Front page | perl.perl5.porters |
Postings from November 2011
Jesse Luehrs <doy@tozt.net> wrote on Wed, 23 Nov 2011 11:27:27 CST: > Well, if nothing else, it's a potentially unnecessary use of > ~500kb/process. Ouch! That's good enough for me. Well, or used to be. Guido recently told the Python folks that having to load 500k of Unicode properties to get regexes to work right on Unicode data didn't really bother him much, and I tend to agree with him. Still, most people think nothing of loading in a megabyte of cruft to do something just to get some syntactic sugar for already-loaded, built-in functionality. These are with v5.14: openbsd% perl -e 'system "ps v$$"; require IO::Handle; system "ps v$$"' PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 9539 S+ 0:00.01 0 0 0 1244 1704 496036 940 0.0 0.3 perl -e system "ps v$ PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 9539 S+ 0:00.06 0 1 0 1476 2580 496036 940 1.0 0.5 perl -e system "ps v$ That's +232 VSZ, +876 RSS. openbsd% perl -e 'system "ps v$$"; STDOUT->autoflush; system "ps v$$"' PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 12282 S+ 0:00.01 0 0 0 1160 1712 496036 940 0.0 0.3 perl -e system "ps v$$"; STDOUT->au PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 12282 S+ 0:00.09 0 1 0 1468 2840 496036 940 4.0 0.5 perl -e system "ps v$$"; STDOUT->au That's +308 VSZ, +1228 RSS. openbsd% perl -e 'system "ps v$$"; $| = 1; system "ps v$$"' PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 2575 S+ 0:00.01 0 0 0 1244 1712 496036 940 0.0 0.3 perl -e system "ps v$$"; $| = 1; sy PID STAT TIME SL RE PAGEIN VSZ RSS LIM TSIZ %CPU %MEM COMMAND 2575 S+ 0:00.01 0 0 0 1244 1740 496036 940 0.0 0.3 perl -e system "ps v$$"; $| = 1; sy That's +0 VSZ, +28 RSS. I know, I know: it's "only" a megabyte, and legibility is more important than performance. Still. I've occasionally wondered whether the compiler could catch that as a special case, but I end up deciding it's not worth the code bloat in the compiler source such a hack would require. Never stopped FORTRAN from optimizing its intrinsics though. Adding cheats to the compiler can be worth more than they appear to cost, because it's a rising tide that raises all ships. Look at all the compiler hacks for detecting the standard Livermore Loops benchmark, for example. Nobody thinks those should be removed for truth and beauty. I just wouldn't want to have to be the one supporting them. --tomThread Previous | Thread Next