Hi,
For loops which are bounded to iterate only 1 time (thus loop's latch doesn't
roll), there is nothing to predictive common, this patch detects/skips these
cases. A test is also added in
gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f for this.
Thanks,
bin
2016-09-01 Bin Cheng <bin.ch...@arm.com>
* tree-predcom.c (tree_predictive_commoning_loop): Skip loop that only
iterates 1 time.
gcc/testsuite/ChangeLog
2016-09-01 Bin Cheng <bin.ch...@arm.com>
* gfortran.dg/vect/fast-math-mgrid-resid.f: New test string.
diff --git a/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f
b/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f
index fcf475d..88238f9 100644
--- a/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f
+++ b/gcc/testsuite/gfortran.dg/vect/fast-math-mgrid-resid.f
@@ -42,3 +42,4 @@ C
! { dg-final { scan-tree-dump-times "Executing predictive commoning without
unrolling" 1 "pcom" { target lp64 } } }
! { dg-final { scan-tree-dump-times "Executing predictive commoning without
unrolling" 2 "pcom" { target ia32 } } }
! { dg-final { scan-tree-dump-times "Predictive commoning failed: no suitable
chains" 0 "pcom" } }
+! { dg-final { scan-tree-dump-times "Loop iterates only 1 time, nothing to do"
1 "pcom" } }
diff --git a/gcc/tree-predcom.c b/gcc/tree-predcom.c
index 096365a..14d53c2 100644
--- a/gcc/tree-predcom.c
+++ b/gcc/tree-predcom.c
@@ -2465,6 +2465,15 @@ tree_predictive_commoning_loop (struct loop *loop)
if (dump_file && (dump_flags & TDF_DETAILS))
fprintf (dump_file, "Processing loop %d\n", loop->num);
+ /* Nothing for predicitive commoning if loop only iterates 1 time. */
+ if (get_max_loop_iterations_int (loop) == 0)
+ {
+ if (dump_file && (dump_flags & TDF_DETAILS))
+ fprintf (dump_file, "Loop iterates only 1 time, nothing to do.\n");
+
+ return false;
+ }
+
/* Find the data references and split them into components according to their
dependence relations. */
auto_vec<loop_p, 3> loop_nest;