arthur2e5

The OKlab definition is probably a bit easier to implement in hardware than CIELAB since there’s no funny “if” branch. Still, the number representation will be more complicated than YUV and such… fp16 maybe?

IIRC author of the TFA has retweeted some shader implementation of OKlab for color interpolation.

Apparently it saves what you do to the address bar. Which is good on its own, but messes up history when you use href.assign() instead of .replace().

https://developer.mozilla.org/en-US/docs/Web/API/Location/re...

The whole HSL and HSV business has never been about hypothetical light sources. HSL and HSV are created from monitor CRT RGB using piecewise functions for the hue. No colorimetric considerations are given at all: it’s all about how to mix an algebraic mean.

I believe the IEEE 754 total ordering is identical to a classical recipe for fast FP comparison.[1] Of course, the handling of NaNs is only identical when the qNaN bit is at the MSB of the significand as recommended.

[1]: http://stereopsis.com/radix.html

Those might be alright. I recall seeing them a lot before Chia was even a thing too. Allegedly cannibalized from laptops (tiny SSD for system + larger HDD for storage configurations) and office desktop computers.

Thought about getting one for swap, but don’t have a free slot.

To add, theanine extraction increases with the time you steep the tea, but so does the caffeine content…

Might be a good idea to go to the PR and mention that. The stale bot has done its awful work… again.

I recall being quite pissed when Chrome decided to remove the pick-character-encoding feature essential for digging through archives of older CJK sites. Odd things keep rolling out of the Chrome team with telemetry as their justification. Ah well, how am I supposed to argue against “data”?

Might be a good idea to retire as “calling a rust library [from Java] with Panama FFI”.

If the author is here: it really won’t hurt to add a short paragraph of introduction about what Panama, the cool new JVM FFI, is.

GitHub Copilot is literally GPT adapted for code. The paper on OpenAI Codex, the stuff powering Copilot, makes it very clear in the abstract. https://arxiv.org/abs/2107.03374

Either you call both AI or you call neither AI.

(A previous version of the comment stated that it was tuned from GPT-3. This is incorrect; the simpler GPT was used for faster convergence.)

For doing these sort of things quickly, try getting the Twinkle gadget (enwp.org/WP:TWINK) — you get a giant searchable list for tags to add. Heck, you can put a BUNCH on at a time!

Also, {{advert}} is a thing.

Certainly off topic, but gotta say I prefer “MADDLESS” over “MADDNESS” for being more… obvious. It still sounds close enough to me, but again I might be just letting my Southern-Chinese ear loose with the n/l confusions…

Pretty big compared to the old RNA secondary structure prediction tools judging from the abstract. Wish someone put it on Sci-Hub… Maybe someone from HN will!

I can totally see a “show HN: webapp that turns your GPU into a discount xeon phi” appearing…

A recent example of videos disappearing is when Radio TV Hong Kong got a new, more pro-Beijing head. The “inflammatory” stuff first went down, followed by anything older than 1 year on YouTube as cover.

The “compile in an old version” thing was IIRC more properly done with linker scripts telling the linker to use an older symbol. I think lsb-gcc does it for you — not that the Linux Standard Base is still widely respected at all.

The CN gov is not addressing the living standard thing as usual, but they are “shedding load” on parents by messing with the curriculum again. There has been harsh enforcement on private tutors[0] and reheated talks about making English non-mandatory.[1] (Let’s just say without the English curriculum I will not learn many of the “skills” that lead to me posting here at all.)

  [0]: https://www.bloomberg.com/news/articles/2021-06-09/why-china-s-cracking-down-now-on-education-tech-firms-quicktake
  [1]: https://www.globaltimes.cn/page/202103/1217396.shtml

And guess what? They are also backpedaling on women’s rights. They are requiring people to jump through extra red tape (managed by the gossipy “community” people!) for getting an abortion and are becoming very conservative with treating uterine fibrosis.[2] Back in the days it was only the ob/gyn requiring family signature to perform hymen-breaking examinations…[3] And relatively barrier-less elective abortions were, well, possibly one of the few upsides of that one-child thing.

  [2]: https://twitter.com/notmychris/status/1429455469039419395
  [3]: https://jezebel.com/why-is-it-so-hard-for-unmarried-women-in-china-to-go-se-1729393595

Cproc+QBE doesn’t seem to be trying to do the linker job itself, so presumably library order (apparently a main point in that PR) does not need to be changed?

AMD and recently Intel both moving to LLVM is a good step in that direction IMO. LLVM also already comes with a dispatcher, so hopefully they aren’t going to… do it again. Personally I’m more excited about being able to port __attribute__ rich code more easily.

(IIRC Intel has tuned down the Cripple AMD thing in MKL around late 2020 by providing a specialized Zen code path. It was slower than what you get with the detection backdoor, but only slightly.)

I have this “ready with tests” PR for almost a year now, and every time I get a notification it’s either the bot or someone who really wants the PR to be merged bumping it up for me. I don’t even care so much about it any more…

Tangential: it would be nice if `@font-face sans-serif { … }` (likewise for other generic families) was a thing in user-injected stylesheets. Especially when you don’t have such options in UI.

What was the original method may I ask? Some… basic RGB mixing?

Compilers are using them,^1 but aren’t necessarily using them enough. They (generally) can’t change alignments of a struct willy-nilly, nor can they just, say, split a big FP reduction into many parallel ones without violating 754. In most loop-related cases however, I would say a nudge in the form of #pragma omp simd goes a long way.^2

And then there’s the genius stuff like SIMD UTF8 decode and whatnot. Compilers don’t magically figure those out. Heck, they sometimes have trouble with shuffles, which is why the author mentions that the new features will help.

^1 Oh, not GCC on -O2 IIRC. There were some talks about turning autovec on that level on like clang does, but I’m not sure it went anywhere… I also think MSVC doesn’t do that.

^2 This pragma also turns on autovec even if the compiler is not told to do that with the flags. Although my original point was about the extra information you can provide with it.

Fast-growing trees turned into biochar in my opinion. Partially because, well, every HN user’s got to have their favorite piece of “proven” old tech.

I have a slight urge to blame this lack of "community" to the institutionalized nature of C's standarization in the style of "Against SQL".[1] Although in this case the barrier is not between research and implementation, but between implementation and standarization.

  [1]: https://scattered-thoughts.net/writing/against-sql/#innovation-at-the-implementation-level-is-gated

(But hey, the defer proposal is here and might actually become a thing... right?)

The struct enumeration thing is certainly useful for homogenous vectors being written as structs, but one of the issues is that it requires the preprocessor to have significantly more knowledge of the underlying language. Would it be a better idea to have something generate X Macros from struct defs instead?

I’ve seen people obfuscating .NET, gotta say seeing python bytecode getting used like this is a first…

God I hope the RenPy game studios don’t see the obfuscator.

The “good” news on this front is that RMS has been allegedly removed from GCC’s steering committee as a side-effect of the flames of April. Emphasis on “alleged”, since some heavyweights say he was never there to begin with.

https://lwn.net/Articles/853230

Somewhere down the thread there are questions about AST, I think.

Cool results and pretty nice formulation (note the page’s emphasis on “no neural net”; see paper/poster link for more). The group can probably make some dewatered versions of NOAA Ocean Expedition videos to show off to the wider Internet.

Patent note is a bit disappointing, but what can I do but respect the author’s decision?

This is not a machine translation issue. The deep fried versions in Chinese make about as little sense as it does in English. Again, this is caused by word-by-word theasurus-running most likely in English.