00:20 <UmbralRaptor> This is why ISO 8601 versioning is good. https://twitter.com/superSGP/status/985629779041890304

00:20 <kmath> <superSGP> when it's definitely your final javascript spec draft https://t.co/bUx4BWiZqg

00:54 <UmbralRaptor> "This email is to inform you that your phone was updated and now has the Google Assistant. You can access the Google Assistant by touching and holding the Home button on your phone."

05:20 <UmbralRaptor> @_@ https://twitter.com/SuperASASSN/status/985710363672039425

05:20 <kmath> <SuperASASSN> "You call this an M-dwarf flare? This is an M-dwarf flare!" - nice research note by @romy_rodz " ASASSN-18di: disco… https://t.co/bCgC6A3oaT

05:20 <UmbralRaptor> !wa 1e36 erg tnt equivalent

05:20 <Qboid> UmbralRaptor: 1000000000000000000000000000000000000 ergs of C_6H_2(CH_3)(NO_2)_3 (trinitrotoluene) --> (unspecified products): (data not available)

05:20 <UmbralRaptor> …

05:22 <UmbralRaptor> !wa 1e36 / 1e7 / 4.184e9 / 1e18

05:22 <Qboid> UmbralRaptor: (((1×10^36)/(1×10^7))/(4.184×10^9))/(1×10^18): 23.90057361376673040152963671128107074569789674952198852772...

05:22 <UmbralRaptor> 24 exatons

05:51 <UmbralRaptor> <@Myke> hobbs: what's your mom's cat's twitter account?

07:55 <egg|zzz|egg> !wpn UmbralRaptor

07:55 <egg|zzz|egg> !wpn whitequark

09:37 <egg|zzz|egg> bofh: so the egg cbrt with Halley is faster on my laptop, but slightly slower on my (sandy bridge) desktop. IACA says it's 2 cycles slower on sandy bridge and 1 cycle slower on haswell (newer IACAs don't have latency analysis so not sure how it fares theoretically on skylake; eggsperimentally it's faster there)

09:47 <egg|zzz|egg> bofh: as the cat said, division gets faster and the trade-offs change in consequence

09:47 <egg|zzz|egg> !wpn whitequark

11:50 <rqou> bofh: random question: there's a bunch of lapack functions that have "blocked" logic wrapping "unblocked" logic

11:51 <rqou> e.g. SGETRF vs SGETF2

11:51 <rqou> can i just always use the "unblocked" version?

11:51 <rqou> afaict the "blocked" version is pessimized on cortex-m anyways

11:51 <rqou> at least for the matrices of interest to me (almost certainly no larger than 4x4)

13:10 <bofh> rqou: yes, I in fact do exactly this

13:11 <bofh> call SGETF2 directly and whatnot

13:11 <bofh> it works quite fine and results in dramatically smaller codesize

13:14 <whitequark> !wpn egg

13:36 <bofh> https://twitter.com/chordbug/status/985800050398973957 ACCURATE

13:36 <kmath> <chordbug> harmony is really funny. because it's like, someone goes "check out my cool rational numbers collection that I've p… https://t.co/d210X6cmdk

14:56 <UmbralRaptor> Why is 2^{1/12) in the rational numbers collection?

15:32 <egg|cell|egg> Bofh: the complaint to ea Nasir isn't on display, some cute envelopes though

15:34 <UmbralRaptor> o_O https://twitter.com/stavvers/status/985842357022220288

15:34 <kmath> <stavvers> A reminder that as Home Secretary, Theresa May also "accidentally" deported almost 50,000 students illegally. She's… https://t.co/UKeZoJnYVF

15:34 <UmbralRaptor> bofh: I have questions.

15:54 <egg|cell|egg> !Wpn UmbralRaptor and bofh

16:06 <UmbralRaptor> !wpn egg|cell|egg

16:11 <egg|cell|egg> Flammalloc

16:36 <egg|zzz|egg> bofh: they sell mugs with text from the rosetta stone at the british museum; that made me think mugs with cuneiform impressed in them would be fun :-p

16:44 <egg|zzz|egg> bofh: other references to ea-nasir: https://books.google.co.uk/books?id=EXUCwoXjUqgC&pg=PA150&lpg=PA150&dq=%22Ea-nasir%22+%22na-ni%22&source=bl&ots=OLCbApIg0p&sig=ra2cqFYD09UJNER3N9eSi0BLEjw&hl=en&sa=X&ved=0ahUKEwielOHgnr_aAhWLD8AKHZwLACsQ6AEIKTAA#v=onepage&q=%22Ea-nasir%22%20%22na-ni%22&f=false

16:48 <egg|zzz|egg> not sure whether that's the same person though, no mention of the copper thing

16:48 <rqou> bofh: so is it just me or does the majority of lapack APIs suck because it doesn't have modern abstractions?

16:48 <rqou> e.g

16:49 <rqou> everything duplicated 4x for different data types

16:50 <rqou> and dragging "LDA" everywhere because it doesn't have lightweight structs

16:54 <egg|zzz|egg> bofh: on the archive in plain text: https://archive.org/stream/CuneiformTextsinTheMetropolitanMuseumofArtVolumeIIIPrivateArchiveTextsfromtheFirstM/CuneiformTextsinTheMetropolitanMuseumofArtVolumeIIIPrivateArchiveTextsfromtheFirstM_djvu.txt

16:55 <egg|zzz|egg> hm, this one seems to trade in garlic, not copper-

16:59 <rqou> 🗡️lapack 🗡️fortran

17:11 <egg|zzz|egg> !u ️

17:11 <Qboid> U+FE0F VARIATION SELECTOR-16 (◌️)

17:47 <egg|zzz|egg> bofh: yeah the tradeoffs change for atlas vs. householder 10, and halley vs. double-newton (lines go sandy to skylake right to left) https://i.imgur.com/i6WAFDl.png

18:06 <egg|zzz|egg> https://twitter.com/stephentyrone/status/985936723938037765

18:06 <kmath> <stephentyrone> @jhripley @RichFelker Floating-point numbers that have low-order bits are unstable; they decay to more stable state… https://t.co/A0IMMYIzya

18:38 <BPlayer> I am kind of screwed. I need to calibrate this sensor, but I can only do it by sending a command from the PC, and it's so sensitive to minimal movement that if I shift a bit on my chair nearby, the sensor (50 cm away from the chair) fails to calibrate properly. So I need to push this button and hold my breath for a minute or so while it calibrates.

18:38 <BPlayer> Every single test run.

19:31 <BPlayer> If I have an euler angle (or do you call that three euler angles?), how do I best rotate a vector by it?

19:38 <UmbralRaptor> Witchcraft.

19:47 <UmbralRaptor> Okay, problem 2 is rather nifty. https://drive.google.com/file/d/1aUtno2n4NVE-_AqTAoaMvUjM0Y7QWUCO/view?usp=drivesdk

20:17 <bofh> http://www.planetary.org/blogs/guest-blogs/marc-rayman/0302-dawn-journal-intricate-maneuvers.html what the actual fuck

20:17 <bofh> I kind of now want to try something like this in Principia.

20:19 <bofh> they literally *reversed Dawn's orbital direction around Ceres* with an ion engine.

20:20 <SnoopJeDi> O_O

20:21 <BPlayer> As long as the ion engine has enough delta-v, why not?

20:21 <BPlayer> And it sure has

20:21 <SnoopJeDi> ion engines don't exactly give you a lot of impulse

20:22 <SnoopJeDi> i.e. the maneuver would probably be "too slow" for a naive approach (reading now)

20:23 <SnoopJeDi> this is more or less what I pictured when bofh said it though. Neato.

22:20 <egg|zzz|egg> !wpn bofh

22:21 <egg|zzz|egg> !wpn whitequark

22:22 <UmbralRaptor> !wpn egg|zzz|egg

22:23 <egg|zzz|egg> !wpn UmbralRaptor

22:51 <bofh> 16:49:09 < rqou> everything duplicated 4x for different data types

22:51 <bofh> this is much more lightweight than using, say, C++ templates, imho.

22:51 <rqou> why?

22:53 <bofh> requires less compiler complexity, which mattered a lot in the 1970s

22:58 <rqou> meh, i don't really care about that nowadays

23:44 <bofh> https://twitter.com/JPMajor/status/986023670740078592 HALP

23:44 <kmath> <JPMajor> We don't know all that much about Puck other than it has water ice on its surface but it does have three named crat… https://t.co/zpYaVapSnL