This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Vectorising pass (was: Re: SIMPLE)
- From: Diego Novillo <dnovillo at redhat dot com>
- To: Steven Bosscher <s dot bosscher at student dot tudelft dot nl>
- Cc: GNU Fortran 95 development-mailing-list <g95-develop at lists dot sourceforge dot net>, gcc at gcc dot gnu dot org
- Date: 07 Jan 2002 12:05:58 -0500
- Subject: Re: Vectorising pass (was: Re: SIMPLE)
- References: <1010100260.1437.3.camel@steven><20020103211536.B22468@tornado.cygnus.com><20020103222705.A26819@tornado.cygnus.com> <1010136130.691.0.camel@steven> <20020104082205.B11927@tornado.cygnus.com> <1010189043.12059.13.camel@steven>
On Fri, 2002-01-04 at 19:04, Steven Bosscher wrote:
> If such a pass is implemented, I think it would be nice to have li-trees
> for vector expressions in a new high-level IR.
>
Do you have anything specific in mind? Would this include explicitly
parallel loops (doall, etc)? What would be the memory semantics?
We need to give enough information to allow the passes downstream to
scalarize the expressions (in case it's not worth doing or the target
doesn't support them or any other reason).
Diego.