It is certainly interesting that deciding whether or not to kill some fleshy humans can be demonstrated to be circumscribed by the halting problem; but it's always a bit irksome to see another proof-of-limitiations-of-turing-complete-system that (either by omission, or in more optimistic cases directly) ignores the distinct possibility that humans are no more than turing complete.

Humans certainly are enormously capable at approximate solutions to brutally nasty problems(eg. computational linguistics vs. the average human toddler); but that is very different from a demonstration that, say, humans possess an Oracle, or are some sort of hypercomputational system, rather than simply being enormously good at hard-but-not-theoretically-intractable problems in certain areas.

In this instance it's especially galling because we've only been philosophizing about acceptable losses, 'just war', legitimate causus belli, 'proportionality', and whatnot for about as long as we've been chucking spears at one another. It's a pure commonplace that a mixture of overkill and underkill is an effectively certain outcome when you go to war. It is interesting that, in principle, kill/no-kill is subject to the halting problem; but has anyone (aside from sleazy assholes hyping 'smart' weapons) ever asserted that kill decisions would be anything but imprecise?

I don't doubt that China could. Half the fun of being a nation state is that you can do all kinds of stuff with no more risk than a stern letter from the WTO. That said, taking action against a private company, selling at a loss out of its own pocket, would likely play differently than taking action against a company being supported by the state to sell at a loss. They could still do it; but the diplomatic angle might be less favorable.

It'd also be interesting to know if they would want to or not: Aside from some very feeble stirrings(I think some of the Loongson 3 MIPS64 stuff was supposed to have hardware assisted x86 emulation; but nobody seems to have heard from that recently), China has basically zero domestic x86 production, so they may well prefer to just get cheap silicon for themselves and more demand for (Chinese-assembled) devices built around cheap Intel silicon.

The question is whether or not you go the way of bnetd, which worked just fine; but couldn't take the legal heat.

(Also, if it's a console, or a PC title with nasty DRM or a 'warden' style thing, convincing it to connect to something that doesn't have the vendor's SSL cert could be a bit of a trick, even if you have a protocol and behavior compatible server.)

Given that a 'tower rig' is basically a server turned on its side, with fewer 40mm fans and some of the classy reliability features cut, that category will take a great deal of killing. On the other hand, the CPU in a server or tower is almost certainly using nearly as many of the power gating, adjustable clock speed, and various other thermal protection and power saving strategies as the mobile CPUs are. Overall efficiency is still going to be lower ('eh, we're on AC, just keep the PSU energized so a USB peripheral can wake the system!' isn't god's gift to brilliant standbye power numbers); but 'mobile' and 'desktop' have been on something of a collision course ever since the P4 flamed out, almost literally, and Pentium M derivatives took over.

It tends to be; but I think regulatory authorities only get nervous if it shows signs of being dangerously effective, or if there is reason to believe that the pockets behind it are deep enough to ignore losses almost indefinitely(as with international dumping/tariff slapfights, where a mixture of xenophobia and the fact that a nation state can typically afford to keep dumping longer than a company can afford to keep competing).

In the case of Intel trying to break into tablets, my understanding is that it's a known matter of fact that Bay Trail parts are being practically given away(along with a nontrivial amount of Intel software work, including an emulator to handle ARM NDK stuff and general porting and polishing to make the x86 Android not look like, say, the blasted hellscape that is MIPS Android); but it is less clear whether Intel has been able to dump hard enough to actually damage competition.

The one product line that they definitely helped bury was Windows RT (which was mostly an unloved bastard child anyway, even before you could cram an x86 into the same chassis, and definitely had no reason to exist afterwards); but that didn't hurt MS much, since the quality of Windows tablets went up. In the wider ARM ecosystem, ARM Ltd, themselves seem to be riding high and unbelievably cheap SoCs continue to pop out of the woodwork.

Their Bay Trail pricing has definitely made x86 Android something you might actually see in the wild, and tablet-Windows something you might actually consider at a sub-Windows Surface price point; but it doesn't seem to have crushed the ARM market very much.

The part that worries me was: "The hardest challenge was explaining the language of the test to a five-year-old. But he seemed to pick it up and has a very good memory."

Sounds like the kid is pretty bright, might well be pretty impressive in a few years; but 'explaining the language of the test' is pretty much a (much easier) equivalent to 'identifying the problem to be solved'.

As an exercise in mental capacity, I'm definitely not going to knock the kid, I certainly wouldn't have managed it at 5, and those capabilities will likely come in handy, I hope for him that they do.

For the MCP, on the other hand, it seems pretty dire that it can be passed by somebody with an excellent memory; but a need to be coached on what the questions mean. Real life is an open book (and/or google) test; but it is notably unsympathetic about telling you what the questions mean, what sort of answer a given question requires, which questions are actually on the test, which answers trigger a surprise exam about disaster recovery 18 months from now...

If somebody is a 'Certified Professional' I'd much rather seem them have an elegant grasp of what the problem is and what the solution should look like; but check the manual for some registry settings, than be conceptually befuddled but have a perfect grasp of the details.

It reminds me of how they used to disable any built-in CD recording features on systems that CD burners; but not Apple-blessed CD burners.

Given the teething issues of SSDs, I don't doubt that an example could be provided of some drives where 'TRIM support' means 'the intern tested it all day on his win7 box and nothing bad happened' and It Would Be Bad if OSX tried to interact with the feature. Aside from that, though, you don't make profits like Apple does without providing a little encouragement to buy high-margin upgrades.

I'm deeply underqualified to tell you how DC-DC converters do work; but only the simplest and nastiest ('linear regulators') step down voltages resistively. Those ones are nice and simple, so they do still show up in low power applications; but their inefficiency rules them out in cases like this. I don't know exactly how it is achieved; but the more sophisticated designs are capable of 90% or greater efficiency, which keeps heat related issues at bay.

As for quoted amperages, high end graphics cards get most of their power from the 12v rail (via one or two 6 or 8 pin auxiliary power connectors, plus what the PCIe slot provides). 12v is the highest voltage easily available inside a standard PC, so is the best choice for feeding a high powered component without too much loss in the cabling between the PSU and the card. However, the only things on a graphics card that actually uses 12v are the fans. The actual GPU chip, the memory, and all the various support components use substantially lower voltages( core voltage for recent GPUs is ~1.1v, GDDR5 is ~1.5v).

Because it would be brutally impractical to transfer substantial power at such low voltages, the conversion from 12v happens on the card, as close to the actual GPU chip as possible. The amperage for the card as a whole is expressed without paying attention to the amperage after this step-down step, since only the 12v draw is externally visible; but with such low vCores, the GPU chip, rather than the card, necessarily draws almost an amp for every watt it dissipates. 200A would definitely be on the high side; but not outrageously so.

Modern core voltages. In the suitably...enthusiastic... segments, a GPU that pulls 180-200 watts isn't hard to find (even without any overclocking, just off the shelf values) and the core voltage is around 1.1v, so there will be a pretty significant current involved. That's why the DC-DC converter tends to be so close to the chip it feeds, resistive losses would be brutal.

My mistake, misread as 'macbook' pro. Still likely to be stuffed full of DC/DC converters. It's been ages since the PSU actually directly powered much of the more demanding silicon.

You can get in-system UPSes for desktops; but they tend to be specialty items (Logic Supply is the one I know of, no relation to them, just the example I could find most easily. Fits in a 2.5in bay, takes standard size Li-ion cells, pretty cute); For anything remotely affordable, the answer seems to be "Buy a laptop". Despite the increasingly overlap between laptops and desktops in the low power and 'all-in-one' segments, I don't know of anything (not sold as a laptop) that offers a fully 'laptop-style' power system, with the option of a nicely integrated battery. Not quite sure why.

No idea if it's the GPU; but high-frequency magnetics are all potential culprits (as are low frequencies, though 50/60Hz is usually 'hum' rather than 'whine'), and a modern laptop is just stuffed with DC/DC converters keeping the various ICs fed, so if it isn't the GPU's support system, it's another fairly similar one.

I understand that high-frequency magnetics are at risk of physical oscillation(the detailed math is right over my head; but all it takes is one part of the part attracting or repelling another part of the part, at least under some input waveforms, and you'll potentially see movement, which easily enough turns to sound); but the seemingly obvious solution is just to pot the magnetics in an adequately thermally conductive epoxy or other encapsulant.

Does anybody know if that just adds too much cost, without performance benefit, and so gets cut during the BOM penny pinching? Do potting compounds have properties that degrade the performance or efficiency of common magnetics? Why is it that, if coil whine is an issue, they aren't just dipping the things in epoxy and calling it a day?

It presumably depends on how your UPS capacity is distributed, as well.

Lead acid is damned heavy, and offers mediocre density; but if you are just going to shove them in the basement, or are building on the cheapest flat land in the middle of nowhere that you can find, that may not be a problem. However, if cabling costs or resistive losses make 'distributed' UPSes, with fewer big battery banks and more, smaller, battery packs powering individual systems (presumably also cutting DC/AC inverter losses out of the equation and providing DC directly to the motherboard) the superior density and lighter weight of Li-ion is much more attractive.

Given the trend toward relatively 'open air' (obviously with filters and stuff, they aren't letting rain get on the racks) designs to minimize cooling costs, I can imagine that engineering around Li-Ion's overtly suicidal tendencies would be doable(if you factor in the markedly better energy density and lower weight, you end up needing a relatively small enclosure, sufficient to resist direct burn-through and with emergency vents that channel the noxious gasses away from anything delicate and into the airflow of the datacenter); but the tendency of Li-ions to be even pricier than APC Blessed Replacement Gellcells, and born to die, seems fairly intimidating.

I assume that Facebook can purchase enough to get all the volume discounts that vendors will offer on 18650s; but I'm genuinely surprised that that's enough, for non-vehicle applications(where weight and bulk are an obvious problem) to beat out the nasty, heavy; but cheap and mature lead acid stuff.