This is the mail archive of the gcc@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: malloc attributes and realloc


Jonathan Lennox <lennox@cs.columbia.edu> writes:
>> ... that is *not* an optimization; the cache behavior of the latter is
>> likely to be worse than the former.
>
> I was referencing the behavior of if-conversion which Roger Sayle described
> in <http://gcc.gnu.org/ml/gcc-patches/2003-12/msg01271.html>.  As I
> mentioned, the code's in GCC already.
>
> I don't know much about the logic of if-conversion, but presumably GCC only
> does this transformation if it's a general win?

It depends on the hardware, actually.  The situation that I know of,
where it was a huge lose, involved a multiprocessor box and a shared
data structure being updated on both CPUs.  Writing the value that
was already there caused the cache line to get bounced back and forth
between the caches.

zw


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