@Huy why are you so nice to ppl in the math chat

but not so nice to us

I am very nice to you

then explain the geodesic ball thing

what geodesic ball thing

I don't remember any diffgeo

I'm doing other things now

see, you're mean

boring calculations with $SL_2(Z) \backslash SL_2(R) / SO_2(R)$

wtf

what's the first two things there

special linear group

no

det = 1

why the thingie

backslash

left/right cosets

wtf is a coset again...

$G/H$ is the set of left cosets $\{gH: g \in G\}$

I see I see

well you do that

so $K \backslash G / H$ is just $\{KgH: g \in G\}$

ok

Ah, it's going to be released Monday. Thanks Reddit.

no leak ?

The tracks are on YouTube.

but

?

But I want to buy it

?

wtf

u have too much money

give me some

I got an iTunes gift card for my bday

I didn't

I have no money

Yeah but you can do math

so f u

not really

f u

you an intellectually rich mofo

Welp

If you have some ideas

Ahahah

Fallout already has a thing to buy all the DLCs

The game isn't even out yet

You're lucky I'm bored, Fallout, else I would wait for the price to drop

cant wait till steam xmas sale ?

So I managed to get my algorithm to work. Been working on it since start of October, (which was when I first did hello world in C). Competitors software takes ~ 14 hours to work ... at the minute I get comparative results in ~ 30 minutes :D

@Slereah : "Man, JD is right". Maths is a vital tool for physics. We can't do physics without it. But maths isn't what physics is.

Well got no games to play currently

Either I wait for Fallout 4 or Dwarf Fortress

@AngusTheMan What kind of a computer are you running your software on? And was the competitor's program precompiled, or did you compile it yourself?

@alarge Im running them both on a linux terminal 12 processor station. Compile both of them before running yes.

@AngusTheMan You sure you got all the compilation flags correct etc? Just feels odd that there would be that much of a difference (or well, if it is a very small research group that made the competiting program, maybe not)

@alarge we are part of the same research group and he has an existing procedure for past ~10/15 years. The group would like to try a new method and hence I said I would give it a go!

@AngusTheMan OK, cool, so the entire algorithm is different then.

You can also try this song :

A traditional belgian song

@alarge Yep, completely new approach. We are now hoping that mixing the two together might give an even better procedure. Even though mine is much faster it doesn\t get quite as good a result as the existing method... so perhaps his is better if you look at it from that angle. But we are both really made up :)

@Slereah Every AAA game has that.

@JohnDuffield Is JD usually right?

I can't wait until the AAA industry collapse

when will that be?

Who knows

We're way overdue for an industry collapse

The last one was in 83

order of magnitude?

what happened then?

That was when every fucking company started making terrible video games

Which culminated with ET on the Atari 2600

A game that was created in 6 weeks and they made 12 million copies

ET was great you ungrateful ...

And then they threw away most of them

I'd like to see you do better

And then basically the whole american video game industry collapsed

Well, home console industry

PC and arcade games were fine

Well, ACS bombed in the UK

But Atari never recovered

maybe this is a sign

The current game industry I don't think can last forever

They make very little money currently because the budgets are so fucking high

^

AAA games cost as much to make as movies

Do you know Review Tech USA?

He's been saying this for a while now

No

I've read similar articles, though

Looking at the profit margins, DLC and microtransactions are what's picking up the bottom line.

Yeah

Fucking GTA had a 500 mil budget

Making shittier and shittier identical games for giant budgets is probably not very sustainable

True, it made that back in a few hours.

Dude, you can't say GTA5 is a shitty game.

C'mon.

@AngusTheMan Cool. Which of the algorithms is more parallelizable? And could you maybe offload them onto the GPU? Is the older code well written, or would it potentially benefit, performance-wise, from a rewrite?

How many of those games are somewhat decent and worth buying, though

Bethesda games.

What, all of them?

Have you played the Elder Scrolls Online

Even Skyrim was a bit disappointing

No.

I was considering buying ESO when it was on sale.

I've heard they patched the shit out of it.

I played during the beta

I'd say don't bother

The beta was crap!

They've massively overhauled it.

But I have no one to play it with.

Eh

No one plays it.

Fallout 3 was a huge letdown for me as someone who used to spend too much time playing the first two.

If it becomes free to play maybe I'll give it a go

I finished Fallout 3 in 3 days

A bit short

Although I'll give them this

Putting in a radio was a smart move

It is quite useful to have in a game where you wander a lot

@0celo7 : Yes, I'm usually right. I make a few mistakes, but not many. And if I don't know, I say so.

@alarge My one is written for parallelisation specifically. However at the minute I have shared the job equally amongst processors and keep having to combine them back etc periodically. I really don't know enough about GPU coding to do it as I have been focused on implementing MPI. I think the fusion of both would be the quickest perhaps. I think the existing code is optimised as best as could be,IMO and is quite well written. It's just the nature of the procedure that isn\t great for parallel

Anybody want to talk physics?

What kind of physics John? :)

Woo

Let's talk about woo

I finally found the difference between compactly generated and compactly constructed

Woo

Kip Thorne has a nice paper on compact generation

Maybe I should answer my own post

@AngusTheMan Well there are the things close to the metal, like if you use AVX efficiently etc that could in principle give you a free 4-8x boost.

@AngusTheMan : anything you like. Electromagnetism, particles, gravity, cosmology. You name it. How about we kick around some unsolved problems in physics?

@alarge You have given me a lot to read up on there! Cheers :)

@JohnDuffield Do you like material science ? What is your favourite sort of area?

By your posts you seem to like gravity I guess?

@AngusTheMan If that stuff is new to you, start with compiler flags though (and ideally, experimenting with different compilers).

@JohnDuffield I would reconsider the "usually" since you're the most downvoted frequent user.

(Or one of)

Just trolls and naysayers

@alarge I don't even know what we are using if I'm honest! I think it is a g++ but I may be mistaken. I think it is old.

@AngusTheMan : my favourite stuff is fundamental physics. Gravity is just one topic amongst many.

It's pretty hard to make a list of causality violating spacetimes to classify because most of them don't have names

The author just goes "Hey look at that spacetime, it is weird, let's laugh at it"

Allegory :

@TanMath I understand that if you find that insulting, your skin is too thin.

@AlfredCentauri I like you.

Just had a PB&J sandwich. Was the first in...many years. I think I'll have another one. This is a magical little bugger of a sandwich.

@JohnDuffield I don't really know that much about that area if I'm honest :/

@AngusTheMan You really should upgrade and make sure you are compiling to your architecture (it'd be a horrid waste if you're not using AVX or even SSE)

The reason you want to upgrade is because compilers have gotten better at autovectorization

You should also experiment with clang, and even possibly with polly, a polyhedral optimizer for clang.

@Slereah : causality-violating spacetimes are a mathematical fantasy. We live in a world of space and motion. We model this using spacetime and worldlines and light cones. But spacetime is a static "all times" block-universe. There is no motion through it. You don't move up a worldline, or around a closed timelike curve. And there is no way you can move such that everything else in this universe not just moves back to where it was, but never moved at all. Hence time travel is science fiction.

Or to be a little blunter: time travel is woo.

@AngusTheMan : ditto for me and materials science.

@JohnDuffield : Please do not talk to me.

@alarge Oh boy I haven't really heard of these things before! I am 100% sure we are not using any of them though. How would I implement these things? Do you have any sources for learning up about them?

@AngusTheMan Well start with "man gcc" and read through what the flags do.

Bascially you want to use -march=native (and -mtune=native) if you're compiling on the computer you will be running on.

@Slereah : I'm not really talking to you. What I'm really doing is explaining to the other guys why the stuff you prattle on about is woo.

@JohnDuffield We all know by now not to listen to you :)

And then I guess you can read about the processor you are running on Intel Ark and Wikipedia and look up the keywords there (if it is even relatively new, it'll support AVX etc)

ugh, I see some issues in the transcript. @TanMath, don't kick people out of the room for holding normal conversations. @AlfredCentauri, no need to be so harsh in response.

and @0celo7, @JohnDuffield, and now apparently @Slereah too, I seem to remember telling you to keep your arguments out of here. (I'm not talking about when I suggested taking your physics discussions to another room.) So please stop.

I do not have any arguments.

I have stopped talking to Duffield a while ago.

So, who wants to talk physics?

@alarge I really appreciate you putting me on to this! I will let my colleague know and maybe he can get his to run a little quicker! Compared to MPI how much quicker do you roughly expect these to go with your recommendations?

@AngusTheMan No, you'll still be using MPI. What you'll enable is a larger instruction set. You do know of assembly, right? So if you don't pass flags to the compiler, it will assume that it can only emit a subset of the instructions your CPU suppors (for maximum binary portability).

Which means that vector instructions such as SSE and AVX will not be enabled. Which is to say that your CPU will be very underutilized.

Ah, the nuclear holocaust

@JohnDuffield Let's talk about string theory

@0celo7 : If you like. What are the strings made of?

dunno

@alarge I will have a read about these different things, at the minute it all seems a bit too complex for my introductory knowledge. However I'll definitely come back to you! :) I have been more focused on getting the algorithm off the page and working. I think next I will look into the optimisation of it all!

@JohnDuffield why should it matter, anyway

@0celo7 : I think string theory started off with a good sentiment, in that a particle like an electron isn't really a point particle. But then things somehow went downhill, and now there are people who say gravity is weak because it leaks off the brane.

@JohnDuffield What's wrong with that

@0celo7 : because explaining something using something that isn't explained is no explanation at all.

@JohnDuffield a theory can't tell you what the fundamental objects are

@0celo7 : What's wrong with that is that it demonstrates no understanding of gravity.

@DanielSank en.wikipedia.org/wiki/Stochastic_cooling for one.

@dmckee Some people from my group are going to ANL in a bit.

@0celo7 : it ought to tell you what the fundamental objects are, otherwise it isn't a fundamental theory.

@JohnDuffield I'm not saying it's fundamental

I'm saying it ought to be. See this answer and always remember this: when a seismic wave runs through the ground, the ground waves. When an ocean wave moves through the sea, the sea waves. So when an electromagnetic wave moves through space, space waves. So the "strings" are made of space. Only they aren't strings, because space is like this ghostly elastic solid. Look:

See that shear stress?

What does that have to do with anything?

@0celo7 Your cat is so cool.

@obe uh, thanks?

What?

Please do this mathematically.

You think I'd be old enough now to know that eating the entire cake was a bad move. But no, I still face these issues :(

@0celo7 : do what mathematically? Tell you that the photon is a singleton linear-propagating electromagnetic wave where E=hc\λ? Or that c=√(1/ε₀μ₀), and that this is akin to the mechanics expression v = √(G/ρ), where G is the shear modulus of elasticity and ρ is density?

I have to go I'm afraid. I'll be back later. Meanwhile: have you ever played a guitar? Take a look at some pictures of the electromagnetic spectrum. What's always the same regardless of frequency?

@JohnDuffield I want you to show me how you can derive Einstein's equations, the Dirac equation, etc. from "spacetime is a lattice of strings".

Damn

Seems my luck getting answers on the Math SE is even worse than on Physics

Can I import Math SE questions on Math Overflow

Seems to not be available unlike PO

MO is only for smart ppl

Hm, is my plane with a handle question Math Overflow material

I am wondering

For a moment I thought the day was flying by since it's 5:30

Then I remember that I woke up like two hours ago

Oh wait, 4

Time does fly

Hey hey ACM

Now that I know what is a compactly generated causality violating region

I must wonder

What is a non-compactly generated one

I can only think of examples where it's compactly generated

Or there is no Cauchy horizon at all

Or, for Kerr, I suspect the Cauchy horizon is due to the singularity

Since I still don't know what "compactly generated" means, I can't help you there :/

Also, I need another weekend after this weekend.

Well the Cauchy horizon is a null hypersurface

Hence it can be entirely generated by null geodesics

@ACuriousMind what

Compactly generated means that the past of those null geodesics remains in a compact region

The so called "fountain"

@0celo7 I need a weekend to recover from this weekend. :P

Basically the Cauchy horizon starts as just one closed null geodesic

@ACuriousMind ah

And then all other null geodesic emerge from it

@Slereah I see

Thusly

Hm

I guess a simple example might be like

Take Minkowski space

Well, for one, it's going to be hard to draw a non-compactly generated one, right?

Cut it off at $t=0$

Then paste Godel space after that point

I guess that would be a simple example

If a bit discontinuous

But I have to follow the Hawking cookbook for spacetimes after all

@Slereah which one is that

The one I just pasted :p

from?

The Hawking-Israel book

@Slereah Well...I think there are very few manifolds at all that aren't obtained by cutting and gluing, so that's not a very specific recipe :P

link?

amazon.com/General-Relativity-Einstein-Centenary-Survey/dp/…

@ACuriousMind Prime manifolds aren't and there's an infinite number of them!

In 4D at least

937 pages

But it would be nice if the manifold was $C^2$ at least

It's not a bad book

A little bit of everything

Although a little bit dated

@ACuriousMind pls i.gyazo.com/105b69cd95e27ef6e5a4b4bbd3a7ea0d.png

what does the last sentence of the first part mean

and why is it relevant

What is even $B$

and $\rho$

@0celo7 What do you not understand about it?

@ACuriousMind are the two $\rho$s there the same?

certainly not

@0celo7 Yes, I think so

@Slereah Riemannian ball of radius $\rho$

@ACuriousMind Ted Shrifin disagreed. Why do you think so?

@0celo7 Because it is worded that way. It might be that they shouldn't, but as it is written, they are the same

Also, the way the corollary statement is worded, it also says that $\rho_0$ is the same for every point.

It's $\exists \rho \forall p$, not $\forall p \exists \rho$.

Here's how I would prove it. From an earlier theorem, we know $B(p,\rho)$ is a normal neighborhood for some $\rho>0$ $\forall p$. So take $\cup_{p\in M}B(p,\rho_p)=M$ and use compactness to get a finite subcover. Then take $\rho_0$ as the smallest such $\rho_p$ in the subcover.

No use in showing it to me because I don't know all the theorems you/the book are referring to

What?

The only theorem you need there is the existence of convex normal neighborhoods.

@0celo7 That does not show the statement you want to prove. You did not show that $B(p,\rho_0)$ is a normal neighbourhood for every $p$.

And that is precisely why the $\rho$ in that proof in the book has to be the same.

@ACuriousMind But $\rho_0$ is the smallest such $\rho$

@0celo7 No, it is the smallest in the arbitrary finite subcover you chose

@ACuriousMind Why you need another weekend? You been partying have you :p ?

@AngusTheMan Yes :D

@ACuriousMind argh

what now

@AngusTheMan Yes, that captures my weekend perfectly.

I'm not in that picture because I'm probably lying under the table, though

you are not there because you were not invited

@ACuriousMind haha oh I bet. I went out for "quiet chem drinks" on friday... good lord

@ACuriousMind how do I prove this now :(

@0celo7 I don't know at this time

@AngusTheMan Is that code for you meeting to take all the drugs you cooked in your labs over the week? ;)

I doubt ACM watches that

Watched the first two seasons, got bored.

It got very boring i though

thought

@DanielSank So... who was the coworker?

sort of like game of thrones ... but somehow I was compelled to keep watching

@ACuriousMind called it

I still like GoT

Breaking Bad was good but I wish they got rid of the family drama

It was the most boring part

His whole family should have been beheaded by the mexican cartel

@0celo7 Oh really?

I have a feeling @Danu is reading the transcript :P

@ACuriousMind Oh really?

@Danu No, I was just joking

@ACuriousMind I'll let it slide this one time...

but don't let me catch you "joking" again!

I explicitly worked out the stereographic projection for once

:)

@Danu 'Kay, from now on you will only catch me masturbate.

@ACuriousMind Oh I am very excited...

@Danu Yes

Overdoing it much?

@Slereah I posted the Skyrim version of that once.

Why would you post a version that is not this one

It is perfect

@Danu I mean, yes to this one.

@ACuriousMind Why are people starring this?

@Danu Beats me

off?

@0celo7 Don't explain my jokes :P

@ACuriousMind then help me with this geometry

@Danu : What manifold is your avatar

@Slereah why don't you help me D:

Help you what

I'm not a topologist man

I cannot topology

this is geometry

ACM is hung over

@DanielSank how is while different than if

@0celo7: careful with your language in math chat, the kids there flag everything

it's a loop

can you make a while loop into an if loop?

If is not a loop.

you need goto for loops :p

@Huy what

@ACuriousMind hmm

derp

so how about them geodesic balls, @ACuriousMind

what I said @0celo7

I got suspended for "are you a troll?" once

while(p) is equivalent to 10: if(p) ; goto 10;

@0celo7 I don't know a proof, and I'm not in the mood to try and think of one, sorry