This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tree-ssa] Generic NRV optimization [was Re: two-element struct performance (was: strict-aliasing and typedefs) ]


In message <1077825678.16194.483.camel@localhost.localdomain>, Diego Novillo wr
ites:
 >The former is what I was hinting at with the idea of moving the RTL
 >expanding logic from expand_return().  It would have to be adapted, of
 >course.
Yup.  Worse yet, you have to look at anything which uses DECL_RTL of
the the return value and make sure it's doing the right thing.

 >The latter is going to prevent NRV on any target that builds a PARALLEL
 >or a REG RETURN_DECL, wouldn't it?
Yes.  But in those cases the return value is going into registers anyway.
Those cases are a smaller benefit than the one we do handle (return value
in memory).

jeff




Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]