Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.
  • Yes. These entries are interesting, at least for me. So far my only parrot "thing" is building it every month on cygwin and "complaining" when I cannot. Lack of time so far for more...but thanks to your posts (and the posts of others here) the day I can really join the effort seems closer. Thanks for your work. Cheers. --stephan
  • Yes.

    And the longer answer - thanks for plugging away at the code, and thanks for describing your findings.

  • Here's another ++ from me.

    And didn't you say something to me on a mailing list that instead of searching for micro-optimizations that I should be looking at algorithmic improvements ;)

    • True... but I haven't found any algorithmic improvements in the past week, and I still believe I can get it faster.

  • For me at least this is very interesting. You could skip the math stuff for my sake, but I guess other people find that interesting.
  • I'm still not up to par yet, but I'm slowly trying to figure out what you are doing, may be send in a patch or two one of these days
  • I guess the missing comments are because most people don't have a lot to add to Perl 6, Parrot and Rakudo entries.

    Ranting, on the other hand...

    (And I, for one, think that 5.5% is a huge gain for one simple optimization).
  • There was a time when I read every journal entry - just as I read every post on PerlMonks. I don't have the time any more. Your parrot posts pass the filter test for my time so please, keep them up.
  • I like your technology rants too though.

    And I agree with the above poster - an overall 5.5% speedup is HUGE. If that were on our infrastructure you'd have saved us over a million dollars.
  • As I try to finish my dissertation your blogs (and other P6 blogs) keep things bearable...
  • Limbic_Region said almost exactly what I wanted to say. I read selected journal entries and yours always pass muster. Keep up the good work.
  • ...One idiom jumped out at me:
            $S0 = concat "PIR", ':"'

    Interesting -- apparently PGE is generating these concat statements whenever there is PIR embedded in a rule. For example:


    token MARK_STATEMENT_END {
            {{ $P0 = match.'to'()
                    $P0 = clone $P0
                    set_global '$!endstmt', $P0
            }}
            <.ws>
    }

    • Based on chromatic's analysis I re-examined the section of PGE that was responsible for generating the concats in the first place, and nearly all of the concatenations it was performing at runtime were constants that could be done at compile-time. I must've been asleep when writing that code.

      So, I refactored the code to avoid the concats altogether, and building rakudo's actions.pl file is now 1.6% faster than it was before (44.363 versus 45.095 seconds on my box). Not a huge speedup, but I'm still happie
  • I, too, enjoy reading these notes. If it would help, I would reply, "Interesting!" to each one in the future. I just usually have nothing much to add.
    --
    rjbs
  • chromatic~

    I strongly prefer these posts to your posts complaining.

    Matt
  • also enjoy them. Would be nice to have in rakudo.org as well.
    --

    --
    xoa