From a243ce2a52a6c62bc0d6be0b756a85dd9c1bceb7 Mon Sep 17 00:00:00 2001
From: Richard Biener <rguenther@suse.de>
Date: Thu, 9 Mar 2023 09:02:07 +0100
Subject: [PATCH] tree-optimization/44794 - avoid excessive RTL unrolling on
epilogues
The following adjusts tree_[transform_and_]unroll_loop to set an
upper bound on the number of iterations on the epilogue loop it
creates. For the testcase at hand which involves array prefetching
this avoids applying RTL unrolling to them when -funroll-loops is
specified.
Other users of this API includes predictive commoning and
unroll-and-jam.
PR tree-optimization/44794
* tree-ssa-loop-manip.cc (tree_transform_and_unroll_loop):
If an epilogue loop is required set its iteration upper bound.
---
gcc/tree-ssa-loop-manip.cc | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/gcc/tree-ssa-loop-manip.cc b/gcc/tree-ssa-loop-manip.cc
index 4ef27bae5152..a52277abdbf5 100644
--- a/gcc/tree-ssa-loop-manip.cc
+++ b/gcc/tree-ssa-loop-manip.cc
@@ -1297,6 +1297,12 @@ tree_transform_and_unroll_loop (class loop *loop, unsigned factor,
}
remove_path (exit);
+
+ /* The epilog loop latch executes at most factor - 1 times.
+ Since the epilog is entered unconditionally it will need to handle
+ up to factor executions of its body. */
+ new_loop->any_upper_bound = 1;
+ new_loop->nb_iterations_upper_bound = factor - 1;
}
else
new_exit = single_dom_exit (loop);
--
GitLab