This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: An unusual Performance approach using Synthetic registers
- From: Alexandre Oliva <aoliva at redhat dot com>
- To: Chris Lattner <sabre at nondot dot org>
- Cc: Andy Walker <ja_walker at earthlink dot net>, gcc at gcc dot gnu dot org
- Date: 29 Dec 2002 08:23:32 -0200
- Subject: Re: An unusual Performance approach using Synthetic registers
- Organization: GCC Team, Red Hat
- References: <Pine.LNX.4.44.0212290229520.1602-100000@nondot.org>
On Dec 29, 2002, Chris Lattner <sabre@nondot.org> wrote:
> However, the true problem is that GCC performs most of its scalar
> optimizations on the RTL representation, which operates on values in
> physical registers. For an machine with few physical registers, such as
> X86, many scalar values are spilled to the stack, impeding optimization.
Except that they're only spilled to stack after register allocation is
completed, and most of the optimizations take place before that, when
we still hope that all pseudo registers will get assigned to hardware
registers.
> The idea here is that you perform optimizations on a representation
> exposing an infinite virtual register set, so scalar stay scalars,
> and optimizations are effective despite the number of physical
> registers in the target.
This is already the case of GCC RTL.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist Professional serial bug killer