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: [ira] patch solving IRA slowness in -O0 mode


Richard Sandiford wrote:
Vladimir Makarov <vmakarov@redhat.com> writes:
As David Miller wrote analogous code was used at least last 10 (may be 15) years in global.c. We could rid off double these days because gcc supports long long now. Long ago 64-bit int support was not guaranteed.

Doesn't look like anyone else had said this yet, so just to be on the safe side...

My understanding is that we still can't assume "long long" exists across
the board, because we need to be compatible with non-gcc C90 compilers.
Uses of "long long" would need to be conditional and would need to have
a pure C90 fallback.

In practice, HOST_WIDEST_INT is a 64-bit integer practically everywhere. (Or everywhere?)


Paolo


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