# Ternary operator performance improvements - Thread: https://externals.io/message/55805 - Original archive: https://news-web.php.net/php.internals/55805 - Started by Arnaud Le Blanc on 2011-10-14T18:08:56+00:00 - Messages: 7 - Reply by email to internals@lists.php.net - More for agents: search with https://externals.io/search.md?q={query} · JSON API at https://externals.io/api · documentation at https://externals.io/llms.txt --- ## #55805 — Arnaud Le Blanc — 2011-10-14T18:08:56+00:00 URL: https://externals.io/message/55805#55805 Hi, I've already posted this patch and it has since been reviewed and improved. I'm re-posting it for discussion before eventually commiting it. The ternary operator always copies its second or third operand, which is very slow compared to an if/else when the operand is an array for example: $a = range(0,9); // this takes 0.3 seconds here: for ($i = 0; $i < 5000000; ++$i) { if (true) { $b = $a; } else { $b = $a; } } // this takes 3.8 seconds: for ($i = 0; $i < 5000000; ++$i) { $b = true ? $a : $a; } I've tried to reduce the performance hit by avoiding the copy when possible (patch attached). Benchmark: Without patch: (the numbers are the time taken to run the code a certain amount of times) $int = 0; $ary = array(1,2,3,4,5,6,7,8,9); true ? 1 : 0 0.124 true ? 1+0 : 0 0.109 true ? $ary : 0 2.020 ! true ? $int : 0 0.103 true ? ${'ary'} : 0 2.290 ! true ?: 0 0.091 1+0 ?: 0 0.086 $ary ?: 0 2.151 ! ${'var'} ?: 0 2.317 ! With patch: true ? 1 : 0 0.124 true ? 1+0 : 0 0.195 true ? $ary : 0 0.103 true ? $int : 0 0.089 true ? ${'ary'} : 0 0.103 true ?: 0 0.086 1+0 ?: 0 0.159 $cv ?: 0 0.090 ${'var'} ?: 0 0.089 The array copying overhead is eliminated. There is however a slowdown in some of the cases, but overall there is no completely unexpected performance hit as it is the case currently. What do you think ? Is there any objection ? Best regards, --- ## #55806 — Dmitry Stogov — 2011-10-14T19:39:18+00:00 URL: https://externals.io/message/55805#55806 · In reply to: #55805 Hi, As I already said, I don't see any technical issues with the patch. I don't like it very mach, but it looks robust, and I don't object about it. Thanks. Dmitry. > Hi, > > I've already posted this patch and it has since been reviewed and improved. > I'm re-posting it for discussion before eventually commiting it. > > The ternary operator always copies its second or third operand, which is very > slow compared to an if/else when the operand is an array for example: > > $a = range(0,9); > > // this takes 0.3 seconds here: > > for ($i = 0; $i< 5000000; ++$i) { > if (true) { > $b = $a; > } else { > $b = $a; > } > } > > // this takes 3.8 seconds: > > for ($i = 0; $i< 5000000; ++$i) { > $b = true ? $a : $a; > } > > I've tried to reduce the performance hit by avoiding the copy when possible > (patch attached). > > Benchmark: > > Without patch: (the numbers are the time taken to run the code a certain > amount of times) > > $int = 0; > $ary = array(1,2,3,4,5,6,7,8,9); > > true ? 1 : 0 0.124 > true ? 1+0 : 0 0.109 > true ? $ary : 0 2.020 ! > true ? $int : 0 0.103 > true ? ${'ary'} : 0 2.290 ! > true ?: 0 0.091 > 1+0 ?: 0 0.086 > $ary ?: 0 2.151 ! > ${'var'} ?: 0 2.317 ! > > With patch: > > true ? 1 : 0 0.124 > true ? 1+0 : 0 0.195 > true ? $ary : 0 0.103 > true ? $int : 0 0.089 > true ? ${'ary'} : 0 0.103 > true ?: 0 0.086 > 1+0 ?: 0 0.159 > $cv ?: 0 0.090 > ${'var'} ?: 0 0.089 > > > The array copying overhead is eliminated. There is however a slowdown in some > of the cases, but overall there is no completely unexpected performance hit as > it is the case currently. > > What do you think ? Is there any objection ? > > Best regards, --- ## #55839 — Ilia Alshanetsky — 2011-10-17T12:52:21+00:00 URL: https://externals.io/message/55805#55839 · In reply to: #55805 Seems like a good patch, +1 from me on inclusion into 5.4/HEAD. > Hi, > > I've already posted this patch and it has since been reviewed and improved. > I'm re-posting it for discussion before eventually commiting it. > > The ternary operator always copies its second or third operand, which is very > slow compared to an if/else when the operand is an array for example: > > $a = range(0,9); > > // this takes 0.3 seconds here: > > for ($i = 0; $i < 5000000; ++$i) { >        if (true) { >                $b = $a; >        } else { >                $b = $a; >        } > } > > // this takes 3.8 seconds: > > for ($i = 0; $i < 5000000; ++$i) { >        $b = true ? $a : $a; > } > > I've tried to reduce the performance hit by avoiding the copy when possible > (patch attached). > > Benchmark: > > Without patch: (the numbers are the time taken to run the code a certain > amount of times) > > $int = 0; > $ary = array(1,2,3,4,5,6,7,8,9); > > true ? 1 : 0        0.124 > true ? 1+0 : 0      0.109 > true ? $ary : 0     2.020 ! > true ? $int : 0     0.103 > true ? ${'ary'} : 0 2.290 ! > true ?: 0           0.091 > 1+0 ?: 0            0.086 > $ary ?: 0           2.151 ! > ${'var'} ?: 0       2.317 ! > > With patch: > > true ? 1 : 0        0.124 > true ? 1+0 : 0      0.195 > true ? $ary : 0     0.103 > true ? $int : 0     0.089 > true ? ${'ary'} : 0 0.103 > true ?: 0           0.086 > 1+0 ?: 0            0.159 > $cv ?: 0            0.090 > ${'var'} ?: 0       0.089 > > > The array copying overhead is eliminated. There is however a slowdown in some > of the cases, but overall there is no completely unexpected performance hit as > it is the case currently. > > What do you think ? Is there any objection ? > > Best regards, --- ## #55840 — Felipe Pena — 2011-10-17T12:57:05+00:00 URL: https://externals.io/message/55805#55840 · In reply to: #55839 2011/10/17 Ilia Alshanetsky : > Seems like a good patch, +1 from me on inclusion into 5.4/HEAD. > > > > […] > > -- > > > > +1 from me too. -- Regards, Felipe Pena --- ## #55842 — Alain Williams — 2011-10-17T13:07:30+00:00 URL: https://externals.io/message/55805#55842 · In reply to: #55805 > Hi, > > I've already posted this patch and it has since been reviewed and improved. > I'm re-posting it for discussion before eventually commiting it. > > The ternary operator always copies its second or third operand, which is very > slow compared to an if/else when the operand is an array for example: Is that why the following does not work as I expected: $dbh = $how == 'r' ? (&$dbh_r) : (&$dbh_w); $dbh is NOT a reference to $dbh_r or $dbh_w. -- Alain Williams Linux/GNU Consultant - Mail systems, Web sites, Networking, Programmer, IT Lecturer. +44 (0) 787 668 0256 http://www.phcomp.co.uk/ Parliament Hill Computers Ltd. Registration Information: http://www.phcomp.co.uk/contact.php #include > > […] > > > […] > Is that why the following does not work as I expected: > > $dbh = $how == 'r' ? (&$dbh_r) : (&$dbh_w); > > $dbh is NOT a reference to $dbh_r or $dbh_w. This is expected; http://docs.php.net/manual/en/language.operators.comparison.php explains it: > Please note that the ternary operator is a statement, and that it doesn't evaluate to a variable, but to the result of a statement. This is why you can't assign the result of the ternary operator by reference. The patch doesn't change this. Best regards, --- ## #55865 — Stas Malyshev — 2011-10-18T21:47:43+00:00 URL: https://externals.io/message/55805#55865 Hi! > It probably does, but it is quite subtle. I was expecting the above to work since > it does in C -- although in C&variable is an address which can be used in an > expression ... the PHP '&' operator is different, in spite of apparent similarities. Actually, in PHP standalone & doesn't even make any sense. PHP has certain operations - assignment by reference (bind two names to one variable) and passing by reference (which is the same thing but cross-scope) and returning by reference (which again is the same thing but in opposite direction). But unlike C there's no operation "taking reference" because while C has pointers PHP doesn't have anything like that. -- Stanislav Malyshev, Software Architect SugarCRM: http://www.sugarcrm.com/ (408)454-6900 ext. 227