http://www.perlmonks.org?node_id=948194


in reply to Re: Question about curious performance of if...elsif block
in thread Question about curious performance of if...elsif block

If you read the base note more carefully you're see that /proc/pid/status contains 36 entries on my system.

yes I am doing the comparisons before returning and I want to so I can measure the difference between executing them and returning immediately and the difference is 5 seconds, not the 3/4 second I would have expected.

understand now?

-mark

Replies are listed 'Best First'.
Re^3: Question about curious performance of if...elsif block
by BrowserUk (Patriarch) on Jan 16, 2012 at 19:17 UTC

    Even with 36 cases, I still see nothing like the 5 second difference you apparently are:

    #! perl -slw use strict; use Time::HiRes qw[ time ]; sub test1 { my $a = $_[0]; # return; if ($a==1) {} elsif ($a==2) {} elsif ($a==3) {} elsif ($a==4) {} elsif ($a==5) {} elsif ($a==6) {} elsif ($a==7) {} elsif ($a==8) {} elsif ($a==9) {} elsif ($a==10) {} elsif ($a==11) {} elsif ($a==12) {} elsif ($a==13) {} elsif ($a==14) {} elsif ($a==15) {} elsif ($a==16) {} elsif ($a==17) {} elsif ($a==18) {} elsif ($a==19) {} elsif ($a==20) {} elsif ($a==21) {} elsif ($a==22) {} elsif ($a==23) {} elsif ($a==24) {} elsif ($a==25) {} elsif ($a==26) {} elsif ($a==27) {} elsif ($a==28) {} elsif ($a==29) {} elsif ($a==30) {} elsif ($a==31) {} elsif ($a==32) {} elsif ($a==33) {} elsif ($a==34) {} elsif ($a==35) {} elsif ($a==36) {} } sub test2 { my $a = $_[0]; return ; if ($a==1) {} elsif ($a==2) {} elsif ($a==3) {} elsif ($a==4) {} elsif ($a==5) {} elsif ($a==6) {} elsif ($a==7) {} elsif ($a==8) {} elsif ($a==9) {} elsif ($a==10) {} elsif ($a==11) {} elsif ($a==12) {} elsif ($a==13) {} elsif ($a==14) {} elsif ($a==15) {} elsif ($a==16) {} elsif ($a==17) {} elsif ($a==18) {} elsif ($a==19) {} elsif ($a==20) {} elsif ($a==21) {} elsif ($a==22) {} elsif ($a==23) {} elsif ($a==24) {} elsif ($a==25) {} elsif ($a==26) {} elsif ($a==27) {} elsif ($a==28) {} elsif ($a==29) {} elsif ($a==30) {} elsif ($a==31) {} elsif ($a==32) {} elsif ($a==33) {} elsif ($a==34) {} elsif ($a==35) {} elsif ($a==36) {} } sub test3 { return if $_[0] > 9; my $a = $_[0]; if ($a==1) {} elsif ($a==2) {} elsif ($a==3) {} elsif ($a==4) {} elsif ($a==5) {} elsif ($a==6) {} elsif ($a==7) {} elsif ($a==8) {} elsif ($a==9) {} elsif ($a==10) {} elsif ($a==11) {} elsif ($a==12) {} elsif ($a==13) {} elsif ($a==14) {} elsif ($a==15) {} elsif ($a==16) {} elsif ($a==17) {} elsif ($a==18) {} elsif ($a==19) {} elsif ($a==20) {} elsif ($a==21) {} elsif ($a==22) {} elsif ($a==23) {} elsif ($a==24) {} elsif ($a==25) {} elsif ($a==26) {} elsif ($a==27) {} elsif ($a==28) {} elsif ($a==29) {} elsif ($a==30) {} elsif ($a==31) {} elsif ($a==32) {} elsif ($a==33) {} elsif ($a==34) {} elsif ($a==35) {} elsif ($a==36) {} } my @dispatch = ( sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, sub {}, ); sub test4 { return if $_[0] > 9; $dispatch[ $_[0] ]->( $_[0] ); } my $start = time; test1( $_ ) for 1 .. 1e6; printf "test1 took: %.5f seconds\n", time - $start; $start = time; test2( $_ ) for 1 .. 1e6; printf "test2 took: %.5f seconds\n", time - $start; $start = time; test3( $_ ) for 1 .. 1e6; printf "test3 took: %.5f seconds\n", time - $start; $start = time; test4( $_ ) for 1 .. 1e6; printf "test4 took: %.5f seconds\n", time - $start; __END__ C:\test>junk37 test1 took: 1.70500 seconds test2 took: 0.30789 seconds test3 took: 0.29240 seconds test4 took: 0.29207 seconds

    With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

    The start of some sanity?

      ok, one more time. try embedding my subruotine in a loop reading /proc/pid/status and measure it. THAT's my point. In a small program the timings are as expected. My question is why the unexpected performance when reading /proc?

      Did you try my second script on your system? Try that without the return statement.

      -mark

        Did you try my second script on your system?

        No. Because it would not run here.

        ok, one more time. try embedding my subruotine in a loop reading /proc/pid/status and measure it. THAT's my point.
        Then it is fairly obvious that the reason for the differences you are concerned about have nothing to do with that subroutine and every thing to do with the code that is calling it. Ie. All your concentration on the subroutine is a complete red herring.

        It seems fairly clear to me that you are misinterpreting the data you are reading. What timing do you see if you comment out the subroutine call entirely?

        But don't bother with "ok, one more time.". Cos I almost certainly wouldn't understand it.


        With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
        Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
        "Science is about questioning the status quo. Questioning authority".
        In the absence of evidence, opinion is indistinguishable from prejudice.

        The start of some sanity?

Re^3: Question about curious performance of if...elsif block
by BrowserUk (Patriarch) on Jan 16, 2012 at 19:11 UTC
    understand now?

    Apparently not.


    With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

    The start of some sanity?