The Nineteenth Byte

« first day (2378 days earlier) ← previous day next day → last day (2479 days later) »

totallyhuman

10:13 PM

dunno

MD XF

Anyone good with C appearing to behave perfectly but really not?

Tobi

@trichoplax when I have 5k objects of the same width, it works well. when i have 5k small objects and oone massive one, as you said, it starts to lag again. I watched a video on quad trees and it has dynamic box sizes, wouldnt that fix this problem?

ATaco

@HyperNeutrino we tried to solve this already, didn't get a solid answer

MD XF

@HyperNeutrino Ha

ATaco

And that was just for a=1

Riley

10:25 PM

@MDXF What's it doing?

MD XF

@Riley Lemme check, I got it to quit seg-faulting but it's not doing anything useful either, brb

Ha. Once again

As soon as I ask someone for help I figure it out. I really need to buy a rubber duck

For some reason I was thinking fseek was like a goto

New Main Posts

Q: When do the lights blink?

Imagine you have two lights. These lights blink on and off at a specific rate: Light 0: Delay 0ms and then blink every 1000ms Light 1: Delay 500ms and then blink every 1000ms Let's simulate these lights for the first 2000ms: 0ms: Light 0 on 500ms: Light 1 on 1000ms: Light 0 off 1500ms: Li...

Riley

@MDXF It is if the 3rd argument is SEEK_SET I think.

MD XF

No I thought it changed the position of the C program it was in >.<

It's like a goto for the file. It's not like a goto for the actual source code

Riley

Oh... Yeah, it doesn't do that.

MD XF

10:33 PM

Yeah. I'm incredibly brilliant. :P

I've been ripping my (small amount of) hair out over this for over a day now

trichoplax

@Tobi I recommend you implement both. This will give you a much better idea of the good and bad points of each, and it will also give you specific measurements for your particular case.

You might even find that a hybrid approach works best, rather than trying to put all objects in the same acceleration structure

ASCII-only

10:53 PM

@Downgoat ???

@MDXF idk maybe

@MDXF wth why, function calls can't do that

I guess macro calls may be able to though

MD XF

I know

ASCII-only

@flawr Also MIT, Apache and GNU criteria?

MD XF

I was just confused aparently

Tobi

its too laggy

only 80 fps

for 10,000 only

Maltysen

@Tobi have you tried turbo.github.io

minxomat made it

you can do you movement operation in parallel on the gpu on all objects

ASCII-only

11:04 PM

> only 80 fps

Tobi

whats the browser support?

ATaco

That webpage looks too good to be a .github.io page, I don't trust it.

ASCII-only

@ATaco It's localhost

ATaco

I'm referring to turbo.github.io

Tobi

omg this is going to be amazing

when i combine quadtrees and turbojs

1,000,000,000 particles

ATaco

11:14 PM

EZ fix, just write the entire thing in Objective C.

Tobi

need for browser

ATaco

Write the entire thing in Objective C, then combine with Emscripten

Tobi

isnt the speed boost for asm.js minimal

does turbo js us web assembly - or what ever its called now?

ATaco

Honestly I'm absolutely clueless, I was just playin' around.

MaybeLater spends 1.7 seconds as a base.

(Although on TIO it spends 0.1 seconds)

ASCII-only

@ATaco ???????

@Tobi no? it says it uses gpu (via canvas) and that's what it uses

ATaco

11:17 PM

It takes a step to parse the tokenizer. I really should precompile it.

Tobi

via canvas?

ASCII-only

@Tobi wasm doesn't offer anywhere near the speedups using the gpu offers

@Tobi yeah? it's the only way to access the gpu (webgl)

Tobi

if they have no GPU, is there any speed up?

ATaco

@ASCII-only This clearly isn't in a format Javascript is very good at reading.

ASCII-only

@ATaco uh yes it is it's called JSON5 and JS is really good at reading it

ATaco

11:19 PM

That part of it is, but look at the strings themselves.

ASCII-only

yeah what about it

Tobi

no speed up?

if no GPU

ASCII-only

@Tobi not really, but most devices have at least integrated gpu

ATaco

They represent the sorts of token that the tokenizer/parser hybrid can match, and before execution they need to parse them into a more flexible format.

Step Hen

@ATaco why not store them in actual JSON?

ATaco

11:20 PM

Easier for me to write them this way, I just need to precomile it, and I fix the problem.

Step Hen

like more layers

ASCII-only

@ATaco -1 disgusting, use a parser generator like jison/pegjs/nearley, or store them in actual JSON like Charcoal stores them in PyON (well idk what to call it)

ATaco

I don't do anything unless I can find a way to reinvent some wheel.

ASCII-only

@ATaco uh you know you can just change the format right?

@ATaco well yeah this is the wrong way to reinvent it, Charcoal uses its own parser and format too you know

ATaco

Yes, but then I need to rewrite my script that uses the tokenizer, which is 90% of the script.

@ASCII-only No such thing as the wrong way in recreational programming.

Zizouz212

11:24 PM

Sup folks

ASCII-only

@ATaco well i guess if you don't care about runtime

ATaco

@Zizouz212 o/

@ASCII-only I thought I made that apparent already :P

Zizouz212

@ATaco I've finally gotten a real app to program in :P

ATaco

What would that be?

Zizouz212

Idle -> PyCharm ;D

ASCII-only

11:29 PM

@ATaco 0/10 seriously you just wasting TIO's CPU time

ATaco

Real time: 0.107 s
User time: 0.090 s
Sys. time: 0.013 s
CPU share: 95.56 %
Exit code: 0

Apparently I'm not.

(Just the CPU share)

RProgN2 uses more CPU time

Real time: 0.311 s
User time: 0.231 s
Sys. time: 0.064 s
CPU share: 94.67 %
Exit code: 0

ASCII-only

@ATaco O_o why does it not take 1.7s on TIO

ATaco

(My testing computer was a $300 laptop)

Also, Jelly uses more resources than both of them

Real time: 0.887 s
User time: 0.760 s
Sys. time: 0.077 s
CPU share: 94.38 %
Exit code: 0

My most efficient language on TIO is TacO, which fails in less than 0.1 seconds on a blank program.

MD XF

Rubber Duck Debugging

For people who find rubber duck debugging to be a very effecti...

New Main Posts

Q: Champernowne up to me

Inspired by this question. Champernowne's constant is an infinite decimal number that consists of "0." followed by all natural numbers concatenated together. It begins like so: 0.123456781011121314151617181920212223242526272829303132333435363738394041424344454647484950, etc. The digits of Chame...

code-golf number kolmogorov-complexity

Downgoat

11:55 PM

PSA: if you are going to post ? or wat those both have 0 semantic meaning so you can just leave them out with same amount of information being communicated.

totallyhuman

wat

Downgoat

@totallyhuman hilarious

totallyhuman

yes very

ATaco

It's almost as good as all these Maybe Later puns I've been hearing.

totallyhuman

no it's much better

and i still you think you practically brought it onto yourself

Downgoat