This is the mail archive of the
mailing list for the GCC project.
Re: Forcing REG_DEAD?
- From: Richard Sandiford <rdsandiford at googlemail dot com>
- To: Richard Hulme <peper03 at yahoo dot com>
- Cc: GCC Development <gcc at gcc dot gnu dot org>
- Date: Sun, 06 Apr 2014 22:05:20 +0100
- Subject: Re: Forcing REG_DEAD?
- Authentication-results: sourceware.org; auth=none
- References: <5341B774 dot 5060104 at yahoo dot com>
Richard Hulme <firstname.lastname@example.org> writes:
> Is there any way of preventing the compiler going with the indirect
> alternative? Can a 'parallel' match be defined in the machine
> description that indicates the '(set (reg/f:HI...' should be discarded?
No, that wouldn't work. The register is only dead if both combinations
take place. Neither combine nor the target code knows when doing the
first combination whether all other uses of the register can also be
Combine just isn't designed for this kind of global optimisation.
It sounds like something that fwprop should do instead.