re PR middle-end/34163 (10% performance regression since Nov 1 on Polyhedron's "NF" on AMD64)

2009-07-03  Richard Guenther  <rguenther@suse.de>

	PR middle-end/34163
	* tree-chrec.c (chrec_convert_1): Fold (T2)(t +- x) to
	(T2)t +- (T2)x if t +- x is known to not overflow and
	the conversion widens the operation.
	* Makefile.in (tree-chrec.o): Add $(FLAGS_H) dependency.

	* gfortran.dg/pr34163.f90: New testcase.

From-SVN: r149207

gcc/ChangeLog

+2009-07-03  Richard Guenther  <rguenther@suse.de>
+
+	PR middle-end/34163
+	* tree-chrec.c (chrec_convert_1): Fold (T2)(t +- x) to
+	(T2)t +- (T2)x if t +- x is known to not overflow and
+	the conversion widens the operation.
+	* Makefile.in (tree-chrec.o): Add $(FLAGS_H) dependency.
+
 2009-07-03  Jan Hubicka  <jh@suse.cz>
 
 	* ipa-pure-const.c (analyze): Update loop optimizer init.

gcc/Makefile.in

@@ -2414,7 +2414,7 @@ omega.o : omega.c omega.h $(CONFIG_H) $(SYSTEM_H) coretypes.h $(TM_H) \
    $(GGC_H) $(TREE_H) $(DIAGNOSTIC_H) varray.h $(TREE_PASS_H) $(PARAMS_H)
 tree-chrec.o: tree-chrec.c $(CONFIG_H) $(SYSTEM_H) coretypes.h $(TM_H) \
    $(GGC_H) $(TREE_H) $(REAL_H) $(SCEV_H) $(TREE_PASS_H) $(PARAMS_H) \
-   $(DIAGNOSTIC_H) $(CFGLOOP_H) $(TREE_FLOW_H)
+   $(DIAGNOSTIC_H) $(CFGLOOP_H) $(TREE_FLOW_H) $(FLAGS_H)
 tree-scalar-evolution.o: tree-scalar-evolution.c $(CONFIG_H) $(SYSTEM_H) \
    coretypes.h $(TM_H) $(GGC_H) $(TREE_H) $(REAL_H) $(RTL_H) \
    $(BASIC_BLOCK_H) $(DIAGNOSTIC_H) $(TREE_FLOW_H) $(TREE_DUMP_H) \

gcc/testsuite/ChangeLog

+2009-07-03  Richard Guenther  <rguenther@suse.de>
+
+	PR middle-end/34163
+	* gfortran.dg/pr34163.f90: New testcase.
+
 2009-07-03  Jan Hubicka  <jh@suse.cz>
 
 	* gcc.dg/tree-ssa/loop-24.c: Update dump file matching; enable -O2.

gcc/testsuite/gfortran.dg/pr34163.f90

+! { dg-do compile }
+! { dg-options "-O2 -fpredictive-commoning -fdump-tree-pcom-details" }
+subroutine trisolve2(x,i1,i2,nxyz)
+integer :: nxyz
+real,dimension(nxyz):: au1
+real,allocatable,dimension(:) :: gi
+integer :: i1 , i2
+real,dimension(i2)::x
+integer :: i
+allocate(gi(nxyz))
+do i = i1+1 , i2
+   x(i) = gi(i)*(x(i)-au1(i-1)*x(i-1))
+enddo
+end subroutine trisolve2
+! { dg-final { scan-tree-dump "Executing predictive commoning" "pcom" } }
+! { dg-final { cleanup-tree-dump "pcom" } }

gcc/tree-chrec.c

@@ -37,6 +37,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "tree-chrec.h"
 #include "tree-pass.h"
 #include "params.h"
+#include "flags.h"
 #include "tree-scalar-evolution.h"
 
 
@@ -1286,7 +1287,19 @@ chrec_convert_1 (tree type, tree chrec, gimple at_stmt,
 
   /* If we cannot propagate the cast inside the chrec, just keep the cast.  */
 keep_cast:
-  res = fold_convert (type, chrec);
+  /* Fold will not canonicalize (long)(i - 1) to (long)i - 1 because that
+     may be more expensive.  We do want to perform this optimization here
+     though for canonicalization reasons.  */
+  if (use_overflow_semantics
+      && (TREE_CODE (chrec) == PLUS_EXPR
+	  || TREE_CODE (chrec) == MINUS_EXPR)
+      && TYPE_PRECISION (type) > TYPE_PRECISION (ct)
+      && TYPE_OVERFLOW_UNDEFINED (ct))
+    res = fold_build2 (TREE_CODE (chrec), type,
+		       fold_convert (type, TREE_OPERAND (chrec, 0)),
+		       fold_convert (type, TREE_OPERAND (chrec, 1)));
+  else
+    res = fold_convert (type, chrec);
 
   /* Don't propagate overflows.  */
   if (CONSTANT_CLASS_P (res))