egg changed the topic of #kspacademia to: https://git.io/JqLs2 | Dogs are cats. Spiders are cat interferometers. | Document well, for tomorrow you may get mauled by a ネコバス. | <UmbralRaptor> egg|nomz|egg: generally if your eyes are dewing over, that's not the weather. | <ferram4> I shall beat my problems to death with an engineer. | We can haz pdf | Logs: https://esper.irclog.whitequark.org/kspacademia
* raptop engages in duolingo shenanigans
<raptop> !U 🥺
<galois> 🥺: U+1f97a FACE WITH PLEADING EYES
e_14159 has quit [Ping timeout: 189 seconds]
e_14159 has joined #kspacademia
egg|anbo|egg has quit [Remote host closed the connection]
<raptop> !wpn
* galois gives raptop a hallucination
<raptop> ...why does the hallucination have 6 wings, and is shrouded in flames?
<whitequark> egg|anbo|egg_: yes sure!
<whitequark> i've tried setting up a matrix homeserver (which works ok) and a bridge (which didn't work for inexplicable reasons)
_whitelogger has joined #kspacademia
<raptop> !8 Should I be worried that whitequarks are dual-use?
<galois> raptop: yes
<whitequark> be worried.
egg has quit [Ping timeout: 189 seconds]
egg has joined #kspacademia
egg|anbo|egg has joined #kspacademia
egg|anbo|egg has quit [Remote host closed the connection]
egg|anbo|egg has joined #kspacademia
<egg|anbo|egg> !wpn whitequark
* galois gives whitequark a MSSM diurnal alligator
egg|anbo|egg has quit [Remote host closed the connection]
egg|anbo|egg has joined #kspacademia
<egg|anbo|egg_> whitequark: what are the coordinates of the things to which I should bridge :-p
egg|anbo|egg has quit [Read error: Connection reset by peer]
egg|anbo|egg has joined #kspacademia
queqiao- has joined #kspacademia
queqiao- has quit [Read error: Connection reset by peer]
queqiao- has joined #kspacademia
<queqiao-> e​ggrobin:matrix.org. — test
<egg|anbo|egg_> test
<whitequark> egg|anbo|egg_: coordinates?
<egg|anbo|egg_> nevermind I was confused
<egg|anbo|egg_> (as in, what should I bridge to)
<whitequark> oh! make a #kspacademia:matrix.org room or something?
<egg|anbo|egg_> yeah I did
<egg|anbo|egg_> right now the bridge bot speaks as itself on the matrix side though, and I wonder whether there is a way to make it smarter there
<whitequark> "You are not invited to this room"
<whitequark> rude
<queqiao-> w​hitequark:matrix.org. — nya
<egg|anbo|egg_> ня
<egg|anbo|egg_> OK at the very least I should prefix messages with IRC usernames if it is going to work like that
<egg|anbo|egg_> test
<queqiao-> e​ggrobin:matrix.org. — test
queqiao- has quit [Read error: Connection reset by peer]
queqiao- has joined #kspacademia
<queqiao-> w​hitequark. — nice
<queqiao-> e​ggrobin. — Hmm, unfortunately the bot isn’t going to work from this side
<queqiao-> e​ggrobin. — !wpn
<egg|anbo|egg_> !wpn
* galois gives egg|anbo|egg_ a significant footnote
<whitequark> why the dot?
<egg|anbo|egg_> oh, I think I went with French typographic style for dialogues
<queqiao-> w​hitequark. — can we... not
<queqiao-> w​hitequark. — it's disorienting
<egg|anbo|egg_> how about this
<queqiao-> w​hitequark. — works :3
<queqiao-> w​hitequark. — thank u
<egg|anbo|egg_> I could do the same on the IRC side, but then it would change the format of the Principia logs since they use the same setting
<queqiao-> w​hitequark. — well, my idea was to drop irssi, so... doesn't really matter?
<egg|anbo|egg_> well, depends on whether you use the whitequark logs
<egg|anbo|egg_> they are good logs
<queqiao-> w​hitequark. — nya~
* egg|anbo|egg_ pets whitequark with a cat
* queqiao- e​ggrobin. — wonders what /me looks like in the other direction
* queqiao- w​hitequark. — meows
<queqiao-> w​hitequark. — it's eaten
<egg|anbo|egg_> nah it does show up as an ACTION (but from the bot)
<queqiao-> w​hitequark. — ohhh
<egg|anbo|egg_> then again I changed the format of the IRC logs in #principia by changing the IRC nick from discord- to queqiao, so I might as well change the formatting as well
<queqiao-> ⟨e​ggrobin⟩ done
<egg|anbo|egg_> I should upgrade the bridge so it knows to send the nick rather than the username
queqiao- has quit [Read error: Connection reset by peer]
queqiao- has joined #kspacademia
<queqiao-> ⟨e​gg⟩ test
<egg|anbo|egg_> much better
<egg|anbo|egg_> test 𒉭
<SnoopJ> !wpn queqiao-
* galois gives queqiao- a starling
<raptop> birb
<SnoopJ> Apt name
<SnoopJ> raptop, 竜学生 == Asuka
<raptop> o_O
<raptop> 竜学生 祖竜 langley?
<queqiao-> ⟨w​hitequark⟩ might make more sense to have /me on Matrix side be emitted as a "* " prefix
<egg|anbo|egg_> yeah, but I don’t know whether there is a setting for that
<egg|anbo|egg_> (I am using this thing) https://github.com/42wim/matterbridge
<galois> title: GitHub - 42wim/matterbridge: bridge between mattermost, IRC, gitter, xmpp, slack, discord, telegram, rocketchat, twitch, ssh-chat, zulip, whatsapp, keybase, matrix, microsoft teams, nextcloud, mumble, vk and more with REST API (mattermost not required!)
<queqiao-> ⟨w​hitequark⟩ that's a lot of networks
<queqiao-> ⟨w​hitequark⟩ wtf is ssh-chat
<egg|anbo|egg_> huh, вк
<SnoopJ> BGP-chat
<raptop> SCPing text files back and forth?
<queqiao-> ⟨w​hitequark⟩ SnoopJ: what was your twitter acc again?
<SnoopJ> whitequark, I am the computational meat responsible for @jgerity
<queqiao-> ⟨e​gg⟩ > <@queqiao:matrix.org> ⟨e​gg|anbo|egg_⟩ yeah, but I don’t know whether there is a setting for that
<queqiao-> ⟨e​gg⟩
<queqiao-> ⟨e​gg⟩ aaaargh, the stupid bridge bot has encoding issues >_<
<egg|anbo|egg_> in a weird nondeterministic way, too
<SnoopJ> !u ⟨
<galois> ⟨: U+27e8 MATHEMATICAL LEFT ANGLE BRACKET
<SnoopJ> ah
<SnoopJ> bad windows font is bad
<queqiao-> ⟨e​gg⟩ ah there is a charset option
<egg|anbo|egg_> test case : ’
<egg|anbo|egg_> nope
<egg|anbo|egg_> test case ’
<egg|anbo|egg_> aaargh
<egg|anbo|egg_> I don’t know how to run into the encoding issue
<egg|anbo|egg_> OK maybe if I reuse the same sentence
<egg|anbo|egg_> yeah, but I don’t know whether there is a setting for that
<egg|anbo|egg_> ~magic~
<egg|anbo|egg_> yeah, but I don’t know whether there is a setting for that
<egg|anbo|egg_> OK now it should work
<egg|anbo|egg_> raptop: this should also fix your recurrent encoding woes in #principia
<raptop> yay
<raptop> !u 🥲
<galois> No info for U+1f972 (I only know about Unicode up to 12.1.0)
<raptop> !u 🪰🪰🪰
<galois> No info for U+1fab0 (I only know about Unicode up to 12.1.0)
<galois> No info for U+1fab0 (I only know about Unicode up to 12.1.0)
<galois> No info for U+1fab0 (I only know about Unicode up to 12.1.0)
<queqiao-> ⟨e​gg⟩ > <@whitequark:matrix.org> that's a lot of networks
<queqiao-> ⟨e​gg⟩
<queqiao-> ⟨e​gg⟩ Sadly no QQ, it might be fun to bridge #principia with something there
<queqiao-> ⟨e​gg⟩ (Principia has a fairly active Chinese userbase amusingly)
<queqiao-> ⟨w​hitequark⟩ nice
<queqiao-> ⟨w​hitequark⟩ oh, finally can post here on my phone
<raptop> yay
<raptop> !8 Should whitequark visit the US to see if cicadas taste better than moths?
<galois> raptop: no
<raptop> bah
<queqiao-> ⟨w​hitequark⟩ hah
<queqiao-> ⟨w​hitequark⟩ can obtain cicadas elsewhere?
<egg|anbo|egg> Not Magicicada sp. tho
<queqiao-> ⟨e​gg⟩ huh, when I speak from another client it pings me here
<queqiao-> ⟨e​gg⟩ ah because the nick has egg twice
<queqiao-> ⟨e​gg⟩ and the nopingnick garbles only the first egg
<raptop> nick-scrambler
<queqiao-> ⟨e​gg|matrix|egg⟩ I suppose this is the obvious solution to this problem
<raptop> hah
<egg|anbo|egg> (insert obvious density matrix pun here)
<raptop> But shouldn't it be |egg>matrix<egg|?
<egg|anbo|egg> hm, the density matrix for the pure state |egg⟩ would be |egg⟩⟨e​gg|
<egg|anbo|egg> iximeow: whitequark: do you happen to know whether there is a way to use intel FMA intrinsics (_mm_fmadd_sd et al.) with clang without making every floating-point instruction VEX-encoded?
<egg|anbo|egg> MSVC is perfectly happy to have those and keep emitting good ol’ SSE2, leaving it to me to do the appropriate CPUID checks, but clang wants -mfma then I end up requiring AVX
<iximeow> ... what does msvc emit for sse2?
<egg|anbo|egg> iximeow: well non-VEX encoded instructions, so it runs on prescott and later
<egg|anbo|egg> addsd etc., as opposed to vaddsd etc.
<iximeow> right, i mean i don't know of a non-vex FMA
<iximeow> addsd/mulsd makes sense though
<egg|anbo|egg> right, but I don’t want the compiler to emit FMA on its own anyway
<egg|anbo|egg> (since 1. that would affect correctness, 2. I need to guard that behind CPUID checks obviously)
<egg|anbo|egg> but if suddenly every addsd and mulsd becomes vaddsd/mulsd, I am dropping Westmere and earlier support just because somewhere behind a CPUID check that will come out false, I have an FMA…
<iximeow> yeah
<iximeow> i don't really know MSVC things so i was mostly perplexed at the notion of SSE FMA :x
<egg|anbo|egg> yeah it is not a thing :D
<iximeow> could you conditionally compile both forms and link two objects?
<egg|anbo|egg> yeah eventually I will have to do that (I would benefit from AVX too anyway)
<egg|anbo|egg> (and obviously using 256-bit stuff requires a whole separate binary because of the VZEROUPPER dance)
<egg|anbo|egg> iximeow: MSVC *will* warn if you use __m256d in an SSE2 binary, but it will even let you do that
<egg|anbo|egg> clang otoh has Opinions and really wants the whole thing to be for one arch apparently
<iximeow> booo
<egg|anbo|egg> iximeow: hmm, there is the question of whether the VZEROUPPER dance would be needed when there is an FMA in the middle of non-VEX SSE2 code
<egg|anbo|egg> it shouldn’t, I think, since nothing here is dirtying YMM (the FMA, though VEX-encoded, only touches the XMM half)
<egg|anbo|egg> iximeow: do you happen to have a Haswell or a Broadwell
<iximeow> i do not in fact
<egg|anbo|egg> hm
<iximeow> i don't think the VZEROUPPER dance should be necessary in that circumstance though
<egg|anbo|egg> yeah, and the lack of warning in MSVC seems to point in that direction too
<egg|anbo|egg> (since MSVC will be grumpy if you utter the words __m256d when building for SSE2)
<egg|anbo|egg> iximeow: oh actually it isn’t even grumpy, I thought it was but apparently it just vzerouppers as appropriate apparently
<galois> title: Compiler Explorer
<iximeow> oh wow
<iximeow> that's awful nice of it
<egg|anbo|egg> (it also likes to move it, move it, but in practice its generated code is actually quite decent; toy examples are toy examples)
<egg|anbo|egg> (also moves are fake actually since Ivy Bridge)
<iximeow> hmmmm i would now like to see a compiler implement pshufb not as an intrinsic but as a series of movb
<egg|anbo|egg> iximeow: speaking of code generation, did I mention https://github.com/mockingbirdnest/Principia/pull/1731#issuecomment-851004013
<galois> title: Remove class Wide by pleroy · Pull Request #1731 · mockingbirdnest/Principia · GitHub
<UmbralRaptop> !8 Does levitating avocado pits enhance the flavor yield of guacamole?
<galois> UmbralRaptop: no
<raptop> "Not a bad way to start your day; with an endless array of cubes!"
<raptop> !wpn
* galois gives raptop a rusty proposal
<raptop> Ah, yes, the tetanus space telescope
egg|anbo|egg has quit [Remote host closed the connection]