From: Jan Hubicka <hubicka@ucw.cz>
To: Richard Biener <rguenther@suse.de>
Cc: Jan Hubicka <hubicka@ucw.cz>, gcc-patches@gcc.gnu.org
Subject: Re: Do not give realistic estimates for loop with array accesses
Date: Wed, 30 Mar 2016 12:36:00 -0000 [thread overview]
Message-ID: <20160330122751.GA97459@kam.mff.cuni.cz> (raw)
In-Reply-To: <alpine.LSU.2.11.1603301325530.13384@t29.fhfr.qr>
>
> You are only changing one place in this file.
You are right. I am attaching the updated patch which I am re-testing now.
>
> The vectorizer already checks this (albeit indirectly):
>
> HOST_WIDE_INT max_niter
> = max_stmt_executions_int (LOOP_VINFO_LOOP (loop_vinfo));
> if ((LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
> && (LOOP_VINFO_INT_NITERS (loop_vinfo) < vectorization_factor))
> || (max_niter != -1
> && (unsigned HOST_WIDE_INT) max_niter < vectorization_factor))
> {
> if (dump_enabled_p ())
> dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
> "not vectorized: iteration count smaller than "
> "vectorization factor.\n");
> return false;
> }
Yes, but one tests only vectorization_factor and other min_profitable_estimate
which probably should be greater than vectorization_factor.
The check above should therefore become redundant. My reading of the code is
that min_profiltable_estimate is computed after the check above, so it is
probably an useful shortcut and the message is also bit more informative.
I updated the later test to use max_niter variable once it is computed.
OK with those changes assuming testing passes?
Honza
* tree-ssa-loop-niter.c (idx_infer_loop_bounds): We can't get realistic
estimates here.
* tree-ssa-loop-unswitch.c (tree_unswitch_single_loop): Use also
max_loop_iterations_int.
(tree_unswitch_outer_loop): Likewise.
* tree-ssa-loop-ivopts.c (avg_loop_niter): Likewise.
* tree-vect-loop.c (vect_analyze_loop_2): Likewise.
Index: tree-ssa-loop-ivopts.c
===================================================================
--- tree-ssa-loop-ivopts.c (revision 234516)
+++ tree-ssa-loop-ivopts.c (working copy)
@@ -121,7 +121,11 @@ avg_loop_niter (struct loop *loop)
{
HOST_WIDE_INT niter = estimated_stmt_executions_int (loop);
if (niter == -1)
- return AVG_LOOP_NITER (loop);
+ {
+ niter = max_stmt_executions_int (loop);
+ if (niter == -1 || niter > AVG_LOOP_NITER (loop))
+ return AVG_LOOP_NITER (loop);
+ }
return niter;
}
Index: tree-ssa-loop-niter.c
===================================================================
--- tree-ssa-loop-niter.c (revision 234516)
+++ tree-ssa-loop-niter.c (working copy)
@@ -3115,7 +3115,6 @@ idx_infer_loop_bounds (tree base, tree *
tree low, high, type, next;
bool sign, upper = true, at_end = false;
struct loop *loop = data->loop;
- bool reliable = true;
if (TREE_CODE (base) != ARRAY_REF)
return true;
@@ -3187,14 +3186,14 @@ idx_infer_loop_bounds (tree base, tree *
&& tree_int_cst_compare (next, high) <= 0)
return true;
- /* If access is not executed on every iteration, we must ensure that overlow may
- not make the access valid later. */
+ /* If access is not executed on every iteration, we must ensure that overlow
+ may not make the access valid later. */
if (!dominated_by_p (CDI_DOMINATORS, loop->latch, gimple_bb (data->stmt))
&& scev_probably_wraps_p (initial_condition_in_loop_num (ev, loop->num),
step, data->stmt, loop, true))
- reliable = false;
+ upper = false;
- record_nonwrapping_iv (loop, init, step, data->stmt, low, high, reliable, upper);
+ record_nonwrapping_iv (loop, init, step, data->stmt, low, high, false, upper);
return true;
}
Index: tree-ssa-loop-unswitch.c
===================================================================
--- tree-ssa-loop-unswitch.c (revision 234516)
+++ tree-ssa-loop-unswitch.c (working copy)
@@ -223,6 +223,8 @@ tree_unswitch_single_loop (struct loop *
/* If the loop is not expected to iterate, there is no need
for unswitching. */
iterations = estimated_loop_iterations_int (loop);
+ if (iterations < 0)
+ iterations = max_loop_iterations_int (loop);
if (iterations >= 0 && iterations <= 1)
{
if (dump_file && (dump_flags & TDF_DETAILS))
@@ -439,6 +441,8 @@ tree_unswitch_outer_loop (struct loop *l
/* If the loop is not expected to iterate, there is no need
for unswitching. */
iterations = estimated_loop_iterations_int (loop);
+ if (iterations < 0)
+ iterations = max_loop_iterations_int (loop);
if (iterations >= 0 && iterations <= 1)
{
if (dump_file && (dump_flags & TDF_DETAILS))
Index: tree-vect-loop.c
===================================================================
--- tree-vect-loop.c (revision 234516)
+++ tree-vect-loop.c (working copy)
@@ -2063,6 +2063,8 @@ start_over:
estimated_niter
= estimated_stmt_executions_int (LOOP_VINFO_LOOP (loop_vinfo));
+ if (estimated_niter != -1)
+ estimated_niter = max_niter;
if (estimated_niter != -1
&& ((unsigned HOST_WIDE_INT) estimated_niter
<= MAX (th, (unsigned)min_profitable_estimate)))
next prev parent reply other threads:[~2016-03-30 12:27 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-30 11:02 Jan Hubicka
2016-03-30 12:09 ` Richard Biener
2016-03-30 12:36 ` Jan Hubicka [this message]
2016-03-30 12:49 ` Richard Biener
2016-03-30 18:52 ` Bernhard Reutner-Fischer
2016-04-07 14:52 ` Tom de Vries
2016-03-30 13:50 ` Bin.Cheng
2016-03-30 14:41 ` Jan Hubicka
2016-03-30 15:30 ` Bin.Cheng
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160330122751.GA97459@kam.mff.cuni.cz \
--to=hubicka@ucw.cz \
--cc=gcc-patches@gcc.gnu.org \
--cc=rguenther@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).