X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/98d0883c8949e6cde8db019bb282d5fa815ffc1a..4e7b7a178890eb8d28edcd2e6ab2763c9a3988e5:/doc/todo/Improving_the_efficiency_of_match__95__glob.mdwn diff --git a/doc/todo/Improving_the_efficiency_of_match__95__glob.mdwn b/doc/todo/Improving_the_efficiency_of_match__95__glob.mdwn index c4d3a8ecb..4e1df3381 100644 --- a/doc/todo/Improving_the_efficiency_of_match__95__glob.mdwn +++ b/doc/todo/Improving_the_efficiency_of_match__95__glob.mdwn @@ -1,3 +1,7 @@ +[[!template id=gitbranch branch=smcv/ready/glob-cache + author="[[KathrynAndersen]], [[smcv]]"]] +[[!tag patch]] + I've been profiling my IkiWiki to try to improve speed (with many pages makes speed even more important) and I've written a patch to improve the speed of match_glob. This matcher is a good one to improve the speed of, because it gets called so many times. Here's my patch - please consider it! -- [[KathrynAndersen]] @@ -20,7 +24,132 @@ Here's my patch - please consider it! -- [[KathrynAndersen]] >>>> is less verbose than Kathryn's patch but also not as >>>> fast; I'm not sure why, tbh. --[[smcv]] +>>>>> I think it's because my patch focuses on match_glob while the memoize patch focuses on `glob2re`, and `glob2re` is called in `filecheck`, `meta` and `po` as well as in `match_glob` and `match_user`; thus the memoized `glob2re` is dealing with a bigger set of globs to look up, and thus could be just that little bit slower. -- [[KathrynAndersen]] + +>>>>>> What may be going on is that glob2re is already a fairly fast +>>>>>> function, so the overhead of memoizing it with the very generic +>>>>>> `_memoizer` (see its source) swamps the memoization gain. Note +>>>>>> that the few functions memoized with the Memoizer before were much +>>>>>> more expensive, so that little overhead was acceptable then. +>>>>>> +>>>>>> It also may be that Kathryn's patch is slightly faster due to using +>>>>>> the construct `$foo =~ $regexp` rather than `$foo =~ /$regexp/` +>>>>>> (probably avoids a copy or something like that internally) -- +>>>>>> this despite checking both `exists` and `defined` on the hash, which +>>>>>> should be reundant AFAICS. +>>>>>> +>>>>>> My guess is that the best of both worlds would be to move +>>>>>> the byhand memoization to glob2re and have it return a compiled +>>>>>> `/^/i` regexp that can be used without further modifiction in most +>>>>>> cases. --[[Joey]] + +>>>>>>> Done, see `smcv/ready/glob-cache` and `smcv/glob-cache-too-far`. +>>>>>>> +>>>>>>> Kathryn's patch is a significant improvement; my first patch on top of +>>>>>>> that is a trivial cleanup that speeds it up a little, and the next two +>>>>>>> patches (using precompiled regexes) have surprisingly little effect +>>>>>>> (they don't slow it down either though, so either omit them or merge +>>>>>>> them, whichever). Detailed benchmark results below. +>>>>>>> +>>>>>>> Moving the memoization to `glob2re` actually seems to slow things down +>>>>>>> again - I suspect the docwiki has few enough mentions of `user()` etc. +>>>>>>> that caching them is a waste of time, but perhaps it's not the most +>>>>>>> representative. +>>>>>>> --[[smcv]] + +[[done]] --[[Joey]] + +-------------------------------------------------------------- + +[[!toggle id="smcv-benchmark" text="current benchmarks"]] + +[[!toggleable id="smcv-benchmark" text=""" +master at time of branch: + + time elapsed (wall): 29.6348 + time running program: 24.9212 (84.09%) + time profiling (est.): 4.7136 (15.91%) + number of calls: 1360181 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 13.24 3.2986 3408 0.000968 Text::Balanced::_match_tagged + 10.94 2.7253 79514 0.000034 IkiWiki::PageSpec::match_glob + 3.19 0.7952 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +`Improve the speed of match_glob`: + + time elapsed (wall): 27.9755 + time running program: 23.5293 (84.11%) + time profiling (est.): 4.4461 (15.89%) + number of calls: 1280875 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 14.56 3.4257 3408 0.001005 Text::Balanced::_match_tagged + 7.82 1.8403 79514 0.000023 IkiWiki::PageSpec::match_glob + 3.27 0.7698 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +`match_glob: streamline glob cache slightly`: + + time elapsed (wall): 27.5753 + time running program: 23.1714 (84.03%) + time profiling (est.): 4.4039 (15.97%) + number of calls: 1280875 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 14.09 3.2637 3408 0.000958 Text::Balanced::_match_tagged + 7.74 1.7926 79514 0.000023 IkiWiki::PageSpec::match_glob + 3.30 0.7646 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +`glob2re: return a precompiled, anchored case-insensitiv...`: + + time elapsed (wall): 27.5656 + time running program: 23.1464 (83.97%) + time profiling (est.): 4.4192 (16.03%) + number of calls: 1282189 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 14.21 3.2891 3408 0.000965 Text::Balanced::_match_tagged + 7.72 1.7872 79514 0.000022 IkiWiki::PageSpec::match_glob + 3.32 0.7678 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +`make use of precompiled regex objects`: + + time elapsed (wall): 27.5357 + time running program: 23.1289 (84.00%) + time profiling (est.): 4.4068 (16.00%) + number of calls: 1281981 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 14.17 3.2776 3408 0.000962 Text::Balanced::_match_tagged + 7.70 1.7814 79514 0.000022 IkiWiki::PageSpec::match_glob + 3.35 0.7756 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +`move memoization from match_glob to glob2re`: + + time elapsed (wall): 28.7677 + time running program: 23.9473 (83.24%) + time profiling (est.): 4.8205 (16.76%) + number of calls: 1360181 + number of exceptions: 13 + + %Time Sec. #calls sec/call F name + 13.98 3.3469 3408 0.000982 Text::Balanced::_match_tagged + 8.85 2.1194 79514 0.000027 IkiWiki::PageSpec::match_glob + 3.24 0.7750 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 + +--[[smcv]] +"""]] + -------------------------------------------------------------- + +[[!toggle id="ka-benchmarks" text="Kathryn's benchmarks"]] + +[[!toggleable id="ka-benchmarks" text=""" Benchmarks done with Devel::Profile on the same testbed IkiWiki setup. I'm just showing the start of the profile output, since that's what's relevant. Before: @@ -54,73 +183,13 @@ number of exceptions: 65 Note that the seconds per call for match_glob in the "after" case has gone down by about a third. K.A. +"""]] -------------------------------------------------------------- -A second set of benchmarks, done by rebuilding the docwiki at commit f942c2db05e4 -like so: - - perl -Iblib/lib -d:Profile ikiwiki.in -setup docwiki.setup --no-verbose - -The docwiki appears to use fewer glob matches than Kathryn's wiki. - -With master: - - time elapsed (wall): 29.6970 - time running program: 24.6930 (83.15%) - time profiling (est.): 5.0041 (16.85%) - number of calls: 1359180 - number of exceptions: 13 - - %Time Sec. #calls sec/call F name - 13.62 3.3629 3406 0.000987 Text::Balanced::_match_tagged - 10.84 2.6773 79442 0.000034 IkiWiki::PageSpec::match_glob - 3.08 0.7598 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 - 3.07 0.7593 29830 0.000025 IkiWiki::bestlink - 2.99 0.7378 10231 0.000072 IkiWiki::PageSpec::match_link - -With my `smcv/memoize-glob2re` branch: - - time elapsed (wall): 30.4931 - time running program: 25.1248 (82.39%) - time profiling (est.): 5.3683 (17.61%) - number of calls: 1439943 - number of exceptions: 13 - - %Time Sec. #calls sec/call F name - 13.19 3.3146 3406 0.000973 Text::Balanced::_match_tagged - 8.41 2.1123 79442 0.000027 IkiWiki::PageSpec::match_glob - 3.97 0.9979 86905 0.000011 Memoize::_memoizer - 3.05 0.7654 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 - 3.02 0.7576 29830 0.000025 IkiWiki::bestlink - -and in a repeated run: - - 8.40 2.0905 79442 0.000026 IkiWiki::PageSpec::match_glob - -With Kathryn's patch as seen in my `smcv/ka-glob-cache` branch: - - time elapsed (wall): 27.7567 - time running program: 22.9941 (82.84%) - time profiling (est.): 4.7627 (17.16%) - number of calls: 1279946 - number of exceptions: 13 - - %Time Sec. #calls sec/call F name - 14.29 3.2867 3406 0.000965 Text::Balanced::_match_tagged - 7.89 1.8136 79442 0.000023 IkiWiki::PageSpec::match_glob - 3.30 0.7577 59454 0.000013 :IkiWiki/Plugin/inline.pm:223 - 3.24 0.7461 29830 0.000025 IkiWiki::bestlink - 3.19 0.7332 143 0.005127 ? IkiWiki::pagespec_match_list - -and in a repeated run: - - 7.84 1.8253 79442 0.000023 IkiWiki::PageSpec::match_glob - ---[[smcv]] - --------------------------------------------------------------- +[[!toggle id="ka-patch" text="Kathryn's original patch"]] +[[!toggleable id="ka-patch" text=""" 
 diff --git a/IkiWiki.pm b/IkiWiki.pm
@@ -155,4 +224,5 @@ index 08a3d78..c187b98 100644
  			return IkiWiki::SuccessReason->new("$glob matches $page");
  		}
+"""]] --------------------------------------------------------------