Scripting Language Startup Overhead | Hacker News Reader

Scripting Language Startup Overhead

git.sr.ht

Nice job. I'd like to add that surprisingly OCaml can be considered a Unix scripting language because it supports both #! shebang scripts and one-shot evaluations e.g. `ocaml -e 'print_endline "hello"'`. In these modes it uses bytecode rather than native compilation.

That is interesting because I thought OCaml only had a native compiler.

(if anyone knows what's up with raku, I'm all ears, because this is a disaster)

Startup time is not the strong point of Rakudo. And it will be a long time before it will.

Beyond startup, performance real 0.04s user 0.03s sys 0.01s

with:

% time real 0.15s user 0.15s sys 0.03s

With NQP taking Those numbers show that real 0.20s user 0.14s sys 0.02s

% time rakudo real 0.16s user 0.15s sys 0.04s

And suddenly, bash 5.2.37 156 busybox ash 1.36.1 35 dash 0.5.12 33 ksh PD KSH v5.2.14 88 mksh R59 75 yash 2.57 93 zsh 5.9 274 busybox awk 1.36.1 19 mawk 1.3.4_20240905 58 gawk 5.3.1 103 nawk 20230909 58 python 2.7.18 629 python 3.12.8 389 perl 5.40 175 raku v2024.10 657 tclsh 8.6 140 jimtcl 0.83 109 lua 5.1.5 83 lua 5.3.6 83 lua 5.4.6 83 luajit 2.1.1716656478 76 nodejs v22.4.1 743 chez scheme 10.0.0 350 petite chez scheme 10.0.0 304 gauche 0.9.15 2530 stklos 2.00 280 guile 3.0.10 1512 guile 3.0.10 (compiler) 3718 ecl 23.9.9 3286 sbcl 2.4.11 241 clisp 2.49.92 399 ccl 1.13 293 definitely has attention of the core developers: that's when hot code gets inlined, and JITted to machine code if you're on Intel hardware. But before that, currently not so much.

The best possible startup time that could potentially be reached in the current implementation of Rakudo, is the bare startup time of NQP in which Rakudo is basically implemented. Compare NQP on an Intel 2.4 GHz i9 on MacOS:

% time nqp -e '' rakudo -e '' about 16MB of memory.

Raku as a language is selected for its features, its community and possibly for its promise. Whether that is a disaster or not, is up to the user.

Finally, if you want fast startup, don't use a scripting language.

>Finally, if you want fast startup, don't use a scripting language.

Huh? Isn't startup speed one of the points of simpler, less optimized interpretation strategies? I mean, unless you compare it to AOT compilation, which is another story. Perl and Tcl, both "full featured" scripting languages, don't have the least issue on the matter.

I'm not the kind to freely rain on parades, especially when Raku is already not a popularity king, but I don't think my use of the word "disaster" was unfounded, even if not "very nice".

> those numbers show that Perl and Tcl, both "full featured" scripting languages, don't have the least issue on the matter.

Anybody using modern Perl will most likely also use Moose, and probably big integers as well. If you compare startup times with Perl / Moose / big integers with Rakudo (on the same Intel hardware as before):

% time perl -MMoose -Mbigint -e '' -e '' Raku starts up faster than Perl! Because Raku comes with a Moose like object system and big integers out of the box, it doesn't need to do anything special to give you those features.

So in a way, you're comparing apples with oranges when you're just comparing bare startup times. Because you are not considering the feature set you get with a bare startup.

So your remark was definitely NOT very nice.

I don't remember where I saw the inspiration for that, but I greatly extended the idea at some point. After a small cleanup today, I thought it could bring some good discussion.

Here are the raw numbers:

  Interpreter                Syscall count  Real time (s; x100)  User time  Kernel time  Max resident memory (KB) 0.07                 0.03       0.04         3596 0.02                 0.01       0.01         1556 0.02                 0.01       0.00         1556 0.05                 0.03       0.02         2376 0.04                 0.02       0.01         1824 0.05                 0.03       0.02         3132 0.08                 0.04       0.03         3792 0.02                 0.01       0.01         1556 0.04                 0.02       0.01         2444 0.06                 0.03       0.03         3568 0.04                 0.02       0.01         2592 0.45                 0.31       0.13         6484 0.74                 0.56       0.18         9636 0.08                 0.02       0.06         5088 15.40                14.49      4.07         148148 0.15                 0.09       0.06         5164 0.12                 0.05       0.06         4052 0.06                 0.03       0.03         2624 0.06                 0.03       0.02         2620 0.06                 0.03       0.02         2644 0.05                 0.03       0.02         2680 1.65                 1.14       0.52         40664 3.70                 2.28       1.42         47568 2.30                 1.42       0.87         31056 3.43                 9.52       1.11         25228 2.78                 2.53       0.25         12368 1.50                 2.02       0.67         16148 5.28                 7.55       1.50         33728 3.61                 5.09       0.91         12268 0.24                 0.06       0.18         19008 0.48                 0.20       0.25         9556 0.36                 0.11       0.21         16456

Reply on news.ycombinator.com