The Nineteenth Byte

New Posts

00:11

0

Q: Give the best Chaitin incompleteness bound

This was originally a pure mathematics question, but I think I've got the best chance for an answer here. The Challenge For concreteness, consider Peano Arithmetic (PA). For some language L of your choosing, construct a computer program of length at most N bits in L such that: If it terminates,...

OldSandboxPosts

01:00

Sandbox posts last active a week ago: Minimum Escaping open-ended-function

l4m2

03:42

How is take input (x,y) meaning y-th smallest integer whose answer is x forbidden by default?

Bubbler

04:39

139

A: Loopholes that are forbidden by default

Adding input or rules which weren't explicitly mentioned in the challenge There was recently a case where an answerer claimed Nowhere does it say the program can't (also) ask the user what the [result] is. Adding additional input or new rules, because they haven't been explicitly forbidden...

Unrelated String

that sounds like a pretty non-trivial equivalence lol

for discussion on something similar check the recent meta thread about decision-problems versus sequences

mousetail

05:32

@lyxal Thanks

New Posts

06:45

CMC: Given a string as input, return how many characters in that string appear in the word "butter". Case insensitive so B and b count as being in the word butter.

lyxal

@NewPosts Vyxal, 46 bits2, 5.75 bytes: ⇩‛₅ġ↔L (vyncoded)

Bubbler

Uiua, 16 bytes: /+∊∶"buterBUTER"

the butter part is not really golfable I think

Unrelated String

07:04

/+∊∶⊂+32."BUTER" ties

also looking at the docs i like how they use "pervasive" for vectorizing

gets the point across without the risk of confusion with hardware vectorization

Bubbler

that word comes from APL terminology

aplwiki.com/wiki/Pervasion

Unrelated String

mmm

Dannyu NDos

Never knew New Posts could propose a CMC.

lyxal

Yeah it's a feature that's been around since mid July

mousetail

07:19

So it seems SE staff learned about the layoffs via the meta post before even being told personally

lyxal

Seems that way yeah

Unrelated String

wow

mousetail

Kaba Kun read the meta post in the night, had a (I imagine) very stressful night, then only learned that they where didn't have a job in the morning because they couldn't log in to anything

Oh Kaba Kun == Catija nvm

lyxal

Yeah :p

mousetail

Me trying to quickly understand and summarize thousands of discord messages

V2Blast is also no longer CM

Neil

07:44

@NewPosts Retina (any version), 9 bytes: Try it online!

@NewPosts Charcoal, 12 bytes: ＩΣＥＳ№bertu↧ι

New Posts

11:16

CMC: Given a number, convert it to base 6, but the digits are butTer. So like b = 0, u = 1, t = 2, T = 3, e = 4, r = 5

mousetail

Have another butter related CMC?

lyxal

@NewPosts Vyxal, 9 bytes: `butTer`τ

Ginger

@lyxal *cough*

well, the downloader's still running

we're on room 202

25 million messages!!!

mousetail

11:32

How much storage is the backup using currently?

Ginger

not sure, lemme check

2.4 gigabytes

which means that the entirety of chat is probably around 5 gigabytes

mousetail

You could fit the whole thing in your RAM

Ginger

yup

it's now working on the Velodrome

mousetail

Are you going to do SO and meta too?

Ginger

12:20

probably

Neil

13:23

@NewPosts Charcoal, 8 bytes: ⍘ＮbutTer

New Posts

14:08

0

Q: How many umbrellas to cover the beach?

You work at a beach. In the afternoon, the sun gets quite hot and beachgoers want to be shaded. So you put out umbrellas. When you put out umbrellas you want to shade the entire beach, with as few umbrellas as possible. Umbrellas come in many sizes. However, larger umbrellas are susceptible to be...

code-golf array

New Posts

15:31

0

Q: Expand Emmet HTML Abbreviations

Emmet is a text-editor and IDE plugin for writing HTML code using much fewer keypresses. You might think of it like a code-golfing HTML preprocessor. Its syntax is based on CSS selectors. Your challenge is to code-golf an implementation of Emmet. Given some Emmet code, output the resulting HTML. ...

code-golf parsing code-generation

Simd

16:37

Which languages are good for writing fast code? C, C++, rust, Julia and .....?

Fortran....

Rydwolf Programs

Assembly :p

Simd

@RydwolfPrograms true .. Any others?

I don't know swift at all for example

Or C#

Rydwolf Programs

Probably not C#

Simd

@RydwolfPrograms thanks

I feel there must surely be another one

There are so many programming languages out there

I forgot nim!

Maybe rpython?

I wonder if this could be a tips question on main

mousetail

17:08

@Simd Zig?

CUDA?

I don't think nim compiles directly to machine code

Simd

17:26

@mousetail interesting! I don't know that at all

@mousetail is that a language?

mousetail

Yes, though also there are other langauges that compile to it

Simd

@mousetail it's very fast. We used to have fastest-code answers in nim here

In the good old days

The Nim compiler supports mainly two backend families: the C, C++ and Objective-C targets and the JavaScript target. The C like targets creates source files that can be compiled into a library or a final executable.

mousetail

A level of indirection will typically make it slightly slower since it makes optimization harder

Simd

@mousetail is it only fast for GPUs?

@mousetail that would make sense

mousetail

It's basically C for GPUs

GPUs are faster than CPUs

Though it depends on the challenge if you can effectively use the parelelism

GPUs are also better at floating point operations than integer operations

Simd

17:32

@mousetail I thought was much more subtle than that

mousetail

It's complex

Simd

I have seen gpu projects where they failed to speed up the code

mousetail

They probably couldn't parelelize enough

or had to transfer too much data

Simd

Could you do this on a GPU? codegolf.stackexchange.com/q/265857/116117

@mousetail right. That seems a big problem

I wish we had some GPU challenges here

mousetail

Not easily, because integers and loops of uneven length

Simd

17:35

@mousetail right. So really it is not easy for a novice to guess when a GPU will help

mousetail

Yea

Simd

@mousetail could you simply run the whole process many times in parallel on a GPU? I am guessing not

mousetail

My answer already runs it on 8 threads in paralel

Simd

I don't understand why it only gets to 4 in that case!

mousetail

No, in GPU all theads share a state so they must execute the same code as the same time. If one thread takes longer all the others will need to wait before being able to advance to the next step

My answer would get less than 1 without multithreading

Simd

17:54

I just remember implementing Kalais method in python/sympy once and I am sure it was a little faster than that

Single threaded because it is python

mousetail

Maybe my answer is just inefficient then

Have you ran my answer on your PC?

Simd

@mousetail I'll do that tomorrow. I have not been able to yet

I am wondering if I can usefully add another bounty

Do you know what "A slightly tighter envelope" means?

Seggan

18:10

@mousetail wait really?

TIL

@mousetail TlLDR divergent branches r bad

mousetail

Relatively

Seggan

i found Futhark to be a really nice lang for GPU stuff

Simd

@Seggan looks cool!

Simd

18:25

We forgot the mythical mojo!

mousetail

Mojo is going to be slower than other options

Ginger

🤣🤣🤣 💯🔥

Unrelated String

^^
making performance improvements over python is a pretty low bar lol

Ginger

gonna set up a bot that sends those emojis every time Mojo is mentioned

Rydwolf Programs

19:03

Technically you're not allowed to do that :p

Ginger

~~technicalities are merely suggestions~~

/srs I know, but it'd be funny

Simd

19:14

@UnrelatedString that is true!

In principle Fortran should be the fastest of all

mousetail

Fortran shouldn't be slower than C or rust, they all don't have costly abstractions

Rydwolf Programs

@Simd I wouldn't necessarily assume that, since C's used so much more, it might have more invested into optimizations

I do believe the main Fortran back-end is GCC-based tho so maybe they have the same optimizations

mousetail

I'd expect any language that builds directly to assembly and has few abstractions to be equally fast

Rydwolf Programs

I think you're underestimating the importance of optimizations tho

It was ages before C was as fast as handwritten assembly wasn't it

mousetail

In theory you can always optimize yourself

Ginger

19:21

isn't that a self-help book?

2

Rydwolf Programs

@mousetail Maybe, but there's some stuff you'd never be able to do without an asm macro

mousetail

Optimizations is more about being able to write more elegant code without worrying about performance than about making already optimized code even faster

Rydwolf Programs

E.g., vectorization. SIMD stuff gets massively complicated, and no language I'm aware of has syntax that would allow you to take advantage of even a fraction of that complexity without compiler optimizations

mousetail

Most languages have intrinsics for that

(which are just assembly with extra steps)

Rydwolf Programs

But if you fall back to assembly or intrinsics for individual instructions, I feel like that's kinda cheaty

Sort of like people saying Python is fast because it can call out to C++ code

mousetail

19:27

I think different opinions are just a different perspective of the same thing

New Posts

0

Q: Write a proof checker for first order logic

This originally came from Give the best Chaitin incompleteness bound, but I realized that there was a fragment of this problem which still potentially gives a good coding challenge. First Order Logic With Equality First, I will need to fix an encoding of first order formulas ('For all x there exi...

code-golf parsing logic

Simd

19:56

@RydwolfPrograms it would be really interesting to see benchmarks. I do know Fortran was designed to be optimizable unlike any other mainstream language

@mousetail the problem with C is that it makes it really hard to optimize

It doesn't have opaque pointers

I mean C does have opaque pointers

Ginger

20:12

@Simd citation for this?

(the part about "unlike any other mainstream language")

mousetail

@Simd You can use the restrict and const keywords to improve optimization of expressions with pointers a bit, but mostly you have to do it yourself

Seggan

@Simd Assembly would like a word with you

Simd

21:04

@Ginger it was a challenge to name one :)

@Seggan run away!

ATaco

22:35

CMC: code-challenge Return the sum of all code-points in a string. Score is the sum of bytes in your solution, lowest wins.

lyxal

@ATaco Vyxal ad, 1 byte, score 67: C

ATaco

Vyncode is slightly better for 22.

lyxal

that's an interesting question

how are fracbytes scored?

and then how are logbytes scored?

Also, if outputting as a character with the corresponding chr value is allowed

@ATaco Vyxal 2.4.1 Ks, 0 bytes, score 0

ATaco

~Recinded~

lyxal

I forgor the flag that sums it :skull:

big brian moment

Dannyu NDos

22:47

I've never succeeded in finding a good quine-related challenge so

ATaco

Fractional bytes are easy, concatenate and re-interpret. Log bytes are designed by the devil.

Ginger

@ATaco ---strike---

ATaco

My formatting was not an error, but intentional style.

Ginger

suuuuure

ATaco

The world is trapped and only I am free

~~I can do anything~~

lyxal

22:49

you [clown around town]

~~stop targetting the goat~~

Ginger

~~that's an awfully long-winded way to say "I forgor 💀 how strikethrough works"~~

lyxal

23:13

for a second, I missed the l in clock

a very unfortunate typo

ATaco

os.rooster()

Neil

23:30

@lyxal a cock-up

2

Seggan

23:41

@ATaco Fig, S \$\log_{256}(96)\$ bytes

Bubbler

CM(ega)C: Repeat to fibonacci in Uiua, but without padding and tag built-in function

thejonymyster

@ATaco can i assume no null bytes in the string or do i have to account for those

ATaco

You have to account for them

thejonymyster

nod

ATaco

Lenguage, Score: 0 is an obvious trick

thejonymyster

23:47

lol

ATaco

@thejonymyster You may take the length of the string first if that makes it easier

thejonymyster

doesnt make it easier in this case, but reasonable

i have a "check if input is exhausted" thing i was just trying to golf out :P

Headass, 9 bytes, score 752: {N)U^}:DP (strings are list of charcodes) self-scoring

theres definitely not a shorter answer in bytes but i wonder how low i can reasonably go by using a different construction

lyxal

𝗖𝗠𝗖: Given 3 strings, return a java header comment with the details filled in. The first string is the file name, the second string is the author name, and the third string is a file description. The template looks like this:

/**
* {𝚏𝚒𝚕𝚎𝙽𝚊𝚖𝚎}
* 𝙰𝚞𝚝𝚑𝚘𝚛: {𝚊𝚞𝚝𝚑𝚘𝚛𝙽𝚊𝚖𝚎}
* 𝙳𝚎𝚜𝚌𝚛𝚒𝚙𝚝𝚒𝚘𝚗: {𝚍𝚎𝚜𝚌𝚛𝚒𝚙𝚝𝚒𝚘𝚗}
* (𝚝𝚑𝚎 𝚍𝚎𝚜𝚌𝚛𝚒𝚙𝚝𝚒𝚘𝚗 𝚖𝚊𝚢 𝚜𝚙𝚊𝚗 𝚖𝚞𝚕𝚝𝚒𝚙𝚕𝚎 𝚕𝚒𝚗𝚎𝚜)
*/

Bubbler

ATaco

It's funny that this (Not this^^^) particular challenge overly rewards boring solutions.

Bubbler

23:54

thank you for reminding how many answers I wrote for the challenge

lyxal

hopefully those shouldn't be reversed because I voted on all the answers

Bubbler

I hope so too :P

Bbrk24

I had a really clever Swift answer and then a boring defer{thing}// ended up being shorter

Dannyu NDos

Reminds me of when I lost 60 rep on Korean Language SE because serial upvotes were undone. :(

Bbrk24

Like compare the short and boring way

var x=0,y=1;defer{print(x)}
(x,y)=(y,x+y)//

vs abusing a parser bug

var x = -1,y=1//"
print(String(repeating:"1",count:y),terminator:"")
(x,y)=(y,x+y)
#if swift(<4)
(#endif
"

lyxal

23:58

@lyxal speaking of votes, I'm 6.6k votes away from overtaking gnat as the most prolific voter :p

5

Bubbler

a lang as mainstream as swift has a parser bug??

Bbrk24

@Bubbler Starting in 4.0 and until it was patched in 5.7, unmatched parentheses inside an inactive #if swift() block may be followed by an ill-formed string literal after the #endif. It's so obscure that probably nobody encountered it for all those years

Transcript for