SPUG: return value of assignment operator

BenRifkah Bergsten-Buret mail.spammagnet at gmail.com
Wed Apr 30 13:43:44 PDT 2008


On Wed, Apr 30, 2008 at 12:52 PM, Yitzchak Scott-Thoennes
<sthoenna at efn.org> wrote:
> On Wed, April 30, 2008 12:13 pm, BenRifkah Bergsten-Buret wrote:
>  > According to my setup* relying on the returned lvalue is less efficient:
>
>
> > Rate increment_separately        increment_lhs
>  > increment_separately 5555556/s                   --                 -28%
>  > increment_lhs        7758621/s                  40%                   --
>
>  You're reading it backwards; the increment_lhs is 40% faster than
>  separately.
>

D'oh.  Yeah, Michael pointed this out to me as well.  I got tripped up
on the -28% and interpreted it as separately taking 28% less time than
lhs.  It's obvious now looking at the number of iterations per second.
 I was somewhat surprised by my initial interpretation.  This makes
more sense to me.

So apparently for this kind of incrementation you can get a
substantial performance boost by relying on Perl returning the lhs.

-- Ben


More information about the spug-list mailing list