Yes, the 2018 iPad Pro has aged well

I’m inclined to agree with much of this. The 2018 model came into my group of hardware and started running circles around my other devices, particularly my third generation Core i5s.

Right now the only good reason I see to upgrade are support for secondary displays and perhaps thunderbolt. Features I desire in my tablet but that aren’t worth the price of a new iPad when everything else is dandy.

Well, if you discount that the base 64GB of storage is getting less roomer, the battery is aging, and I’m sometimes tempted by the idea of an cellular capable iPad. Those are all interesting points, but still sufficient for the time being 😛

Signs of getting old

Random things that make me feel like I’m getting old: having seen entire generations of hardware make the transition between hot, sexy, powerful computers and being the oldest equipment this side of a dumpster.

Also, it’s probably a good thing that I don’t have room for a server rack in my closet.

Having spent part of my day studying network captures that made me ask if a network attached blender was involved, I’m reminded of my awesome router.

About a decade ago, after a stream of typical home routers and much rage, I decided to buy an RT-AC68U. At about $242 and change it was like pulling teeth bare handed, to pay that much for a router. But it quickly became worth every penny when it achieved the level of more reliable than our ISP, like literally the neighborhood had network outages more often than there has been need to reboot the thing.
I reckon, if the savings in pain and suffering haven’t caused it to pay for itself yet, the whole reliable long lived service life thing has. One of my favorite equations for equipment is cost divided by age, by which that router has cost me about $24 and change per year. By contrast, the $65-$75 routers it replaced were only lasting a few years before they either failed or failed to cope with the ever increasing number of devices connected. Ahh, there is such a thing as a good router.

Rimuru Restoration – With a screwdriver, I stab at thee

Despite it only being only a few years old, motherboards compatible with Rimuru’s processor are largely gone and needless to say, my AsRock isn’t re-obtainable as far as local resources go. That’s been the sad trend IMHO, that yes, desktops are still pretty modular, no the parts won’t be worth a fart tomorrow. But alas, that’s a different issue.

Deciding that the motherboard is the root of the problem based on my multimeter readings and the screwed up power behavior, I debated two courses of action: decommission Rimuru in favor of a laptop, as it was already expected to be my last desktop build; or attempt to fix things with replacement parts. The upside of the later is that it is the minimal cost option, the former that it’s the less likely to piss me off.

Rimuru is now rocking an Asus motherboard a generation forward. A small fortune and the better part of my day later, everything seems to be operational. Fortunately, re-activating Windows licenses purchased from Microsoft’s own store are still not too terrible to deal with motherboard replacement.

In the process, I’ve also decided to ditch the humongous air cooler and get a liquid cooler, cue kraken, stage left. When I originally designed Rimuru, I had considered liquid cooling and decided to stick with what I know. Well, I decided if I was going to be replacing a motherboard, there was going to be something a lot smaller hovering over the processor getting int he way of my hands, or I was going to drive a spike through the board. So, liquid cooler it be.

Now if there was just a solution for the raging headache ^_^.

Rimuru facing decommissioning

Before the holidays, I had the problem that Rimuru would power on in a brain dead state — fans would spin up, most buses would power their components, and so on. But it wouldn’t POST or reach BIOS. Just brain dead. The only way to turn it on or off was the cord and kill switch on the PSU, and trying to hold the power switch on the case would just act like a reset and then total brain death with running fans.

Tried all the good jumping off points: trying to boot off integrated graphics, reseating and walking RAM sticks, etc. Then I took my multimeter to the power supply’s ATX main and CPU power pins, and that seemed to be fine.
At that point the only thing I hadn’t reseated was the CPU, and I’d have to go to Micro Center for fresh thermal gloop anyway if I did that. So, I decided to drop it off and see what they could figure out. Today, I got my machine back. Reseated memory, POSTS. Well, I hoped their touch was more magical than mine. But no dice.
When I had asked, it was unknown if it fresh thermal paste was used, since the guy doing initial processing had tried reseating the CPU and swapping in another of the same model with no luck. Reseated the processor with fresh thermal gloop, and I’ve decided that I am taking a new policy on this. If the CPU cooler is connected and it don’t catch fire, I’m not effing remounting CPU coolers for this–it’s just to much of a pain in the ass in such cramped spaces, and my hands ain’t gettin’ tinier. If I ever build another desktop, there better not be a huge ass heatsinkage over the processor or I think I’m outsourcing to someone fitting that description.
Anyhow, the machine wouldn’t boot and was doing the same thing as when I dropped it off. Gave it a kick and numerous elevated heart rate notifications later while choking on my urge to go Incredible Hulk on the decommissioning, I eventually came back to try a few things.
Tried loading a single RAM channel up and with some fiddly, I got into the OS. To do so, I had to pull the front panel header and rub a nail clipper on the pins. Eventually, I put in the other RAM channel but it didn’t seem to make a difference what I did with the memory.
Depending on how I manually short the reset and power switch pins, the system either hangs as before, goes into brain death, or does a sorta reset and may boot or die.
Fiddling around, I noticed that USB ports may or may not have power. I.e., plug in a keyboard while in BIOS, walk it through the ports toggling num lock, cycle back and what the fuck there’s no power, and then again and now it works. Stuff like that. In some cases the keyboard would draw power enough to light the num lock LED but wouldn’t toggle, and then in some cases it would just be flashing all three LEDs at the top for a while and then go dead.
Anything resembling a “Proper” shutdown like the OS would do leaves the machine brain dead. Have to pull the power and fiddle pins trying to find the right fiddling and timing to boot. Trip the right way and the processor fan cranks and the case fan turns off, and it’s total death until power pull.
So at this point, I think a technician who’s able to test piece by piece and determine what is failed would be needed. That’s beyond my skill without a schematic, and the tiny as hell components to desolder and resolder would make the repair work beyond my ability even if I did have a schematic to work off.
Based on its behavior, I’m inclined to believe the power management chip is screwed up or something has gone awry with the path of power causing it to just “Leak” into systems that should remain unpowered until initialized properly.
To garner a second opinion, I think I may buy one of those PSU tester things to double check my power supply’s readings vs my multimeter. Which basically means, motherboard if it’s not the power supply. If that’s the case, given how much issue it is to get ahold of another motherboard of that model; I may opt to decommission  Rimuru.

One of the things I’ve been wondering for a while now is how the performance of macOS’s EXFAT driver is representative of its peers. It’s notably slower than what you would see in NT, but not so bad until you go from the sequential 1M to random 4K part of my choice benchmarks. Once you hit the randoms, it goes form “I wonder if that’s lack of optimization in the driver, or the I/O system design” to abysmal. But to be fair that is the worst performing metric anywhere, and I’m more interested in the sequential performance. 

Well, having a nice shiny (or should I say, mat?) Samsung T7 Shield that was on sale, I decided to do a little test cycle. EXFAT, FAT32, HFS+, and APFS. This drive is designated for Time Machine duty, so I have no need for it to remain on a interoperable file system.

Using AmorphousDiskMark 4.0.

EXFAT as formatted out of the box:
Test - Read MB/s Write MB/s
SEQ1MQD8 - 586.42 691.32
SEQ1MQD1 - 594.45 690.05
RND4KQD64 - 21.75 13.68
RND4KQD1 - 21.70 13.48
FAT32 as formatted MS-DOS (FAT32) from Disk Utility:
Test - Read MB/s Write MB/s
SEQ1MQD8 - 516.03 690.32
SEQ1MQD1 - 596.97 691.80
RND4KQD64 - 21.56 13.64
RND4KQD1 - 21.50 13.51
HFS+ as formatted Mac OS Extended (Case-sensitive, Journaled) from Disk Utility
Test - Read MB/s Write MB/s
SEQ1MQD8 - 612.39 820.77
SEQ1MQD1 - 578.25 691.00
RND4KQD64 - 120.48 55.44
RND4KQD1 - 18.33 14.70
APFS as formatted APFS (Case-sensitive), after converting from MBR to GPT from Disk Utility.
Test - Read MB/s Write MB/s
SEQ1MQD8 - 733.22 818.84
SEQ1MQD1 - 617.40 684.06
RND4KQD64 - 121.67 55.13
RND4KQD1 - 21.27 13.83

This makes me suspect the performance lossage is more to do with how optimized the FAT drivers are. I should really repeat this with one of my USB flash drives where the performance sucks to begin with, but I don’t want to spend all day on this :^o).

Migration to macOS has been relatively successful so far. Juggling work and dogs and the need to occasionally vegetate, it took a couple weeks to get Shion properly setup.

For me one of the key problems with switching between Mac and PC has been the great modifier shift. The annoying kind of things when I come home and start using Mac shortcuts on my PC, or go to work and start using PC shortcuts on my Mac.

As a work device, my issued MBP has mostly been a case of IDGAF in terms of PC vs Mac. Contemporary OS X and its successors-thus-far, are suitable BSD under the hood with GNU sprinkled on top that it’s basically a non-issue. On Windows, I would be using Windows Subsystem for Linux and SSH. On Mac, well it’s native enough unless it needs to be Linux ELFs. Like NT, it comes with some nice to have GUI software but most of what I care about can be found in the Terminal.

As a home device, I’m finding it fits quite nicely. It does the desktop things that better maintained Linux distributions and Windows systems do, and it provides most of the goodness I’d get out of running FreeBSD or Debian. More importantly I don’t find myself !@#$%ing mixing up the command and control keys ^_^.

Outside of Direct3D based games the majority of software that I care about is cross platform, often with GNU/Linux as the primary platform if one could be defined. So, basically everything I want to run either runs on unix, NT, and Mac systems already; or it’s tied to POSIX APIs longer than Linux and OS X have been around, or it’s unlikely to run on anything that doesn’t do Big Honking DirectX GPUs.

Thus: Rimuru’s intended mission profile is what it was chiefly built for. Playing video games, converting videos, and cursing those times when compiling on NT is a thing. Meanwhile Shion takes over the more secretarial domain of general productivity and desktop computing.

So far, one of the unexpectedly nice things about macOS: I can use my Bluetooth keyboard to wake my MacBook Air. No putzing required. I’d like to assume my desktop could pull that off with fiddling with the power management options for the front ports or motherboard root controller, and/or the Bluetooth USB dongle. But my relationship with NT and things USB/BT is one of pain and suffering, so I’m less inclined to putz with that.

Further iterating my Gateway Station concept, I’ve tossed the Anker 555 in the closet and hooked up a TS4 from CalDigit. Under macOS, I’m finding that this works flawlessly and resolves the “Well, if I just use a second cable for power” issue I observed with the hub.

And then there’s windows (>_<).

One reason that I opted to get the newer TS4 is its love for 10 Gbit/s ports versus the older TS3 Plus. Basically the ports are either rated for 10 Gbit/s USB or 40 Gbit/s TB. Another reason is because of the Thunderbolt 4 / USB 4 capability made me wonder if it would be both backwards and forwards compatible with my desktop.

Migrating All The Things ™ to the CalDigit appears to work well enough for my purposes. I’ve no need for the DisplayPort or Ethernet port on the desktop side. All the USB ports appear to function when plugged into my ASMedia controller, and the only issue observed is that USB drives won’t work. I am unsure if this is due to power negotiation, or drivers. Under macOS, I can basically plug any damn thing in  without issue. Part of me is tempted to swap out the USB 3.1 10 Gbit/s card for a Thunderbolt card, and part of me just does not want to know what the fuck Intel’s NT drivers for that are like ^_^.

When I decided to consider a dock, I decided if I was going to spend the big bucks, which my previous solution was meant to avoid, that I was going to make sure it was cable of being the heart of my desk setup. Such that it could be the Single Point of Truth in connectivity instead of just the break out. That way if the desktop side of the coin proved sufficient I could do that, and if not, I could retain the previous configuration for Gateway Station. Compared to the Windows issues my previous arrangement had, it’s been a bigger issue finding room for the dock on my desk.

Considering that I can use Rimuru’s front panel USB ports when storage drives are required, and have USB-C extension cables that could be routed to the spare port on the ASMedia card, I’m not particularly concerned about my observations so far. My goal with this transition was to have my laptop become the core for the non gaming stuff and relegate my desktop to being focused on gaming. Thus far that’s working. The test that remains is to determine how reliable this turns out on the NT side.

And to remember to turn off my speakers so they don’t fallback to Bluetooth pairing mode, if I leave it connected to Rimuru instead of Shion.