15 interpreters vs. gawk

[mss]

Just thinking aloud here...

For a project I created here:

<http://www.topcat.hypermart.net/topic.html>

Used several interpreters, & gawk beat them all, with respect
to speed/size, & in some cases, by a couple orders of magnitude
(though sed was close). But admittedly, this was a DSL project
of the sort [g|m]awk excels at. Still, awk is a great language
for me =)

Interpreters (in no particular order) were:

- AWK*
- SED
- Scala
- Rhino
- Clojure
- Common Lisp
- Pike
- Lua
- Perl
- Ruby
- Python
- BCX
- SNOBOL4
- Fortran
- QBASIC
- Chipmunk BASIC

* Thanks for the help pk, Janis & Ed, I appreciate it.

--
later on,
Mike

http://www.topcat.hypermart.net/index.html

[addre...@invalid.invalid]

I had a look at the site- no graphs of of runtimes?

Also- care to add the luajit to the list? It's getting good press

T

[mss]

<addre...@invalid.invalid> wrote:

> I had a look at the site- no graphs of of runtimes?

Was not a competition, just an observation...
None of the times were 'bad' on the Unix/Linux
side of things. Not near so great under Windows however;
subsystem apps (text mode that is) seem to have only gotten
slower with each new release of Windows.

Here are my informal results using lua 5.1.4/gawk v3.17
under Slackware:

lua invocation: 'time lua topic.lua notes < topic-spec.txt'

lua results:

real 0m0.005s
user 0m0.004s
sys 0m0.000s

gawk invocation: 'time gawk -f topic.awk -v TPC=notes < topic-spec.txt'

gawk results:

real 0m0.003s
user 0m0.000s
sys 0m0.000s

Here, I'm judging the best time by adding the real/user/sys times together.
An aggregate run of 3 times with the tests above is about the same on my end.

> Also- care to add the luajit to the list? It's getting good press

Ahh.. nifty. Well I'll certainly look into it when I can.
Good luck with luajit, seems like an interesting project.