of hacking my way through other companies’ wasteful lack of focus or product cohesion. Your post on APC just happened to trigger this particular rant.

* Ok, I just looked again and the first thing I see at APC is an ad saying “Want to turn APC products into Recurring revenue?". I know the intention is to sell ‘managed services’ but my immediate interpretation is that they have figured out a way to milk their own customers with post-sales water torture costs.

But I stand by my complaint about their ridiculous product line. I would be willing to bet that they now have three or four hundred UPS models just in their Back-UPS and Smart-UPS categories, and that’s not getting into their enterprise ’solutions'.

MPG: indeed, it is a very confusing product line with hundreds of options.

I spent 20 minutes trying to find the right internal replacement battery for the SmartUPS 1000XL (it is not listed in specifications or data sheet!). I am pretty sure that the right replacement for the APC SmartUPS 1000XL is the APC Replacement Battery Cartridge #7, RBC7. I did find the external add-on battery supplement page.

APC sucks in a lot of ways: the web site can be extremely slow, it is very hard to find the right battery (and no distinction between internal and external). Tech support inquiries require serial numbers—well it’s not see easy to get a serial number when a 50 or 75 lb device is in a tight space that does not make it visible. I might look at TRIPP LITE and see how that goes because evey time I have to deal with battery issues it chews up hours.

Which Camera System / Lenses Should I Get?
✓ Get the best system for your needs the first time: diglloyd photographic consulting.

Bill Atkinson’s “Photo Card” for iPhone/iPad: Tangible Internet-Age Postcards

Bill Atkinson is Mr. Hypercard, of Apple fame. He is a color expert and brilliant photographer. Today I had the pleasure of his company on several topics, including his latest creation.

Bill showed me his latest creation, the iPhone/iPad app “Photo Card”, available on the Apple Store. Other platforms are coming, e.g., Android, and I am trying to persuade him to do a web interface for computer users like me.

It started simply enough—Bill showed me one of his postcards—printed and sent through the mail. The card is very durable and aside from holding it to see sheen to reveal the printed-on postal service processing stuff (or a UV light), it looks like it was just custom made and was never posted.

It’s the kind of thing you could not do half as well at home: I was astounded at the quality of the laminated card with excellent color. They’re way more good enough to frame—and no backing/support is needed in a frame so you can see front and back of the card. backing/support is needed in a frame so you can see front and back of the card.

More info

CNET: Apple legend Bill Atkinson's new mission: Save the postcard

Twit.TV part 1 and Twit.tv Part 2 and Twit.TV Part 3

How it works

First, you need to create an account and buy credits.

You choose your own image, your own stamp, and you can even add a QR reader for a voice recording. Very slick, very well thought out. For example, just entering the zip code alone looks up the city and state, saving time on addressing.

A preview after editing is shown below. The fish picture (mine) will be the front of the postcard. The stamp is a real postage stamp made with my own image, the bike is a graphic just for fun, the smaller fish picture is yet another picture of mine, and the yellow/blue thing will contain a QR code with a recorded voice message up to a minute long.

Two things from my POV: (1) the images have to be on the phone to be used by the app, which for me means copying and syncing to the phone first—a hassle. This is of course NOT an issue for shots made with the iPhone and already on the phone. (2) I would like to make cards on my Mac in an app or web browser because it is far more efficient for me to work on a computer, where all my photos and contacts are stored.

Bill Atkinson’s iPhone app 'Photo Card', preview of final card

Below, a not yet finished card.

Bill Atkinson’s iPhone app 'Photo Card', editing view

FOR SALE: Lloyd’s Apple Laptops, NEC Display

Cleaning house—recently bought a new 15" MacBook Pro.

Contact Lloyd

  • Mid 2012 MacBook Pro 13" 2.9 GHz Intel Core i7, 16GB / 480GB OWC Mercury Extreme Pro SSD / Intel HD Graphics 4000 512MB with charger $825. OWC sells a similar model used for $1200. Speedy little laptop fast enough to run all my web sites (it was a spare).
  • NEC EA244UHD 4K display $650 (sells new for $1049). See my review.
    A very nice 4K display (see my review), but I’m just not using it any more because of iMac 5K. Never saw many hours of operation, so backlight should have long life. Would make a terrific primary display for space constrained environments and/or an excellent 2nd display.
  • Late 2013 MacBook Pro Retina 15" 2.6 Ghz, 16GB / 512GB / NVidia GeForce GT 750M 2GB (top of the line except 512GB SSD) with charger in Apple box $1200 / BO. Original owner, was covered by AppleCare — no issues. Has a minor dent and scratches on case, keys worn.
    This was my workhorse laptop for the past 3 years and it just keeps going strong. Local buyers welcome to inspect firsthand. Still a strong performer even beating the 2016 model on some Photoshop tests. SOLD
  • 2011 MacBook Pro 13" 2.3 GHz Intel Core i5, 4GB / 200GB enterprise-grade SSD with charger $350. SOLD
    No speed demon, but a solid computer for the kids.
ThunderBay 4 - The Speed To Create. The Capacity To Dream.
Storage Wishlist…

Adobe Memory Usage, with no apps in use

See my Mac wish list.

Earlier today I wrote about macOS BloatWare.

Adobe Memory usage with no Adobe apps running

Now it’s Adobe’s turn to be spanked: I take a dim view of products using substantial memory when absolutely nothing is running.

These Adobe processes are just there running all the time even when no apps are in use, and they take up 132MB of real memory. On 4GB or 8GB machines, this is not a trivial consideration and if every vendor did this, pretty soon a GB or two of memory would be scarfed up to no useful purpose.

Adobe ought to have one bootstrap daemon using at most 10MB, one that can fork off all this overhead cruft when an app is actually in use. As shown, memory usage rises to justa bout doubles for these overhead processes when an Photoshop is in use—a huge amount for what is likely a very basic task.

SSD Upgrade for MacBook Pro Retina
Internal SSD Wishlist…

macOS BloatWare (and what feels like SpyWare)

See my Mac wish list.

Recently I’ve found that Photoshop CC 2017 on macOS 10.12.2 will no longer complete the diglloydMedium benchmark in 16GB of memory without a slowdown—compare 2013 results to 2015 results to 2016 results—the trend is substantially slower at present. What I observe here in 2016/2017 is a lot more memory swapping, which necessarily slows down the test.

So something has bloated in the system to push it beyond the point at which 16GB is adequate, causing that memory swapping: (a) the benchmark runs as fast in 2016/2017 as it did in 2013 when enough memory is availble and (b) Activity Monitor shows lots of swapping going on. See 2013 Mac Pro results, which show that time to execute in 2013 is within 1.1% of the 2016 figure. Ditto for the iMac 5K — but both have 64GB.

I think it likely that the slowdown is mainly due to a bloated macOS memory footprint, though I cannot rule out Adobe getting sloppy witih a lot of memory usage overhead.

We don’t have to look far for seriously bad software design: here is macOS Finder using nearly 5GB of real memory. This is a design obscenity, given that Apple sells a lot of laptops with 8GB total system memory.

5GB memory usage by macOS Finder

Why does all this stuff run that I do not want, some of which I have even turned off?

All of these processes consume memory and at some point, consume CPU time.

I was looking through running processes and wondering how much crapware and bloatware I might find. But off isn’t actually off in some cases, like Siri.

Qué Siri, Sera isn’t an acceptable answer. All of these were running on my 2013 Mac Pro, all take memory and none of these processes offer functionality I want active. Some like “CallHistoryPluginHelper” raise privacy concerns, and that’s not idle speculation.

Use "sudo launchctl list" to list active background processes. Alternatively, one can select all in Activity Monitor and copy/paste as plain text, but the two lists use different names for some processes.

The list below reflects turning off everything that I don’t want that macOS allows to be turned off, so the actual list for most users may be significantly larger/longer. As well, there are more daemon processes running than just these.


AirPlayUIAgent — I never use AirPlay
AirPlayXPCHelper — I never use AirPlay
airportd — I never use wireless on my Mac Pro
CallHistoryPluginHelper — raises privacy concerns: what call history? From my phone?
CallHistorySyncHelper — raises privacy concerns: what call history? From my phone?
callservicesd — for what exactly? Phone call lists again?
com.apple.photoanalysisd.plist — privacy concern to be analyzing photos without consent
findmydeviced Find My Mac is turned OFF
mediaremoteagent — I doubt this does anything I want.
mediaremoted — I doubt this does anything I want.
mobileassetd — I doubt this does anything I want.
parentalcontrolsd — with a single user and no parental controls set?
photolibraryd — I don't want Photos library
Photos Agent — I don't want Photos running
Sirioffensive - I have Siri turned OFF
Wi-Fi — I never use WiFi on my Mac Pro
wirelessproxd — I don't use or want wireless
WirelessRadioManagerd — I don't use or want wireless

If Console.app is opened, it’s a mess: a steady stream of sputum is ejected constantly. You’ll see this on every Mac, it’s voluminous, and a fresh system install brings the same mess. A very small sample below. A lot of this is normal, but a lot of it repeats constantly, and it’s all from Apple stuff—and what is “CSSM Exception” from Spotlight (mdworker) and why does it spew constantly? It makes looking for real issues far more difficult.

default	13:15:43.802085 -0800	gamed	GKClientProxy: clientForBundleID:
default	13:15:43.802128 -0800	gamed	GKClientProxy: updateIfRecentlyInstalled
default	13:15:43.850130 -0800	Core Sync	TCP Conn 0x60000018aeb0 event 1. err: 0
default	13:15:43.850159 -0800	Core Sync	TCP Conn 0x60000018aeb0 complete. fd: 31, err: 0
default	13:15:43.850295 -0800	Core Sync	TCP Conn 0x60000018aeb0 starting SSL negotiation
default	13:15:43.930815 -0800	Core Sync	TCP Conn 0x60000018aeb0 SSL Handshake DONE
default	13:15:46.033359 -0800	gamed	GKClientProxy: clientForBundleID:
default	13:15:46.033419 -0800	gamed	GKClientProxy: updateIfRecentlyInstalled
default	13:15:46.033859 -0800	gamed	GKClientProxy: clientForBundleID:
default	13:15:46.033882 -0800	gamed	GKClientProxy: updateIfRecentlyInstalled
default	13:15:46.100246 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 0, EGID: 0
default	13:15:46.931828 -0800	symptomsd	NDFSM: auto bug capture is administratively OFF, ignoring symptom with key: 421890
default	13:15:47.537797 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 0, EGID: 0
default	13:15:47.538055 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 20, EGID: 20
default	13:15:47.538475 -0800	opendirectoryd	Client: , UID: 501, EUID: 501, GID: 20, EGID: 20
default	13:15:47.546837 -0800	opendirectoryd	Client: , UID: 501, EUID: 501, GID: 20, EGID: 20
default	13:15:51.187559 -0800	CommCenter	#watchdog #I Callback Watchdog: checkin 328
default	13:15:51.187636 -0800	CommCenter	#watchdog #I Server Watchdog: checkin 328
default	13:15:56.345479 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 0, EGID: 0
default	13:15:56.345596 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 0, EGID: 0
default	13:15:56.345729 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 20, EGID: 20
default	13:15:56.345783 -0800	opendirectoryd	Client: , UID: 0, EUID: 0, GID: 89, EGID: 89
default	13:15:56.346063 -0800	opendirectoryd	Client: , UID: 501, EUID: 501, GID: 20, EGID: 20
default	13:15:56.346153 -0800	opendirectoryd	Client: , UID: 89, EUID: 89, GID: 89, EGID: 89
default	13:15:56.355784 -0800	opendirectoryd	Client: , UID: 89, EUID: 89, GID: 89, EGID: 89
default	13:15:56.355894 -0800	opendirectoryd	Client: , UID: 501, EUID: 501, GID: 20, EGID: 20
default	13:15:56.359875 -0800	opendirectoryd	Client: , UID: 89, EUID: 89, GID: 89, EGID: 89
default	13:15:56.359905 -0800	opendirectoryd	Client: , UID: 501, EUID: 501, GID: 20, EGID: 20
default	13:15:56.379043 -0800	mdworker	subsystem: com.apple.securityd, category: csresource, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.379083 -0800	mdworker	subsystem: com.apple.securityd, category: csresource, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.379258 -0800	mdworker	subsystem: com.apple.securityd, category: dirval, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.379300 -0800	mdworker	subsystem: com.apple.securityd, category: dirval, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.379560 -0800	mdworker	subsystem: com.apple.securityd, category: unixio, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.379644 -0800	mdworker	subsystem: com.apple.securityd, category: unixio, enable_level: 0, persist_level: 0,
default_ttl: 0, info_ttl: 0, debug_ttl: 0, generate_symptoms: 0, enable_oversize: 0, privacy_setting: 2, enable_private_data: 0
default	13:15:56.612652 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612654 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612700 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612699 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612758 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612758 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612797 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.612797 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.616884 -0800	trustd	cert[2]: AnchorTrusted =(leaf)[force]> 0
default	13:15:56.616921 -0800	trustd	cert[2]: AnchorTrusted =(leaf)[force]> 0
default	13:15:56.618896 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.618909 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.618945 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.618968 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.619007 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.619010 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.619046 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.619046 -0800	mdworker	CSSM Exception: -2147411889 CSSMERR_CL_UNKNOWN_TAG
default	13:15:56.625643 -0800	trustd	cert[2]: AnchorTrusted =(leaf)[force]> 0
default	13:15:56.625651 -0800	trustd	cert[2]: AnchorTrusted =(leaf)[force]> 0

Mike writes:

MacOS X, I will remain on Yosemite, has always been full of bloat and it is deteriorating fast and steadily.

SIP making it even harder to find the necessary ticks and switches to be halfway back in control of your own machine.

With 16 GB RAM in my MacBook Pro the system eats 25-30% of RAM after a cold boot just idling in Finder.

Deleting or unloading all those services is in my eyes a must. There are daemons active that are meant for features that aren't even supported on my hardware. Airdrop, anything iphoney, like handoff, cloud-nonsense: the lists just keeps growing.

After going through all these launchctl agents and and demons (sic!) – that are sorely lacking in documentation; just to add insult to injury – I am now happy about a system that 'just' eats away 10-15% RAM at idle. That is about half of both RAM-sticks the machine came with.

Of course you have to avoid Mail.app and Safari, Contacts and just about anything that Apple chose to bundle. They are mostly useless or downright dangerous for my data.

Then try to look at what cfprefsd does, writing plist files into your user preferences directory every minute for things like Dock and Finder that only are in use. Not changing anything in there respective preferences. coresymbolicationd's data directory, coreduetd's /var folder (Even when it was never used, never was supported on my machine, it managed to accumulate 5 GB of worthless data when I found out about it.) The always growing out of hand and often self.corrupting cache-files, etc. The only example I know of that actually got better from 10.9 to 10.10 was /private/var/log/asl keeping itself in manageable and sensible proportions in Yosemite. Keeping in mind that there are now even more examples of these crazy writes to disk one keeps wondering if this is equally detrimental for ssd's as the recent Firefox-session-store 'scandal'? Once set up I rarely change prefs in my apps. And yet. Cfprefsd writes 10 GB a day 'for me' and this is untenable _and_ untameable.

So I keep asking myself and now I am asking you: For how long will you remain in this sinking boat and swallow all those salty drinks Apple keeps serving us?

Come WWDC security updates for Yosemite will stop. If they announce a single new feature instead of a very, very large list of grievances fixed along with an apology for all those crazy stupid design decisions of the past years (starting with .DS_Store), then I will jump ship.

MPG: the natives are getting restless.

Constant writes to disk are a source of wear and tear on SSDs; the flash cells on SSDs can be written so many times. Writing 1 byte or ten is the same as writing an entire 512K block (of flash memory). It is perhaps an argument for never booting off an expensive SSD, just to avoid the constant write-harrassment of the SSD by the OS.

USB-C Dock for MacBook

4 USB3 ports, 1 USB-C port, SD card reader, gigabit ethernet, audio ports, HDMK 4K port!

Safari Seems Less Stable in macOS 10.12.2

See my Mac wish list.

Safari has been crashing more lately, and I suspect macOS 10.12.2 is involved, though it might just be that macOS Sierra is the turd it has felt like from the start.

Today, I just wanted to print one page from a web site—it’s not asking a whole lot.

Process:               Safari [30427]
Path:                  /Applications/Safari.app/Contents/MacOS/Safari
Identifier:            com.apple.Safari
Version:               10.0.2 (12602.3.12.0.1)
Build Info:            WebBrowser-7602003012000001~2
Code Type:             X86-64 (Native)
Parent Process:        ??? [1]
Responsible:           Safari [30427]
User ID:               501
Date/Time:             2017-01-10 08:43:53.862 -0800
OS Version:            Mac OS X 10.12.2 (16C68)
Report Version:        12
Anonymous UUID:        3A1B3248-8914-452A-A38E-AF944F6FE8CA
Sleep/Wake UUID:       652ED10F-8234-48DA-A5DA-00E009C4C044
Time Awake Since Boot: 26000 seconds
Time Since Wake:       3300 seconds
System Integrity Protection: enabled
Crashed Thread:        33
Exception Type:        EXC_BAD_ACCESS (SIGSEGV)
Exception Codes:       KERN_INVALID_ADDRESS at 0x0000000000000018
Exception Note:        EXC_CORPSE_NOTIFY
Termination Signal:    Segmentation fault: 11
Termination Reason: Namespace SIGNAL, Code 0xb
Terminating Process: exc handler [0] Thread 33 Crashed: 0 libobjc.A.dylib 0x00007fffbcc32b5d objc_msgSend + 29 1 com.apple.AppKit 0x00007fffa5bb3aed -[NSView(NSLayerKitGlue) _drawViewBackingLayer:inContext:drawingHandler:] + 1717 2 com.apple.AppKit 0x00007fffa5bb3432 -[NSView(NSLayerKitGlue) drawLayer:inContext:] + 80 3 com.apple.QuartzCore 0x00007fffadc76314 CABackingStoreUpdate_ + 3740 4 com.apple.QuartzCore 0x00007fffadd93464 ___ZN2CA5Layer8display_Ev_block_invoke + 75 5 com.apple.QuartzCore 0x00007fffadd930c5 CA::Layer::display_() + 1803 6 com.apple.AppKit 0x00007fffa5bb22ca _NSBackingLayerDisplay + 577 7 com.apple.AppKit 0x00007fffa5ba5531 -[_NSViewBackingLayer display] + 885 8 com.apple.QuartzCore 0x00007fffadd86f6e CA::Layer::display_if_needed(CA::Transaction*) + 572 9 com.apple.QuartzCore 0x00007fffadd87099 CA::Layer::layout_and_display_if_needed(CA::Transaction*) + 35 10 com.apple.QuartzCore 0x00007fffadd7c878 CA::Context::commit_transaction(CA::Transaction*) + 280 11 com.apple.QuartzCore 0x00007fffadc73631 CA::Transaction::commit() + 475 12 com.apple.QuartzCore 0x00007fffadc73aab CA::Transaction::release_thread(void*) + 589 13 libsystem_pthread.dylib 0x00007fffbd73b50f _pthread_tsd_cleanup + 544 14 libsystem_pthread.dylib 0x00007fffbd73b249 _pthread_exit + 152 15 libsystem_pthread.dylib 0x00007fffbd73bda8 pthread_exit + 30 16 com.apple.Foundation 0x00007fffa99d2dd7 +[NSThread exit] + 11 17 com.apple.Foundation 0x00007fffa9967c81 __NSThread__start__ + 1263 18 libsystem_pthread.dylib 0x00007fffbd739aab _pthread_body + 180 19 libsystem_pthread.dylib 0x00007fffbd7399f7 _pthread_start + 286 20 libsystem_pthread.dylib 0x00007fffbd7391fd thread_start + 13
Cycling

The GPU remains a Science Fair Project

See my Mac wish list.

Update 10 Jan: Adobe contacted me (well, I sent a link to this post to Adobe!), and they are going to look into the crash I'm seeing. I have provided an actions file and script to drive it. Now if only they would also agree to do something to address the GPU scaling headache.

Three+ years ago, the 2013 Mac Pro was released by Apple with graphics drivers rife with bugs. It took 6 months to get the drivers to a usable state. Adobe actually added GPU support for sharpening, then had to take that support out. What I never understood is how Adobe could ship those code changes without proper testing, the proof of that being self evident in the undoing.

Graphics driver bugs caused by half-baked Apple code are one thing—and I do pin the lion’s share of the blame on Apple. [It is my understanding that AMD engineers actually developed the drivers for the Mac Pro, but it’s Apple’s responsibility to ensure quality control.]

GPU problems with Adobe Camera Raw in Photoshop CC 2017,
macOS 10.12.2

But graphics driver bugs aside, Adobe could at least address what it can, that is, Adobe fails to fix severe usability bugs that result from enabling the GPU—for over two years now. Adobe was/is well aware of this issue, because I communicated directly, the issue was acknowledged as “hard to fix”. But even though I suggested minimally invasive user interface changes that could at least sidestep the scaling problem, this was not done either. Well, what is the point of a GPU if it results in unreliable operation of any kind?

And now GPU bugs persist to this day and now have taken a turn for the worse.

With the release of macOS 10.12 Sierra and Adobe Photoshop CC 2017, Photoshop has become more unstable than it has been in years. My 2013 Mac Pro, 2013 and 2015 and 2016 MacBook Pro and iMac 5K all have shown this GPU problem, as shown. Not always, but at launch or other times, and often enough to be disturbing.

Just today, something hard-crashed my Mac Pro (hard power off required), an extremely infrequent event (months without such a crash)—I was using Photoshop/ACR. I suspect the crash was a GPU driver crash in kernel space.

I don’t know where the fault lies (Apple or Adobe), but it is pathetic that one of my test suites (Lens Filters) cannot be completed on any Macs reliably, and never can run successfully on 2 of the 5 Macs.

Chad M writes:

I just upgraded my 2013 Mac Pro to Sierra over the holidays while I had down time.

I’m using an AJA LHI through ThunderboltBox and Final Cut X 10.3.1. I am getting kernel panics now (must be around a dozen).

I’m suspecting a driver issue of some kind – and your GPU thoughts have me leaning to look more in that direction. Maybe the next OS update will help. I should never have updated. So frustrating!!

DIGLLOYD: frustrating indeed for professionals trying to get work done. A kernel panic is of the same bug level as my 'hard freeze' I reported: something nasty happening at a level that kills the operating system.

Martin writes:

I read all your articles, and especially the more nerdy stuff about apple. Normally most of the people writing about Apple have nearly no clue about this more nerdy in deep stuff.
but I have to say something about your GPU disillusionment.

I’m into this GPU stuff since the late 90 doing 3D visualization and animation with extremely special and expensive cards, GPU acceleration before it was on the mass market.

And people like me know all these weird problems you experience.

It may sound bold and offensive but the 2013 MacPro with AMD GPUs and Adobe is the worst possible scenario. AMD drivers are known from the beginning as not suitable for professional work, besides they always had this FireGL pro line.

When you’re lucky that the drivers are stable, you will somehow experience display or rendering (calculation errors), especially working with fine graining values or extremes.
This is the reason nearly every pro in the 3D business or doing extreme compositing / video stuff is equipped with NVidia cards. The bad thing, the quality of their drivers is declining. In most of the pro software products is CUDA better integrated, integrated at all, more stable and a lot faster.

AMD looks always sweet on the spec sheet but they never deliver these speeds or stability. And the last 20 years there was never a stable and bug-free OpenGL driver from ATI/AMD, they are known for bad OpenGL.

The bad stuff, a lot of Adobes GPU implementations are also founded on OpenGL, like Lightroom's develop module - ADOBE confirmed this official that LR uses OpenGL. So every Mac now with a DGPU has an AMD GPU. Adobes Mercury Engine also relies on a lot of NVIDIA stuff - Premiere renders x times faster on a GeForce — is there is now other problem - ADOBE’s GPU implementations are really bad - MEDIAENCODER loves to show it uses CUDA but renders on your slow CPU.

But there are other software companies who have working GPU code - like Phase One. Run Capture One Pro on an 5 year old Intel i5 PC with a cheap GTX 970 — this is so much faster than Lightroom on the most badass Mac or PC money can buy. But Phase One also suffers from bad AMD drivers on Mac Mac and PC—just take a look and their user forums, a lot of people mourn that the software is so slow on their newest hardware.

At the end it relies on the software and Apple is constantly kicking us pros into our balls with their decision not to use NVIDIA.

MPG: the survey at BareFeats.com shows that 80% want NVIDIA cards, not AMD.

NuGard KX Case for iPhones and iPads
Outstanding protection against drops and impact!
Excellent grip for wet hands, cycling, etc!

2.5K or 4K or 5K Display for Image Editing and Viewing?

See my Mac wish list.

This essay at diglloyd.com goes into all the considerations in choosing a display: 2.5K or 4K or 5K Display for Image Editing and Viewing?. It applies equally to Windows users also.

See also Can a 2016 MacBook Pro support an 8K display?.

OWC ThunderBay 4 20TB RAID-5 $1299!
4TB to 40TB, configure single drives or as RAID-5, RAID-0, RAID-10.
Now up to a whopping 40 Terabytes

Crapware in macOS

See my Mac wish list.

I just thought that screen shot below captures the hubris run amok at Apple: the disregard for users that manifests itself at every level of the software and hardware product stack.

Crapware that I cannot remove is a dissatisfier (applies to iOS also), whether it is Chess.app or Garbage Band (oops, did I mis-spell it?) or the other Apple apps I never use and never will use.

Chess.app is a required application in macOS Sierra

On a related note, I have turned Siri OFF. The arrogance of Apple running the Siri process anyway is part and parcel of the attitudinal problem.

macOS: Siri runs even when off

When Siri is off, it’s still running. Any attempt to disable it in the plist is rejected by special security privileges that cannot be bypassed.

macOS: Siri runs even when off and cannot be disabled, even in the plist

I thought that use launchctl might work, but it doesn’t work to unload it.

sudo launchctl unload /System/Library/LaunchAgents/com.apple.Siri.plist
/System/Library/LaunchAgents/com.apple.Siri.plist: Could not find specified service

See also:

Greg M writes:

macOS: Siri runs even when off

I enjoy your comments about how bad the software Apple produces.

I opened up Contacts this morning and got this dialog box. All I did was open up Contacts, did no operations within.

MPG: sheer genius.

 

ThunderBay 4 - The Speed To Create. The Capacity To Dream.
Storage Wishlist…

Can a 2016 MacBook Pro support an 8K display?

See my Mac wish list.

The Dell UltraSharp 32 Ultra HD 8K Display will hit the market in early spring, raising the question of whether any Macs can drive that glorious 33.2 megapixels.

The GPU power along with the bandwidth required to drive four 4K displays is just emerging, but the 2016 MacBook Pro looks like it might be capable enough.

Apple specifications for the 2016 MacBook Pro state:

Up to two displays with 5120-by-2880 resolution at 60Hz at over a billion colors
Up to four displays with 4096-by-2304 resolution at 60Hz at over a billion colors

A 4K display is 3840 X 2160; thus four displays at 4096 X 2304 offer 13.7% more bandwith than required to drive an 8K display. Of course, an 8K display would have to be the only display, unless a 2017 Mac Pro arrives that can drive six or eight 4K displays. But with a MacBook Pro, presumably the internal display remains usable, so a dual display system looks possible.

Using two Thunderbolt 3 cables, it should be possible to drive an 8K display on the 2016 MacBook Pro in 10 bit color*, at least in terms of the required bandwidth, using a trick similar to how a 5K display was possible in previous Macs: drive each half of the display with one of the Thunderbolt cables, that is, Multi Stream Transport. The same trick might work for 8K. Perhaps Apple or Dell will weigh in on this possibility soon.

* The Apple specifications indicate “billions” which means 10-bit color.

Potential for an iMac 8K

Since a 32-inch 8K panel now exists, and the 2016 MacBook Pro already has the bandwidth to drive an 8K display, an iMac 8K is technically possible with today’s technology.

An iMac 8K could use a custom graphics solution to push those 32 megapixels, just as the iMac 5K does. There is no need for wait for standards to evolve to support external 8K support; it can be done internally with a proprietary solution. The main issue is cooling, but Apple should be able to engineer an appropriate GPU cooling system. And since the 2016 MacBook Pro is an even greater cooling challenge, and it can drive four 4K displays, it hardly seems like an iMac 8K should have any barriers to supporting an 8K built-in display.

The Dell UltraSharp 32 Ultra HD 8K Display will apparently sell for $4999. Even assuming that Apple could use its considerable volume to get that price down to $2000 or so, it would make an iMac 8K with the necessary GPU power something hitting the $6K to $8K range in price. So even though it appears to be ideal for an 32" iMac 8K, cost is likely to push off such an iMac. Then again, Apple has surprised us in the past, and could do so again, particularly with its failure to innovate for several years now.

ThunderBay 4 - The Speed To Create. The Capacity To Dream.
Storage Wishlist…

Dell UltraSharp 32 Ultra HD 8K Display Suggests that an iMac 8K is Technically Feasible

See my Mac wish list.

See also Too-High Pixel Density on 5K and 8K Displays Impedes Image Assessment.

While I saw firsthand the LG 5K display for the 2016 MacBook Pro, unfortunately I did not chance upon the Dell 8K 32-inch display which was announced at CES.

7680 X 4320 pixels = 33.2 megapixes in 32" form factor, aspect ratio 1.77:1

A 4K display is like tiling together four HD displays and thus requires 4X the bandwidth of a conventional 1920 X 1080 display. An 8K display is like tiling together four 4K displays for a whopping 33.2 megapixels (7680 X 4320).

A a full resolution Nikon D810 image is 7360 X 4912, so its entire width fits with room to spare on an 8K display, although the aspect ratio is too tall to fit vertically. The pixel density of an 8K display has to be an awesome viewing experience, but high pixel density is problematic for assessing image sharpnesss.

The $4,999 UP3218K is slated to go on sale on March 23rd. It has a refresh rate of 60 Hz and will work with zero Macs and only a few select PCs with specialized video hardware, and even then GPU performance will need some time to catch up. The display supports 100 percent Adobe RGB, sRGB, and Rec709 color gamuts.

HDMI 2.1 (announced at CES) supports 8K, but Apple doesn’t even support HDMI 1.3 yet, so external 8K displays are likely to take a few years to become more than a special rarity.

However, at least in bandwith terms, 8K might already be possible; see: Can a 2016 MacBook Pro support an 8K display?

Potential for an iMac 8K

Since a 32-inch panel now exists, it seems ideal for an iMac 8K which could use a custom graphics solution to push those 32 megapixels—no need for wait for standards to evolve to support external 8K support; it can be done internally just as with the iMac 5K.

The Dell display will apparently sell for $4999. Even assuming that Apple could use its considerable volume to get that price down to $2000 or so, it would make an iMac 5K with the necessary GPU power something hitting the $6K to $8K range in price. So even though it appears to be ideal for an iMac design, pricing is an obstacle that could put off an iMac 8K for a year or more.

4TB Internal SSD
for 2013 Mac Pro
Free how-to videos and tools included, 3-year warranty

LG 5K Display for 2016 MacBook Pro

See my Mac wish list.

See also Too-High Pixel Density on 5K and 8K Displays Impedes Image Assessment.

Yes, you want one. It looks fantastic in person, shot below from CES. But at present it works only on the 2016 MacBook Pro, via a Thunderbolt 3 cable, as shown below. OTOH, for about $1700 you can get a 5K display from Apple with a free computer.

I am NOT recommending it for work where evaluating image sharpness or very fine tweaks is needed—that’s why I still use the NEC PA302W for my image evaluation—the pixel density is way too high for that type of detail work. But like the iMac 5K viewing experience, the LG 5K is aweseome for viewing images, with its 14.7 megapixel display.

I did not chance upon the Dell 8K 32-inch display which was announced at CES, but a 32-inch panel now exists, it seems ideal for an iMac 8K which could use a custom graphics solution to push those 32 megapixels. An 8K display requires the bandwidth of four 4K displays and isn’t going to fly with any Mac (yet).

 
LG 5K display for 2016 MacBook Pro
__METADATA__

OWC Envoy Pro Thunderbolt 3 2TB SSD

I’m in SSD heaven!

I just saw a production version of the OWC Envoy Pro Thunderbolt 3 2TB SSD. About twice the size of the current OWC Envoy Pro EX (my go-to drive for travel backup and other on-the-go purposes), the new OWC Envoy Pro nonetheless fits in the palm of my hand. The larger size is all for the better, because it allows for very high capacity, cooling fins for sustained ultra high performance and a robust aluminum case.

  • 2TB capacity with a form factor that implies a 4TB version should emerge.
  • Thunderbolt 3.
  • Tested with DiskTester: this external SSD not only rocks on small transfers (about as fast as anything I’ve ever tested), it hits 1300 MiB/sec on writes and 2000 MiB/sec reads. Externally. That performance blows away the 1TB internal SSD in my 2013 Mac Pro, at twice the capacity. Practically speaking, more speed is of no practical concern for anything but the most specialized or obscure situation.
  • Pricing TBD.

I see this drive as a breakthrough device: 2TB (and maybe later, 4TB) capacity, desktop grade SSD performance, outstanding portability. I can’t think of anything ever before that combines all these attributes, with no downsides at all. Many people could carry all their Stuff on one of these drives and never look back. The possible catch: is it backward compatible with Thunderbolt 2 via the Apple Thunderbolt 3 to Thunderbolt 2 adapter? It should be, but this is to be confirmed.

The OWC Envoy Pro Thunderbolt 3 2TB SSD should be available for purchase “before NAB” according to OWC (that means April or so).

2TB OWC Envoy Pro Thunderbolt 3 SSD attached to 2016 MacBook Pro
2TB OWC Envoy Pro Thunderbolt 3 SSD attached to 2016 MacBook Pro

A Powerful Argument for a Small Laptop

I grabbed dinner at the airport on the way to CES. I asked Andalé Taquería if El Presidente Burrito trumped El Supremo, but they feigned ignorance. Personally I think both are quite a mélange of ingredients.

Anyway, I’m sitting on a Virgin America 737 or whatever this narrow-bodied thing is, and I’m back in cattle class. There is barely (barely!) enough room to open my 15-inch MacBook Pro, with the likelihood of it being damaged if the passenger ahead of me reclines his seat—so I politely asked for a warning if he were to be so inclined (double entendre intended).

If I flew regularly, I’d buy a MacBook just for airline travel. In the “old days” a 15-inch laptop was not a problem. On Virgin America, it’s a bad choice size in this godawful sardine can. But there were few options left when I bought my ticket.

Update: the return flight on Jet Blue was vastly better for leg space. And for $25 I sat in seat 4D where I could actually stretch my legs out, and my pack would fit under the seat ahead of my. Night and day better vs Virgin America.

Cycling

OWC “DEC” for 2016 MacBook Pro

Update 4 Jan: I’ve seen the prototype of the OWC DEC first hand. The DEC bolts onto the bottom of the MacBook Pro after removing the bottom shell of the case. The result is a seamless integration with all the key ports I need (gigabit ethernet, USB-A 3.1, SD card slot), delivering what feels in the hand very much like the 13" 2012 MacBook Pro in thickness (but in 15" size).

The convenience of this form factor should not be underestimated by any power user: no dongles, the classic ports right built-in, and from what I understand—potentially configurable for all sorts of things like extra battery power, gobs of SSD storage, possibly even hard drives. In short, a really powerful platform could make a MacBook Pro into a portable desktop grade machine, particularly if/when Apple introduces the Kaby Lake chipset to the MacBook Pro line.

Another factor is data safety: by having a 2nd (or 3rd drive) in the DEC, pros that cannot lose data can backup the internal drive conveniently (think video footage that cannot be re-shot and whose loss could break a reputation after ruining that wedding or whatever)—remember the soldered-on SSD is a lot of downtime if the MBP goes down. Of course, an external backup should also be used, and there the OWC Envoy Pro is ideal.

The conceptual design is just about limitless, affording possibilities for huge SSD storage (my guess is that up to 8TB will be possible before long, OWC says up to 4TB initially), extra battery power should be configurable too—how about 20 or 30 hours of battery life? That too is my guess but the volume of the DEC should be able support something like that and/or a mix of SSD and battery boost. Other potentialities are there too. I don’t see anything that precludes all sorts of terrific expansion capabilities, all by extending the thicknesss of the MacBook Pro to a total height little different from what used to be the standard thickness.

The OWC DEC will be available for purchase in Spring 2017, according to OWC.

Form factor of the OWC DEC is little different from what used to be the standard MacBook Pro thickness. Connector at right (prototype, final will be more elegant) couples the Thunderbolt 3 port of the MacBook Pro to the DEC.

Later units (not prototypes) will have an SD card slot and USB port on the right side, and other ports (gigabit ethernet, more USB ports) on left side. OWC hints at other potential ports as well.

Prototype of OWC DEC attached to 15" 2016 MacBook Pro
Production version to polish fit and finish, plus space gray or silver options

Sleek enough to travel with because the form factor is little different from what used to be the standard MacBook Pro thickness.

OWC Makes MacBooks ‘Pro’ Again With Game-Changing OWC DEC

Tuesday, January 3rd, 2017

OWC today announced the OWC DEC, the first-ever expansion solution designed specifically for the 2016 Apple MacBook Pro. The OWC DEC is the finishing touch to a MacBook Pro, allowing users to take their laptops to the next level, extending the life and the functionality of the already-exceptional 2016 MacBook Pro. The DEC is the latest in OWC’s long line of upgrades for nearly every Mac made and continues the company’s legacy of providing the best solutions for Mac owners.

The OWC DEC attaches completely flush to the bottom of the 2016 MacBook Pro, providing additional flash storage and connectivity, ultimately increasing the performance of MacBook Pros. With the OWC DEC, users will get an enhancement path for their 2016 MacBook Pro and MacBook Pro with Touch Bar to keep their system upgradeable for the long term in a clean and integrated fashion. On top of the core functional benefits, the OWC DEC also touts a sleek design. When installed, the OWC DEC and MacBook Pro will be as thin as a 2012 MacBook Pro, allowing this advanced solution to retain the attractive light weight design that users favor.

This patented solution has a range of features, which include:

  • Up to 4TB of additional Flash/SSD storage (for a maximum of 6 TB, including factory capacity)
  • SD Card Slot/Multi-Media card slot
  • USB 3 Type A Ports for standard USB cabled devices
  • Gigabit Ethernet
  • And other features to be announced at a later date

Larry O’Connor, OWC Founder and CEO, will personally showcase the OWC DEC prototype in Las Vegas this week. The OWC DEC enables the new MacBook Pro to be updated and expanded in order to operate at peak performance and ultimately be relevant for years longer than may otherwise be expected.

“We’ve been developing this concept for over three years and feel now is the perfect time for this ground-breaking product,” said O’Connor. “Storage space is always at a premium on a laptop and without the potential to upgrade the factory drive, the OWC DEC brings that option back in addition to cleanly supporting the gear we all rely on.”

The release of the OWC DEC comes on the heels of a significant year of new product announcements from OWC, including the OWC Thunderbolt 3 Dock, OWC USB-C Dock, Aura SSD and the newly released Mercury Elite Pro Dual mini and Mercury Elite Pro Dual.

OWC will ship the highly-anticipated OWC DEC in Spring 2017. OWC Upgrades are available from MacSales.com and other fine retailer and e-tailers in the U.S. and around the world. For more information, please visit www.owcdigital.com/DEC.

Envoy Pro mini - In Motion There Exists Great Potential
SSD Wishlist…

Desktop CPU Performance at a Standstill?

As I wrote in Computers Are not Getting Faster in a Meaningful Way, GPU is Half-Baked Tech, Too Many Software Developers Suck, computer performance has hardly changed in 3+ years. All while most software continues to make poor use of CPU cores, so a double whammy from two different directions.

ArsTechnica comments on the CPU situation in Intel Core i7-7700K Kaby Lake review: Is the desktop CPU dead?:

With identical performance to Skylake, Intel brings desktop performance to a standstill.

The Intel Core i7-7700K is what happens when a chip company stops trying. The i7-7700K is the first desktop Intel chip in brave new post-"tick-tock" world—which means that instead of major improvements to architecture, process, and instructions per clock (IPC), we get slightly higher clock speeds and a way to decode DRM-laden 4K streaming video. Huzzah.

Update 04 Jan: are at least a few reasons why the Intel 'Kaby Lake' release is significant:

  • Improved graphics performance.
  • My understanding is that the Kaby Lake 'H' series supports 32GB memory, thus making a MacBook Pro with 32GB of DR 23000 DRAM possible. But whether the power draw is viable on a laptop is unclear (meaning what we could expect from Apple, given the rationalizations seen with the Nov 2016 MacBook Pro).
  • The i7-7920HQ 3.1 GHz (turbo boost to 4.1 GHz, 4 real CPU cores) might be suitable for a MacBook Pro.
  • The i7-7700K 4.2 GHz (turbo boost 4.5 GHz) shoudl be suitable for an iMac. This perhaps is the “standstill” point—that’s only 5% faster than the 4.0 GHz iMac 5K that sits on my desk today—at the cost of a 95 watt TPD.

Even if CPU performance is stuck in 3rd gear for now, all is not lost. In particular, Apple could go back to a real “pro” desktop by making it a big box again, with features like this. I bet that it would be a best-seller Mac Pro:

  • A choice of 4 to 18 CPU cores, or whatever the limit is currently. Even better: dual CPUs again.
  • 8 memory slots accepting up to 256GB. Stretch goal: 12 or 16 memory slots for up to 384GB.
  • 12 Thunderbolt 3 ports, split across 3 busses.
  • Support for three external 5K displays.
  • Single outrageously fast GPU.
  • Two 16x PCIe slots for optional 2nd or 3rd GPU and/or extra PCIe SSD, etc.
  • Internal SSD options of 1TB or 2TB or 4TB, option for 2nd SSD with same capacity options, running at 3 GB/sec.
  • Space for two internal hard drives, thus allowing 20TB internally in addition to the SSDs.
  • Nice to have: 4 USB3 ports to keep connectivity hassles down with useful legacy devices (mouse, keyboard, camera card readers, USB3 drives, etc).

Why not? Apple could re-establish itself as a serious player for high-end performance. Make it a box, but very quiet box, a big cylinder if some nitwit demands that as the tradeoff, but make it robust.

BareFeats.com has a a very interesting survey highly relevant to the above points: 63% prefer a tower form factor, 62% prefer six or more CPU cores, 34% prefer 128GB or more memory, few users want AMD GPUs (NVIDIA strongly preferred), 73% want support for at least two displays, 39% want 5K support, 43% want at least 4TB of internal storage. And so on. Clearly there is a desire for a high end machine.

OWC Thunderbolt 2 Dock
Review of Thunderbolt 2 Dock

Apple Core Rot: PDF Support

Is this incompetence, disdain, or all-out contempt for developers and users? MPG tends to favor the “incompetence and mismanagement” theory, but those generally require some degree of the other two, that is, to allow it to be approved and ship to customers.

Apple Insider: Apple's changes to macOS PDF handling stymie third-party developers, cause data loss

Apple's changes to how it handles PDFs in macOS Sierra are causing problems with third-party utilities, with the most profound issue potentially causing the removal of an optical character recognition layer from user's files.

Adam Engst from long-time Mac journal TidBITS noted that while problems were widely publicized at Sierra's launch with ScanSnap scanner software, other issues have persisted after two updates, and in some cases gotten worse.
...

According to reports collated by Engst, and confirmed by AppleInsider, Apple has re-written the PDFKit framework in macOS 10.12 Sierra, and implemented a common core with iOS and macOS. However, this has caused some serious issues with software that relies on Apple's PDFKit.
...

MPG: data loss supports the “disdain and contempt” theory, but does not rule out sheer incompetence.

Note the “common core” thing—a very dangerous trend for future APIs in terms of reliability, compatibility and data integrity particularly since Apple seems to have no idea what unit testing is.

Whose data of any kind is safe when Apple has no qualms about rewriting APIs that damage user files? Such actions show a profound disrespect for users, although leaving RAID users hung out to dry is arguably far worse.

On the Apple Core Rot topic, see one more view in Apple’s 2016 in review.

Monitoring Most Everything: iStat Menus

 
iStatMenus: CPU and Network Status

Last week I reported on cleaning out dust to forestall component failure.

A program like iStatMenus can help figure out if a Mac is overheating; take a baseline when new or after cleaning and periodically check those temperatures against temperatures some months later (assuming the same room temperature). Or, if the Mac’s fans are running too loudly or too frequently under no apparent load situation, check out the temperature of the components.

BJango Software also offers iStat for Mac. Maybe it can run locally only, but it seems to be designed for remote monitoring via iStat Server.

While iStatMenus has a wealth of information, and has a smooth and polished interface—it’s a terrific tool—there are two things that bother me which are peculiar to my own setup.

  • I would prefer to have all this information in one window for two reasons: (1) see everything at once, and (2) locate the window on whichever display I want to, (3) unclutter y menu bar. Well, it’s iStatMenus for a reason so that is by design. Still I would like this info in a window which would reside on my 4K display for the same reason all my palettes in Photoshop are off my main screen.
  • With my dual display setup, the main screen is standard (non Retina) resolution. The very small type is not os easy to read on standard-res displays. The menu bar must be on that display (it is not a viable option to put a menu bar on the other display and/or to make the 4K 2nd display be the primary, for my own reasons). So I am stuck with small type that is unfriendly to my eyes and cannot be captured with retina quality in a screen shot either. The foregoing is inapplicable to my iMac 5K or to my MacBook Pro.

Below, all the sensor measurements on my 2013 Mac Pro.

iStatMenus: Sensor measurements on 8-core 3.3 GHz 2013 Mac Pro

Configuring the “Combined” menu

I like the combined menu because I can mouse over each area as submenus. If configured to use separate menus in the menu bar (each area of reporting having its own separate menu), then each menu has to be clicked on separately.

See my previous comments: a modal information display is not nearly to my linking as being able to see all this stuff in one larger status display.

iStatMenus: “Combined” menu with Memory submenu
Thunderbolt 3 Dock
Must-have expansion for 2016 MacBook Pro
Thunderbolt 3 • USB 3 • Gigabit Ethernet • 4K Support • Firewire 800 • Sound Ports

2016 MacBook Pro: Test Results vs 2015 MacBook Pro and 13" 2016 MacBook Pro + a Big Surprise

Mac wish list •  all 2016 MacBook pro models at B&H Photo • all 15" Apple MacBook Pro 2016 models •  all 13" Apple MacBook Pro 2016 models. MPG gets credit if you buy through those links.

Deal on top-spec 2015 MacBook Pro at OWC.

I encountered a stunning surprise when testing: the 13" 2016 MacBook Pro dual-core 2.4 GHz beat out all but one of the other Macs on my #1 most important Photoshop task. And I had to give the other Macs several tries to get better times; the times usually were slower.

That’s insane! But I ran and re-ran the test—same deal. So to paraphrase Mark Twain, there are lies, damn lies, and benchmarks. All that matters is how well a computer works for one’s own particular workflow.

This amazing performance is from the 13" 2.4 GHz / 16GB / 512GB non-touchbar model, whose Turbo Boost mode hits 3.4 GHz, which is only 0.1 GHz slower than the fastest processor Apple offers in the 13" model. But those faster base-clock-speed models require the touchbar, and the touchbar is a non-starter for me and others. Plus, the touchbar model has a significantly smaller battery and slotted (not soldered-on) SSD. I also found that the 13" model has the least trouble cooling itself, as judged by fan noise (not much even under load). Plus the 13" model does not slow down under some loads like the 15" model does.

While all of the machines showed some variability in test results (laptops in particular), multiple runs confirm the dominance of the 13" 2016 MacBook Pro in this real-world workflow challenge. The graph below takes the fastest time from each machine. Other tests show the 13" model to be be slower, but quite a feisty ankle biter on many tests. If anything, it shows that software plays a huge role in using or not using CPU cores efficiently.

2016 MacBook Pro vs other Macs: Photoshop filters

2016 MacBook Pro: Test Results Updated to include 2015 MacBook Pro and 13-inch 2016 MacBook Pro

Mac wish list •  all 2016 MacBook pro models at B&H Photo • all 15" Apple MacBook Pro 2016 models •  all 13" Apple MacBook Pro 2016 models. MPG gets credit if you buy through those links.

Deal on top-spec 2015 MacBook Pro at OWC.

2016 MacBook Pro: Photoshop Filters

2016 MacBook Pro vs other Macs: Photoshop filters
SSD Upgrade for MacBook Pro Retina
Internal SSD Wishlist…
SSD Upgrade for MacBook Pro Retina
Internal SSD Wishlist…

diglloyd.com | Terms of Use | PRIVACY POLICY
Contact | About Lloyd Chambers | Consulting | Photo Tours
Mailing Lists | RSS Feeds | Twitter | Retina Image Control
Copyright © 2008-2017 diglloyd Inc, all rights reserved.