The h Bar

« first day (1132 days earlier) ← previous day next day → last day (4095 days later) »

user54412

07:00

Ahhh, that amplituhedron paper finally came out: arxiv.org/abs/1312.2007

Brandon Enright

@ChrisWhite I see N=4 symmetry mentioned all the time but I have no idea what it is. Is there a wiki article or something else I can read to give me at least a basic layman's understanding of it?

user54412

@BrandonEnright you're asking the wrong person

Brandon Enright

@ChrisWhite is there an english term / name for it that I can at least google?

n=4 doesn't help much

user54412

I wish I knew - when physicists start speaking in groups and symmetries, I get quite confused myself

Brandon Enright

@ChrisWhite That makes two of us :-)

Which is too bad because my hobby involves a lot of (simple) group-theory.

user54412

07:20

\begin{anecdote} Once upon a time, I was taking a particle physics course. It was pretty low-level as such things go, giving us rules for Feynman diagram calculation (I'm really great at that now) but giving no explanation for what we were doing.

user54412

One day the prof did some hand waving with representations (basically a Clebsch-Gordan transformation, I later realized). I asked how he got the result, and all he said was, "That's just group theory. Moving on..." At this point I had two solid years of abstract algebra under my belt, and nothing he was doing resembled group theory in any way, either because it wasn't or because the notation was obfuscated.

user54412

From that day forward, whenever a physicist hand waves something away as "just group theory," I immediately stop listening, since I already know I won't get a deep understanding from what they're saying. I figure there's a critical disconnect between math and physics here. I learned the math way, and I'm too old and tired to learn how to bridge that gap. One day, perhaps, someone will explain it to me. \end{anecdote}

Nick

07:51

@chris: Can I put that in my book or will you put it in yours? Either way, it has to go into a book.

7 hours later…

Physics Meta

14:32

Q: Error with the form "Your Answer"

I would like to answer this question: Applications of the Spectral Theorem to Quantum Mechanics but I cannot: I insert my answer in the form "Your Answer", I click on "Post Your Answer" and then after a while (even 5 minutes) I receive an error message. Could someone help me? Thanks in advance....

support

3 hours later…

Kyle Kanos

17:12

Today is a good day!

Crazy Buddy

@KyleKanos Why so?

Kyle Kanos

@CrazyBuddy My hydro code collection is complete!

And the bestest one is running nicely on my computer right now

Crazy Buddy

@KyleKanos hydro what?

What's that?

Kyle Kanos

@CrazyBuddy hydro = hydrodynamics. Eulerian fluid stuff

Crazy Buddy

No, I meant "hydro code" - I thought that's some kind of software :D

Kyle Kanos

17:15

Ooo.

No, I just have a collection of hydrodynamic codes. And I finished getting all the (relevant) good ones

AstroBEAR, MPIAMRVAC, CLAWPACK, Athena, Pluto, etc

Crazy Buddy

ah... You're an astrophysics Ph.D fella. Right?

Kyle Kanos

Yep :D

Crazy Buddy

Fluid dynamics plays a role?

Kyle Kanos

Yep. Highly ionized plasma = fluid

Crazy Buddy

oh, that one... certainly is :)

Why are you collecting? Any paper publishing or something?

(projects are meant for me...) :P

Kyle Kanos

17:18

Mostly collecting b/c I want to write my own good one in the future and want to see what is good/bad about each one to make a better one.

Crazy Buddy

oh, nice... ;-)

Kyle Kanos

But I am currently working on a project that requires 2d r-phi coordinates with adaptive mesh refinement and there's only 2 codes that do that, and one really blows chunks to use, the other is pretty nice but restrictive in availability

Crazy Buddy

ah...

@KyleKanos Don't use much technical terms, I won't be able to grasp it :P

Kyle Kanos

Sucks for me that the bad one is written in Fortran (what I'm proficient in) and the good one is in C (what I'm not so good at)

@CrazyBuddy Hmm...not sure how I can re-word that without using something related to "blows chunks"

Crazy Buddy

What? Not good in C?

@KyleKanos nah, just kidding :D

I thought that C was much better than older ones...

Kyle Kanos

17:22

Yeah, I've spent 8 years mastering Fortran 90+ and now have to use C for this project, which I haven't really used since 2004ish.

C is as old as Fortran

But it's mostly syntax differences, e.g. DO loops vs FOR loops

Crazy Buddy

@KyleKanos 8 years? That should be exaggerating :D

Well, honestly I dunno other programming any better than C/C++ :)

Kyle Kanos

@CrazyBuddy I can give 2 reasons why Fortran 90+ > C/C++: (1) elemental functions and (2) C = A+B works for any size array (assuming all three have the same shape & dimensions)

But 8 years isn't an exaggeration :(

Crazy Buddy

hehe :D

Ooops... I gotta go now. Dinner time :D

Kyle Kanos

Dinner? It's 12:33 pm :D

Crazy Buddy

@KyleKanos Nah, it's just 11 here ;-)

I have such unusual times :P

user54412

18:18

Anyone else catch this? Nobel winner declares boycott of top science journals

Kyle Kanos

@ChrisWhite Wow...interesting

1 hour later…

user54412

19:37

@KyleKanos cool

user54412

actually, if you have any opinions about athena, good or bad, I'd be interested

user54412

I'm working on developing a new version of it for a specific problem, and I'm curious how it stacks up to other codes in terms of useability, etc.

Kyle Kanos

@ChrisWhite My only problem with it is its lack of AMR

Now I don't really use it that much, but I have borrowed its curvilinear PLM for AstroBEAR

user54412

@KyleKanos yeah - AMR is hard

user54412

I hope to avoid it as much as possible ;)

Kyle Kanos

19:43

But I think the problem modules are pretty straight-forward enough to implement your own problem without much difficulty

@ChrisWhite I concur, which is why I use someone else's code that already has AMR enabled, all I have to do is write how many mesh levels I want

Some people, unfortunately, take the "easy" way out and integrate Boxlib, Chombo, or Paramesh into their codes. This just kills the ability to parallelize efficiently

user54412

I think that's why Athena doesn't have AMR - I know Jim Stone wouldn't put something into the code that wouldn't work well in parallel, and so the only option is to sink a lot of effort into making it work right

Kyle Kanos

Yeah, it can be a pain. I think Udo Zeigler spent like 5 or 6 years developing his AMR for NIRVANA. But it's pretty good. And it works in all coordinates

Romain Teyssier's Ramses has its own AMR engine, he spent a few years developing that. AstroBEAR has their own engine, they spent 3 years and 4 grad students & 2 post-docs to get that done that quickly (but they were also basing it heavily on Chombo).

Actually, I think Jim hasn't put the effort into the AMR because he's spent the last couple of years writing a hybrid PIC (kinetic ions, fluid electrons) code.

Wait...aren't you at Princeton?

user54412

@KyleKanos yep, Jim is my thesis adviser :P

Kyle Kanos

ooooo. Lucky duck.

So he's having you work on Athena & not Pegasus?

user54412

20:02

he's doing a lot of things simultaneously - my niche doesn't involve newfangled plasma microphysics, probably for the best

Kyle Kanos

@ChrisWhite I do recall you asking for a reference on GR-hydro....I suppose that's the next step for Athena?

user54412

yeah, there is definitely a trend these days with merging GR and MHD, and we figure Athena should be part of that

Kyle Kanos

Well, my $0.02: don't be afraid to ask someone for their code, but make sure you mention what you intend to do with it

2 hours later…

Anonymous

21:48

Hello?

hwlau

@Anonymous hi there

ManishEarth

22:36

@ChrisWhite Looking at the title I assumed it would be a mathematician. Surprised it was a Medicine guy. Then I remembered that there is no Nobel for math.

1 hour later…