On Tue, 4 Oct 2016, Joseph Myers wrote:
> On Tue, 4 Oct 2016, Richard Biener wrote:
>
> > On Tue, 4 Oct 2016, Joseph Myers wrote:
> >
> > > On Tue, 4 Oct 2016, Richard Biener wrote:
> > >
> > > > Possibly. Though then for FP we also want - abs (a) -> copysign (a,
> > > > -1).
> > >
> > > For architectures such as powerpc that have a negated-abs instruction,
> > > does it get properly generated from the copysign code? (The relevant
> > > pattern in rs6000.md uses (neg (abs)) - is that the correct canonical RTL
> > > for this?)
> >
> > I have no idea - given that there is no copysign RTL code I suppose
> > it is the only machine independent RTL that can do this?
> >
> > The question would be whether the copysign optab of said targets would
> > special-case the -1 case appropriately.
>
> Why -1? Any constant in copysign should be handled appropriately. I'd
> say that (neg (abs)) is probably better as a canonical representation, so
> map copysign from a constant to either (abs) or (neg (abs)) appropriately.
>
> Then in the case where abs is expanded with bit manipulation, (neg (abs))
> should be expanded to a single OR. I don't know whether the RTL
> optimizers will map the AND / OR combination to a single OR, but if they
> do then there should be no need to special-case (neg (abs)) in expansion,
> and when (abs) RTL is generated a machine description's (neg (abs))
> pattern should match automatically.
Ok, that's a good point -- I'll do copysign (X, const) canonicalization
to [-]abs() then. Hopefully that's a valid transform even if NaNs and
signed zeros are involved.
For now I've installed the requested min(a,-a)->-abs(a) pattern.
Bootstrapped / tested on x86_64-unknown-linux-gnu.
Richard.
2016-10-05 Richard Biener <rguent...@suse.de>
PR middle-end/55152
* match.pd (min(a,-a) -> -abs(a)): New pattern.
* gcc.dg/pr55152-2.c: New testcase.
Index: gcc/match.pd
===================================================================
--- gcc/match.pd (revision 240738)
+++ gcc/match.pd (working copy)
@@ -1291,6 +1302,13 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
&& (! ANY_INTEGRAL_TYPE_P (type)
|| TYPE_OVERFLOW_UNDEFINED (type)))
(abs @0)))
+/* min(a,-a) -> -abs(a). */
+(simplify
+ (min:c @0 (negate @0))
+ (if (TREE_CODE (type) != COMPLEX_TYPE
+ && (! ANY_INTEGRAL_TYPE_P (type)
+ || TYPE_OVERFLOW_UNDEFINED (type)))
+ (negate (abs @0))))
(simplify
(min @0 @1)
(switch
Index: gcc/testsuite/gcc.dg/pr55152-2.c
===================================================================
--- gcc/testsuite/gcc.dg/pr55152-2.c (revision 0)
+++ gcc/testsuite/gcc.dg/pr55152-2.c (working copy)
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-options "-O -ffinite-math-only -fno-signed-zeros -fstrict-overflow
-fdump-tree-optimized" } */
+
+double g (double a)
+{
+ return (a<-a)?a:-a;
+}
+int f(int a)
+{
+ return (a<-a)?a:-a;
+}
+
+/* { dg-final { scan-tree-dump-times "ABS_EXPR" 2 "optimized" } } */
