00:42 avuserow left, avuserow joined 02:11 topnep left 02:12 topnep joined
disbot8 <simon_sibl> I changed the code a bit and added 30x the number of jobs to do Raku is at 400ms and Python stays at 20ms maybe the hashmap is a big performance hole here ? the python code uses array index which could be faster but makes the code less readable I think 02:48
<simon_sibl> somehow using index instead of hash its even slower 😆 02:54
avuserow simon_sibl: have you tried the profiler? a quick run of termbin.com/znm3 shows some slowness in the `printf` call, and it looks like you could use `say` instead 02:59
disbot8 <simon_sibl> oh I removed the printf for fairness yes, since the python code doesnt print 03:00
avuserow ah, I have not compared vs the python version, just took a quick look out of curiosity :) 03:01
disbot8 <simon_sibl> I am not yet used to Raku profiler, but when I check allocation the biggest one is Hashmap (somehow its even bigger when using the Array version so only keeping the hash for the result) 03:02
<simon_sibl> the GC doesnt take much time
<simon_sibl> it says The profiled code ran for 99.62ms does it mean the program ran for 99ms ? 03:03
<simon_sibl> because when I use my shell time and run the program I get 490ms
avuserow other culprits in the profile are the call to `min` which seems like it could be a single call to `sort` and then using `pop` or a for loop.
disbot8 <simon_sibl> the issue is that, its required to sort every time after the loop since it changes the time and priority 03:04
<simon_sibl> I am sure there are some math possible to calculate that in advance
avuserow I would expect the profiler will under-report since it's within raku and does not account for some start-up time, nor does it account for its own runtime :)
disbot8 <simon_sibl> but the Python code also sort during the loop
<simon_sibl> whats all the <anon> ? it couldnt find the name of the sub or something ? 03:05
avuserow I think that's unnamed blocks
03:59 stanrifkin_ joined 04:02 stanrifkin left
disbot8 <simon_sibl> makes sense xD 04:54
06:04 SEric left 06:05 SEric joined 06:22 Guest5 joined 06:23 Guest5 left
disbot8 <simon_sibl> is there a go to module for GUI ? 07:54
<rcmlz> If execution speed is the most important metric, why not compare Raku/Python to e.g. Rust? Perhaps compared to Rust booth appear indistinguishable„slow“ ;-) Everything is relative … (I did some experiments giving some - IMHO good - Python code that did some brute-forcing to ChatGPT asking for an equivalent Rust-version. The Rust version of my toy-brute-force problems was much, much, much faster.) IMHO the 08:17
strengths of Raku are -O-fun due to so much Batteries-included thingd that are fun to explore (and Grammars.)
<simon_sibl> its the balance of human coding speed and software speed, Rust is very slow for human to code, Python and Raku are fast for human to code, otherwise in the benchmark I could add assembly/Zig/C xD 08:28
<simon_sibl> just since Raku and Python are somewhat similar in human coding speed, thats why I compare those two 08:29
<simon_sibl> what does the -O- means before fun ?
<rcmlz> Yes, I aggree. Raku’s mottos: "Raku is designed to make the easy jobs easy, without making the hard jobs impossible." "There Is More Than One Way To Do It - now with even more -Ofun (optimized for fun) added." 08:33
<rcmlz> -Ofun … I like that 08:34
<rcmlz> Web-based? Or Linux/Mac/Windows-native? 08:35
<rcmlz> In Raku I personally never did any GUI-things, only CMD-Line-Tools or Jupyter-Notebooks. Always wanted to try HARC approach from @librasteve , but do not habe a usecase yet. 08:38
<rcmlz> Raku-Jupyter-Notebook come in two flavours - with or without AI included. raku.land/zef:bduggan/Jupyter::Kernel raku.land/zef:antononcube/Jupyter::Chatbook 08:45
<simon_sibl> oooh jupyter nice, I will give a try 09:19
<simon_sibl> web would be okay
<simon_sibl> native better I guess, but everything is web now xD
09:20 greenfork left, greenfork joined, KOTP left 09:21 KOTP joined 09:42 hankache joined 09:45 hankache left 09:58 hankache joined
disbot8 <rcmlz> In case you use Nix and want to try Jupyter::Chatbook: github.com/rcmlz/Raku-Nix-Shells 10:39
10:40 hankache left
disbot8 <simon_sibl> oh I use Guix, but it works well 10:40
<rcmlz> Rakudo from 2022 ... packages.guix.gnu.org/search/?query=rakudo 10:44
<rcmlz> and here is HARC: harcstack.org/ 10:48
<rcmlz> rakujourney.wordpress.com/2025/03/...arc-stack/ 10:49
<rcmlz> but as said, I only use command line and Juypter - so can not comment on native GUI od HARC 10:50
<simon_sibl> oh I used Rakudo star to get the lastest version on my Guix system 11:05
11:59 topnep left 12:02 topnep joined 12:04 topnep left 12:05 topnep joined
disbot8 <rcmlz> Good choice. I had a look at the definition and that really needs some attention to update to the latest releases … (zef version 0.12 …etc.) I do not know how big guix user community is, but it would help Raku, if that would be updated. codeberg.org/guix/guix/src/branch/.../perl6.scm 12:27
<simon_sibl> I will try to update it, but I am not a master of Guix packages 13:38
<simon_sibl> so far I could get moarvm, nqp and nqp-configure to their last version 14:32
<simon_sibl> trying to build rakudo 2025-06-1
<simon_sibl> anyone with experience could help me figure out what happens with those tests ? termbin.com/vibz 14:40
<simon_sibl> I noticed they build the moarvm without dynasm, could that be a reason of those failures ? 14:42
16:18 topnep left, topnep joined 17:38 sivoais left, sivoais joined
disbot8 <jubilatious1_98524> @librasteve Here's an interesting/fun graphic on the Perl PA&A Discord (discusses HTMX): discord.com/channels/8609007676654...4879911956 19:09
<jubilatious1_98524> @librasteve Here's an interesting/fun graphic on the Perl PA&A Discord (discusses HTMX): discord.com/channels/8609007676654...4879911956 19:10
20:29 topnep left 20:31 topnep joined 22:35 topnep left 22:36 topnep joined