On Fri, Oct 24, 2008 at 09:40:52PM -0400, Hans-Peter Nilsson wrote:
> Sounds like a perfect use for the new feature attribute-enabled;
> skipping alternatives depending on TARGET_... See md.texi
> "@subsection Disable insn alternatives using the @code{enabled}
> attribute".
Thanks for the pointer
On Sat, 25 Oct 2008, Joern Rennecke wrote:
> Moreover, the .md file also doesn't get more readable if I replace
> a three character constraint string with a multiword iterator invocation...
> for something like eight out of 14 alternatives for some instructions.
Um yeah, there is that...
> FWIW,
On Fri, Oct 24, 2008 at 07:06:29AM -0400, Hans-Peter Nilsson wrote:
> You could use iterators... except you'd probably want to iterate
> over text, not RTX codes or machine modes. Maybe a good reason
> to introduce generalized (or maybe just text) iterators!
>
> (define_iterator i [foo bar])
> (d
On Fri, 17 Oct 2008, Joern Rennecke wrote:
> On Fri, Oct 17, 2008 at 10:46:41AM -0700, Ian Lance Taylor wrote:
> >
> > I think you could achieve the same result by writing multiple
> > define_insn patterns and using the instruction predicate.
>
> Yes, I could. But that would quadruple my machine
On Fri, Oct 17, 2008 at 10:46:41AM -0700, Ian Lance Taylor wrote:
>
> I think you could achieve the same result by writing multiple
> define_insn patterns and using the instruction predicate.
Yes, I could. But that would quadruple my machine description,
which is already 169 KB, and it would mak
Joern Rennecke <[EMAIL PROTECTED]> writes:
> ARCompact has three-address instructions, but it has a number of
> shorter two-address instruction codes.
>
> When optimizing or speed, I think it would generally be best to
> choose the three-address alternatives for input reloads, since this
> allows
ARCompact has three-address instructions, but it has a number of
shorter two-address instruction codes.
When optimizing or speed, I think it would generally be best to
choose the three-address alternatives for input reloads, since this
allows more reload inheritance. OTOH, when optimizing for siz
