public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc/devel/omp/gcc-13] openmp: Fix number of iterations computation for "omp unroll full"
@ 2023-07-27 11:57 Frederik Harwath
  0 siblings, 0 replies; only message in thread
From: Frederik Harwath @ 2023-07-27 11:57 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:e59a5a62eb52bf1dac06f85f45e71426c5423b8c

commit e59a5a62eb52bf1dac06f85f45e71426c5423b8c
Author: Frederik Harwath <frederik@codesourcery.com>
Date:   Wed Jul 26 18:27:33 2023 +0200

    openmp: Fix number of iterations computation for "omp unroll full"
    
    gcc/ChangeLog:
    
            * omp-transform-loops.cc (gomp_for_number_of_iterations):
            Always compute "final - init" and do not take absolute value.
            Identify non-iterating and infinite loops for constant init,
            final, step values for better diagnostic messages, consistent
            behaviour in those corner cases, and better testability.
            (gomp_for_constant_iterations_p): Add new argument to pass
            on information about infinite loops, and ...
            (full_unroll): ... use from here to emit a warning and remove
            unrolled, known infinite loops consistently.
            (process_omp_for): Only print dump message if loop has not
            been removed by transformation.
    
    gcc/testsuite/ChangeLog:
    
            * c-c++-common/gomp/loop-transforms/unroll-8.c: New test.

Diff:
---
 gcc/omp-transform-loops.cc                         | 94 ++++++++++++++++------
 .../c-c++-common/gomp/loop-transforms/unroll-8.c   | 76 +++++++++++++++++
 2 files changed, 146 insertions(+), 24 deletions(-)

diff --git a/gcc/omp-transform-loops.cc b/gcc/omp-transform-loops.cc
index e7a038c2672..281951d9697 100644
--- a/gcc/omp-transform-loops.cc
+++ b/gcc/omp-transform-loops.cc
@@ -153,20 +153,27 @@ subst_defs (tree expr, gimple_seq seq)
   return expr;
 }
 
-/* Return an expression for the number of iterations of the outermost loop of
-   OMP_FOR. */
+/* Return an expression for the number of iterations of the loop at
+   the given LEVEL of OMP_FOR.
+
+   If the expression is a negative constant, this means that the loop
+   is infinite. This can only be recognized for loops with constant
+   initial, final, and step values.  In general, according to the
+   OpenMP specification, the behaviour is unspecified if the number of
+   iterations does not fit the types used for their computation, and
+   hence in particular if the loop is infinite. */
 
 tree
 gomp_for_number_of_iterations (const gomp_for *omp_for, size_t level)
 {
   gcc_assert (!non_rectangular_p (omp_for));
-
   tree init = gimple_omp_for_initial (omp_for, level);
   tree final = gimple_omp_for_final (omp_for, level);
   tree_code cond = gimple_omp_for_cond (omp_for, level);
   tree index = gimple_omp_for_index (omp_for, level);
   tree type = gomp_for_iter_count_type (index, final);
-  tree step = TREE_OPERAND (gimple_omp_for_incr (omp_for, level), 1);
+  tree incr = gimple_omp_for_incr (omp_for, level);
+  tree step = omp_get_for_step_from_incr (gimple_location (omp_for), incr);
 
   init = subst_defs (init, gimple_omp_for_pre_body (omp_for));
   init = fold (init);
@@ -181,34 +188,64 @@ gomp_for_number_of_iterations (const gomp_for *omp_for, size_t level)
       diff_type = ptrdiff_type_node;
     }
 
-  tree diff;
-  if (cond == GT_EXPR)
-    diff = fold_build2 (minus_code, diff_type, init, final);
-  else if (cond == LT_EXPR)
-    diff = fold_build2 (minus_code, diff_type, final, init);
-  else
-    gcc_unreachable ();
 
-  diff = fold_build2 (CEIL_DIV_EXPR, type, diff, step);
-  diff = fold_build1 (ABS_EXPR, type, diff);
+  /* Identify a simple case in which the loop does not iterate. The
+     computation below could not tell this apart from an infinite
+     loop, hence we handle this separately for better diagnostic
+     messages. */
+  gcc_assert (cond == GT_EXPR || cond == LT_EXPR);
+  if (TREE_CONSTANT (init) && TREE_CONSTANT (final)
+      && ((cond == GT_EXPR && tree_int_cst_le (init, final))
+	  || (cond == LT_EXPR && tree_int_cst_le (final, init))))
+    return build_int_cst (diff_type, 0);
+
+  tree diff = fold_build2 (minus_code, diff_type, final, init);
+
+  /* Divide diff by the step.
+
+     We could always use CEIL_DIV_EXPR since only non-negative results
+     correspond to valid number of iterations and the behaviour is
+     unspecified by the spec otherwise. But we try to get the rounding
+     right for constant negative values to identify infinite loops
+     more precisely for better warnings. */
+  tree_code div_expr = CEIL_DIV_EXPR;
+  if (TREE_CONSTANT (diff) && TREE_CONSTANT (step))
+    {
+      bool diff_is_neg = tree_int_cst_lt (diff, size_zero_node);
+      bool step_is_neg = tree_int_cst_lt (step, size_zero_node);
+      if ((diff_is_neg && !step_is_neg)
+	  || (!diff_is_neg && step_is_neg))
+	div_expr = FLOOR_DIV_EXPR;
+    }
 
+  diff = fold_build2 (div_expr, type, diff, step);
   return diff;
 }
 
-/* Return true if the expression representing the number of iterations for
-   OMP_FOR is a constant expression, false otherwise. */
+/* Return true if the expression representing the number of iterations
+   for OMP_FOR is a non-negative constant and set ITERATIONS to the
+   value of that expression. Otherwise, return false.  Set INFINITE to
+   true if the number of iterations was recognized to be infinite. */
 
 bool
 gomp_for_constant_iterations_p (gomp_for *omp_for,
-				unsigned HOST_WIDE_INT *iterations)
+				unsigned HOST_WIDE_INT *iterations,
+				bool *infinite = NULL)
 {
   tree t = gomp_for_number_of_iterations (omp_for, 0);
-  if (!TREE_CONSTANT (t)
-      || !tree_fits_uhwi_p (t))
+  if (!TREE_CONSTANT (t))
     return false;
 
-  *iterations = tree_to_uhwi (t);
-  return true;
+  if (infinite &&
+      tree_int_cst_lt (t, size_zero_node))
+    *infinite = true;
+  else if (tree_fits_uhwi_p (t))
+    {
+      *iterations = tree_to_uhwi (t);
+      return true;
+    }
+
+  return false;
 }
 
 static gimple_seq
@@ -519,10 +556,18 @@ full_unroll (gomp_for *omp_for, location_t loc, walk_ctx *ctx ATTRIBUTE_UNUSED)
 {
   tree init = gimple_omp_for_initial (omp_for, 0);
   unsigned HOST_WIDE_INT niter = 0;
-  if (!gomp_for_constant_iterations_p (omp_for, &niter))
+  bool infinite = false;
+  bool constant = gomp_for_constant_iterations_p (omp_for, &niter, &infinite);
+
+  if (infinite)
+    {
+      warning_at (loc, 0, "Cannot apply full unrolling to infinite loop");
+      return NULL;
+    }
+  if (!constant)
     {
       error_at (loc, "Cannot apply full unrolling to loop with "
-		     "non-constant number of iterations");
+		"non-constant number of iterations");
       return omp_for;
     }
 
@@ -1589,8 +1634,9 @@ process_omp_for (gomp_for *omp_for, gimple_seq *containing_seq, walk_ctx *ctx)
   if (!dump_enabled_p () || !(dump_flags & TDF_DETAILS))
     return;
 
-  dump_printf_loc (MSG_NOTE | MSG_PRIORITY_INTERNALS, transformed,
-		   "Transformed loop: %G\n\n", transformed);
+  if (transformed)
+    dump_printf_loc (MSG_NOTE | MSG_PRIORITY_INTERNALS, transformed,
+		     "Transformed loop: %G\n\n", transformed);
 }
 
 /* Traverse SEQ in depth-first order and apply the loop transformation
diff --git a/gcc/testsuite/c-c++-common/gomp/loop-transforms/unroll-8.c b/gcc/testsuite/c-c++-common/gomp/loop-transforms/unroll-8.c
new file mode 100644
index 00000000000..d49d7c42c87
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/gomp/loop-transforms/unroll-8.c
@@ -0,0 +1,76 @@
+extern void dummy(int);
+
+void
+test1 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = 101; i > 100; i++)
+    dummy (i);
+}
+
+
+void
+test2 ()
+{
+#pragma omp unroll full
+  for (int i = 101; i != 100; i++)
+    dummy (i);
+}
+
+void
+test3 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = 0; i <= 0; i--)
+    dummy (i);
+}
+
+void
+test4 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = 101; i > 100; i=i+2)
+    dummy (i);
+}
+
+void
+test5 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = -101; i < 100; i=i-10)
+    dummy (i);
+}
+
+void
+test6 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = -101; i < 100; i=i-300)
+    dummy (i);
+}
+
+void
+test7 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = 101; i > -100; i=i+300)
+    dummy (i);
+
+  /* Loop does not iterate, hence no warning. */
+#pragma omp unroll full
+  for (int i = 101; i > 101; i=i+300)
+    dummy (i);
+}
+
+void
+test8 ()
+{
+#pragma omp unroll full /* { dg-warning "Cannot apply full unrolling to infinite loop" } */
+  for (int i = -21; i < -20; i=i-40)
+    dummy (i);
+
+  /* Loop does not iterate, hence no warning. */
+#pragma omp unroll full
+  for (int i = -21; i > 20; i=i-40)
+    dummy (i);
+}

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-07-27 11:57 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-07-27 11:57 [gcc/devel/omp/gcc-13] openmp: Fix number of iterations computation for "omp unroll full" Frederik Harwath

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).