namibj

There's actually little that changed in a way too fundamentally to matter other than _perhaps_ getting the async load-from-global-to-shared-memory DMA memcpy that avoided blocking register file space as target buffers for in-flight read-from-global operations. Shared after all is just a partition of L1d$ since iirc Volta (since they offered non-fixed/at-launch-requested expanded shared capacity support), so it made sense to provide this not-just-a-hint "prefetch into this user-managed slice of what is otherwise L1d$": it's AFAIK basically just some special load-like units that ask special L1d$-miss-fill units to deliver to a now-explicitly-specified target location in the non-automatic-cache partition of the local SRAM and signal completion in otherwise fairly normal local semaphore/barrier fashion.

The major difference is that this doesn't have a natural moment to transform/touch the values after read from global and before storage to shared.

Otherwise, tiled MMA (gemm) kernels where normal even in Maxwell days (after the classic K80, before the P100; Maxwell is when H.265 support landed).

IMHO there should be extra incentives for BIPV just to the amount that would offset the classic shingles underneath because not just doing a simple barn roof (single gable, two pitch) or triangle roof (single pitch) with the slope entirely covered by solar glass of usual 400~800 Wp size modules is where a big part of the excess wasted cost is from. Just make sure to allow structures that allow module-sized parts of the grid to be replaced with human occupancy windows in a nice and simple way.

Do you have any pointers like that solroof that don't bother trying to be inactive anymore?

My benchmark there is a ventilated attic with insulation between attic and house, using traditional glass roof structure filled with frameless glass-glass panels instead of human-below rated laminated safety glass.

Like, it can't be that hard to do better when you're not trying to be substantially fancier than a classic gable roof, it's just that the panels need to not be shear-loaded much, and the structure doesn't have to be so pretty from underneath as traditional glass roofs.

Solar is REALLY CHEAP. And provided you keep existing central European gas heating infrastructure around for a while, you can basically just wait out the really good energy storage by using existing caverns you pre-fill with methane to keep your people from freezing. If you're not curtailing a substantial fraction of PV yield (yearly) in central Europe that's a sign there way not enough capacity yet.

Built facades and roofs out of glass-glass PV laminate. We have the technology from glass roofs/facades; you just add glass-catching-mesh/insulation below because you can't use the insulated multi-pane window glass construction with safety lamination and solar cells all three together.

Says electrically 3x2.

What they mean is that the cost per bit both capex and opex/power is worse for 10G than 25G for a while now as long as you talk about new hardware.

We're at the point where 25GBaud PAM4 is being replaced by 50GBaud PAM4. That's 50 to 100 Gbit/s.

But iirc the use of PAM4 for the faster ones than "only" 25Gbit/s lanes is a hindrance to managing bottom-barrel price-per-bit. PCIe 3 was 8, PCIe4 was 16, and PCIe 5 is 32 GBaud with a line code basically like the 10+ Gbit/s Ethernet links (well, it's 66b/64b for Eth and 130b/128b for PCIe).

10m is extended reach copper, you can do about triple the range of 100m with approximately the same transceiver analog prowess.

Modern cable trenching at least if you're not hitting rocks is to take a wet vac and a pressure washer and just cut a slit, make sure you got ergonomics sorted and a place to dump the sludge for drying (c.f. kiddie pool, or one of those pools that rely on the top inflated ring to keep the otherwise loose bag of water from spilling...except made from geotextile or something rock/dirt friendly that'll filter the sediment letting the water seep past) before you backfill.

Use simple roof geometry. Use frameless panels in rails/frames made for keeping the building dry under exposed glass roof. Put regular insulation underneath, don't just expose it.

Profit.

Just hard to get the stuff it seems, mostly because the market has a fetish for early-gen retrofit/independent solar panels, as far as mounting goes.

Look at where the Barmen gas power plant is in Wuppertal, Germany to see....

They are like bulbs on an old-school Christmas tree light string, all in series with just one insulated conductor between.

B(+)Trees do actually admit a fast intersection (they offer a way more powerful projected-to-shared-keyspace mutual index join, technically it's even able to do antijoin but that'll actually modify iteration more than a very genericalized inner join; basically whenever you look at a key in any one of the involved indices you project it to a shared keyspace before doing the comparison-based-search things):

You get cache locality from the upper layers, and for navigation basically `let mut head = keyspace.min(); 'outer: while(!cursors[0].finished()){ for(&mut cursor in cursors.iter_mut()) { let new_head = cursor.seek_to_target_or_next_after_if_none_match(head); if (head != new_head) {continue 'outer; }} /* passed all without seeking past target on any one */ output_fun(head, cursors.iter().map(|x| x.val())); }`. If you want you can do the inner loop's seeks concurrently, which helps if those are IO latency bound and you can afford to waste absolute IOPS on eagerly doing that. You'll want to locally compute the max() of those returned and assign that to `head`. Imagine the cursors are lambda-parametrized to feel like they operate on the projected shared keyspace.

If the keys are a bitstring prefix suited to a binary prefix trie you can actually intersect that way, it's beyond worst case optimal when multiple key columns are involved. Sadly any simple implementation strategies of those algorithms have prohibitive external-memory-machine coefficients for their nominally poly-logarithmic IOPS, due to involvement of combinatorial explosion / curse of dimensionality in search tree/trie structures. They do work though. C.f. "Tetris-LoadBalanced"/"Tetris-Reordered".

The latter even tames one index containing "all" even numbers and the other "all" odd numbers, well, matters more if you involve 3+ columns :D

Speculative decoding is already gambling.

There is one for the f64 FMA recycling IFMA from AVX512 they have for bignum libraries;it's a 52 bit unsigned multiply and accumulates either the low or the high output halves into a 64bit accumulator.

It's surely no 64 bit but it's much more than 32 bit. And it's giving you access to the high halves so you can use it to compute 32x32->64 on vector even if only half as packed as that could be.

No they isolate L and N not PE.

None of the DDR2 and onwards memories have anywhere near enough bandwidth to meet refresh frequency on each bit by you even just reading it in a loop.

The refresh that we do is run in parallel on the memory arrays inside the RAM chips completely bypassing any of the related IO machinery.

First reaction to warning tone should always be to (safely!) stop and assess.

Considering that the persons involved can't be expected to not be deaf, or functionally so via e.g. headphones, and thus you always have to be able to brake anyways. Running onto a driving lane (be it bikes or cars doesn't matter) without looking especially if the direction you didn't look just gave an audible warning is always reckless.

In Germany it's illegal to drive bikes that assist beyond 28km/h (about 20mph) in what are true bike paths (which can be built as lanes! And, notably, they can be marked as virtual-lane-shared (pictogram side by side with a vertical divider) or as true shared (pictogram above and below at a horizontal divider), if pedestrians are also allowed to use them.

An ancient gas-e-bike rating is allowed on them outside city limits but iirc those bikes are exceedingly rare since even before e-bikes became truly mainstream.

RTL-SDR-grade fleet doing passive radar (using radio/TV OTA broadcasts) isn't actually that new; but pretty much any detailed reports have caught self-censoring after TLA visitors came by.

And what about the CA?

Upgrade to a better one in initramfs?

u32::from_be_bytes

u32::from_le_bytes

u32::from_ne_bytes the n stands for native

The latter can definitely afford a support contract.

JS numbers behave much more like C's definition of signed overflow being UB as it's signed numbers are effectively like 51-ish bit with a SEPARATE sign bit and non-assiciative behavior when overflow happens.

Any Ethernet (well, any RJ45 you expect in a home/office) has to have at least 1500V isolation from the RJ45 wire to anything metal that can be touched or is a connector on the device. A PoE-only device with no electrical connectors besides the RJ45 can just use a very cheap RJ45 port with integrated magnetics and PoE allowance (tiny bit bigger wires and a center pin exposed, less than 50ct more than the cheapest RJ45 with integrated magnetics) and a cheap buck from 40~80V to e.g. 5V.

Oh, and a cheap bridge rectifier and some signaling resistors to take care of input polarity and signal to the source that we in fact want the approximately 50V that could hurt a device not made for it.

Central Europe has more buried electrical systems and nothing is safe until it's actively grounded so hard it'll arc flash the idiot who broke you LOTO before you even feel a clear tingle.

The 800W is about grid management impact limitation to levels that do not warrant the utility imposing any "but we first have to upgrade the substation before we can get you your local transformer with the higher speed EV charging and McMansion winter full heat pump setup" delays before you are allowed to turn it on/grid-tie it.

Ehhhhh, what's your problem with booster seats? Those are just to get the shoulders into the design zone as needed for the typical 3-point seat belt.

It's based on size (shoulder height measured in sitting position) not age.

I think some 5-point seat belts allow a much wider design zone so won't need boosters by the time the children outgrow the front-facing bucket seat.

Iirc classic WRT-841 and similar "300Mbit WiFi" generation 4-100Mbit-ports Wi-Fi routers had the CPU attached via an on-SoC gigabit link to a vlan capable switch that has the 4 100BASE-TX ports exposed.

The fix for not doing that by hand is to get a robot to do it, given the applicator is human-held, a human-strength Kuka with enough reach to cover the area it can handle before the applicator needs refurbishment of some sort which would give a good opportunity to move the robot to a new section of the heat shield.

Just fork git and patch that out? Can't be that hard just ask the agent for that patch. Don't need to update often either, so it's ok to rebase like twice a year.