egg|nomz|egg changed the topic of #kspacademia to: https://gist.github.com/pdn4kd/164b9b85435d87afbec0c3a7e69d3e6d | Dogs are cats. Spiders are cat interferometers. | Космизм сегодня! | 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.
awang has quit [Ping timeout: 186 seconds]
<bofh> fuck x86_64 SIMD intrinsics holy shit
<egg> [a-z]?mmintrin
<bofh> basically.
<bofh> I usually just include immintrin.h on the rare occasion I need 'em.
<egg> context: #1722
<Qboid> [#1722] title: Implement some operations of R3Element using intrinsics | According to IACA the throughput of these operations goes from 3.00 to 2.00 cycles on Haswell, and this is reflected by a gain of up to 40% on the benchmarks of the polynomials. As expected, the effect is more pronounced for Estrin evaluation.... | https://github.com/mockingbirdnest/principia/issues/1722
<egg> bofh: so do we need to find a mathematician who was born in the right year for the new year version names from now on :-p
awang has joined #kspacademia
<bofh> e1
<bofh> egg: I concur and agree.
<egg> that seems tricky
<egg> bofh: for starters figuring out which sign is Εὐκλείδης's seems hard :-p
<egg> although if we manage to get to Euler by next year, we have our pig...
<bofh> Use a different mathematician for that year... also huh, so you do.
<egg> bofh: well we also want to do mathematicians in alphabetical order of (using some ill-defined collation algorithm)
* UmbralRaptor 🔪 Simbad for putting HTML tags in a TSV file.
<bofh> Eew.
e_14159 has quit [Ping timeout: 182 seconds]
e_14159 has joined #kspacademia
<awang> !u Εὐκλείδης
<Qboid> U+0395 GREEK CAPITAL LETTER EPSILON (Ε)
<Qboid> U+1F50 GREEK SMALL LETTER UPSILON WITH PSILI (ὐ)
<Qboid> U+03BA GREEK SMALL LETTER KAPPA (κ)
<Qboid> U+03BB GREEK SMALL LETTER LAMDA (λ)
<Qboid> U+03B5 GREEK SMALL LETTER EPSILON (ε)
<Qboid> U+03AF GREEK SMALL LETTER IOTA WITH TONOS (ί)
<Qboid> U+03B4 GREEK SMALL LETTER DELTA (δ)
<Qboid> U+03B7 GREEK SMALL LETTER ETA (η)
<Qboid> U+03C2 GREEK SMALL LETTER FINAL SIGMA (ς)
<awang> egg: pig?
<X> Quantum physics for a babies is a book and they represent atoms with a Bohr model!
<X> The horror.
<UmbralRaptor> More importantly, how did they get a He- ion?
icefire has quit [Quit: Leaving]
* UmbralRaptor resists RTing the Principia announcement with a doge meme.
tawny has quit [Quit: 「Roundabout」 - To Be Continued]
tawny has joined #kspacademia
<kmath> <dosnostalgic> IWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEACATIWANTTOBEAC… https://t.co/Y8iKzVYI7Q
<egg> UmbralRaptor: why resist
StCypher has joined #kspacademia
StCypher has quit [Ping timeout: 186 seconds]
APlayer has joined #kspacademia
icefire has joined #kspacademia
tawny has quit [Ping timeout: 383 seconds]
icefire has quit [Quit: gone]
awang has quit [Ping timeout: 207 seconds]
awang has joined #kspacademia
egg|wfh|egg has joined #kspacademia
<kmath> <bofh453> 恭喜发财!!!
APlayer has quit [Ping timeout: 383 seconds]
egg|wfh|egg has quit [Quit: Web client closed]
APlayer has joined #kspacademia
<bofh> UmbralRaptor: https://en.wikipedia.org/wiki/Chinese_New_Year#Greetings see the second entry
APlayer has quit [Read error: Connection reset by peer]
<UmbralRaptor> bofh: Google translated it as just "congratulations"
APlayer has joined #kspacademia
<bofh> UmbralRaptor: google translate is not very good at times. I'm actually confused how it pulls that off, did it just ignore the last two Hanzi?
<UmbralRaptor> ``\_(^_^)_/``
<kmath> <AnalysisFact> 'The shortest path between two truths in the real domain passes through the complex domain.' -- Jacques Hadamard
APlayer has quit [Ping timeout: 182 seconds]
<kmath> <SpaceGeck> This is what happened when I was doing a content-aware fill to clean up some cosmic rays, but unwittingly had my se… https://t.co/FnI9HDJxhr
APlayer has joined #kspacademia
StCypher has joined #kspacademia
tawny has joined #kspacademia
tawny has quit [Ping timeout: 207 seconds]
tawny has joined #kspacademia
<APlayer> egg: What can cause an integrator to look fine on the first iteration and spit out total nonsense numbers on the second?
<APlayer> Do you have tips on how to track this down? Because I am lost in all those formulas
<SnoopJeDi> it's easier to look at the problem the other way around APlayer and think about what properties of integrators allow any iteration to be not-garbage
<APlayer> All of them?
<SnoopJeDi> i.e. what properties of the underlying formulae are you assuming/exploiting that allow you to extrapolate
<APlayer> There are too many formulae to do this right now
<APlayer> I am currently walking through every single of them and seeing if they work
<APlayer> Many don't, BTW
<SnoopJeDi> well, I'm not saying pick apart every bit of your integrator, if that's what you interpreted it as
<SnoopJeDi> instead I'm suggesting don't ask "why isn't this working" vs. "why do I expect this to work at all?"
<SnoopJeDi> the difference can be subtle but the devil's in the details with numerics
<SnoopJeDi> and you can probably review the source of your method (RKF45 right?) to review what strokes it takes in building its approach
<SnoopJeDi> APlayer, also, what is your test case? do you have a simple test case that you're validating with (possibly the thing you're asking about?)
<APlayer> I have a script that is made exclusively to simulate a rocket launch. Three forces acting, engines, aerodynamics and gravity. But the numbers are gibberish
<APlayer> Terminates with a NaN error because I am performing a square root on a negative number
<APlayer> Dividing by zero, even, sorry
<APlayer> But the cause is similar
<SnoopJeDi> well, if you want to test your *integrator*, ignore all the fluff and test something you can understand. Even a single pendulum can be helpful since the answer is nearly analytic
<SnoopJeDi> at least, assuming you've written it in a way that allows for that
<APlayer> I can't just remove all that stuff and code new one for a pendulum
<APlayer> Those formulae are like half the integrator
<SnoopJeDi> this is kOS I assume?
<APlayer> Yes
<SnoopJeDi> I'm not super familiar but yea I guess there's not a good way to organize things there from what I know
<APlayer> Well, it's a simple language
<APlayer> Both, in terms of learning curve and in terms of features
<SnoopJeDi> anyway, the answer to your original question is "most things" but I'd suspect a non-mathy error the most of all in this context (wrong sign somewhere etc.)
<APlayer> I guess I'll just have to track some number through this jungle and see where it gets broken
<SnoopJeDi> yea, "cranking the handle" for a single iteration of your integrator shouldn't be that hard to do by hand to compare against what your integrator is doing
<APlayer> Ohh, calculations by hand are a nice idea
<APlayer> Let me make that
<APlayer> I'll follow the integrator by hand and see where it stops making sense, then
<SnoopJeDi> LOL this SVG icon provider charges $39.99/yr to be able to select a color for the SVG path from their little color picker.
<SnoopJeDi> (in their defense, their icons are also CC-BY so it's not like they're preying on *everyone*)
<APlayer> SnoopJeDi: Notepad++?
<SnoopJeDi> or Inkscape (or CSS directly in the browser, or...)
<SnoopJeDi> trying to find a decent SVG of a pint glass for our happy hour shirts (the bar we frequent the most offered to print some for us on their dime)
<APlayer> CSS support for that is somewhat lacking compared to other properties
<SnoopJeDi> in what way?
<APlayer> Not many browsers do it well
<APlayer> Also, saw this one yet? https://openclipart.org/detail/219885/Beer-Mug
<SnoopJeDi> I haven't, but CC0 is rather helpful for this use case
<APlayer> CC0?
<SnoopJeDi> The license that site requires authors to release their work under, https://creativecommons.org/publicdomain/zero/1.0/
<SnoopJeDi> It's a very concise but very clear public domain declaration
<APlayer> Ah
icefire has joined #kspacademia
<APlayer> Problem #1: I failed to calculate the atmosphere velocity given the position on the planet
<Qboid> [#1] title: Quantity experiment | This seems useful. It might well lead to a switch to C++.... | https://github.com/mockingbirdnest/principia/issues/1
<SnoopJeDi> Found something pretty nice on that site for a starting point, thanks APlayer!
<SnoopJeDi> !wpn APlayer
* Qboid gives APlayer a flammable Bloch anticommutator
<APlayer> !wpn SnoopJeDi
* Qboid gives SnoopJeDi a pyro flying squirrel
<APlayer> Woah! I too want flying pyro-squirrels!
<iximeow> !wpn APlayer
* Qboid gives APlayer a Hittite tombstone
<iximeow> i tried
<APlayer> !wpn iximeow
* Qboid gives iximeow an ellied pre-warp hydrofluorocarbon
<iximeow> still stumped at how cool magnetorquers are
<SnoopJeDi> they are pretty neat
<SnoopJeDi> and happily people are no longer hostile to the reality that you can't really get great 3-axis stability with *just* those, so CubeSat designs that include reaction wheels are pretty common nowadays AFAICT
APlayer has quit [Ping timeout: 182 seconds]
<SnoopJeDi> oh spectacular, that site already includes the URL and original author in the metadata \o/
<bofh> Nice!
awang has quit [Ping timeout: 186 seconds]