Bug 83043 - [11/12/13/14 Regression] FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times graphite "2 loops carried no dependency" 1 (found 0 times)
Summary: [11/12/13/14 Regression] FAIL: libgomp.graphite/force-parallel-1.c scan-tree-...
Status: NEW
Alias: None
Product: gcc
Classification: Unclassified
Component: tree-optimization (show other bugs)
Version: 8.0
: P4 normal
Target Milestone: 11.5
Assignee: Not yet assigned to anyone
URL:
Keywords: testsuite-fail
: 83082 (view as bug list)
Depends on:
Blocks:
 
Reported: 2017-11-19 09:35 UTC by Tom de Vries
Modified: 2023-07-07 10:32 UTC (History)
2 users (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed: 2018-01-10 00:00:00


Attachments
force-parallel-1.c.150t.graphite (4.94 KB, text/plain)
2017-11-19 09:48 UTC, Tom de Vries
Details
force-parallel-1.c.154t.parloops2 (4.64 KB, text/plain)
2017-11-19 09:48 UTC, Tom de Vries
Details
force-parallel-1.c.228t.optimized (1.94 KB, text/plain)
2017-11-19 09:49 UTC, Tom de Vries
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Tom de Vries 2017-11-19 09:35:02 UTC
https://gcc.gnu.org/ml/gcc-regression/2017-11/msg00290.html :
...
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times graphite "2 loops carried no dependency" 1 (found 0 times)
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times graphite "2 loops carried no dependency" 1 (found 0 times)
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times optimized "loopfn" 8 (found 12 times)
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times optimized "loopfn" 8 (found 12 times)...

Regressions on trunk at revision 254889 vs revision 254872
Comment 1 Tom de Vries 2017-11-19 09:48:02 UTC
Created attachment 42648 [details]
force-parallel-1.c.150t.graphite

instead of "2 loops carried no dependency" 1 we have:
...
$ grep loops force-parallel-1.c.150t.graphite 
0 loops carried no dependency.
0 loops carried no dependency.
0 loops carried no dependency.
...
Comment 2 Tom de Vries 2017-11-19 09:48:43 UTC
Created attachment 42649 [details]
force-parallel-1.c.154t.parloops2
Comment 3 Tom de Vries 2017-11-19 09:49:06 UTC
Created attachment 42650 [details]
force-parallel-1.c.228t.optimized
Comment 4 Tom de Vries 2017-11-19 09:57:48 UTC
Regressing commit: r254888.

...
commit 7ae0128a031e2fd2f324f3853560f10e604c1812
Author: hubicka <hubicka@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Fri Nov 17 17:47:36 2017 +0000

        * predict.c (determine_unlikely_bbs): Set cgraph node count to 0
	when entry block was promoted unlikely.
        (estimate_bb_frequencies): Increase frequency scale.
	* profile-count.h (profile_count): Export precision info.
	* gcc.dg/tree-ssa/dump-2.c: Fixup template for profile precision
	changes.
        * gcc.dg/tree-ssa/pr77445-2.c: Fixup template for profile precision
	changes.


    git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@254888 138bc75d-0d04-0410-961f-82ee72b054a4
...
Comment 5 Richard Biener 2017-11-20 09:13:08 UTC
Possibly some optimize_loop_nest_for_speed check changes.
Comment 6 Jakub Jelinek 2017-11-21 12:23:29 UTC
*** Bug 83082 has been marked as a duplicate of this bug. ***
Comment 7 Andrey Guskov 2017-11-30 10:56:38 UTC
Confirmed on Silvermont.
Comment 8 Jakub Jelinek 2018-02-15 15:52:27 UTC
This is just graphite, shouldn't it be P4?
Comment 9 Jan Hubicka 2018-02-15 16:01:43 UTC
> This is just graphite, shouldn't it be P4?

I would say so. Indeed it is most probably some loop becoming cold and thus quite harmless bug.  I will look into it.
Comment 10 Tom de Vries 2018-02-21 07:56:59 UTC
https://gcc.gnu.org/ml/gcc-regression/2018-01/msg00362.html :
...
New passes:
  ...
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times graphite "2 loops carried no dependency" 1 (found 0 times)
FAIL: libgomp.graphite/force-parallel-1.c scan-tree-dump-times optimized "loopfn" 8 (found 12 times)
...
Comment 11 Martin Liška 2018-03-14 18:04:50 UTC
I also believe it's not P1, thus adjusting to P4.

Tom I would add:

diff --git a/libgomp/testsuite/libgomp.graphite/force-parallel-1.c b/libgomp/testsuite/libgomp.graphite/force-parallel-1.c
index 0393356f9f2..01ac124f0a0 100644
--- a/libgomp/testsuite/libgomp.graphite/force-parallel-1.c
+++ b/libgomp/testsuite/libgomp.graphite/force-parallel-1.c
@@ -1,3 +1,5 @@
+/* { dg-options "-fno-ipa-partial-inlining" } */
+
 void abort (void);
 
 int x[10000000];

Because IPA fn split does not play any role in the test. Then I would adjust expected output.
Comment 12 Jakub Jelinek 2018-05-02 10:11:09 UTC
GCC 8.1 has been released.
Comment 13 Jakub Jelinek 2018-07-26 11:23:36 UTC
GCC 8.2 has been released.
Comment 14 Jakub Jelinek 2019-02-22 15:25:39 UTC
GCC 8.3 has been released.
Comment 15 Jakub Jelinek 2020-03-04 09:52:10 UTC
GCC 8.4.0 has been released, adjusting target milestone.
Comment 16 Jakub Jelinek 2021-05-14 09:49:30 UTC
GCC 8 branch is being closed.
Comment 17 Richard Biener 2021-06-01 08:09:35 UTC
GCC 9.4 is being released, retargeting bugs to GCC 9.5.
Comment 18 Richard Biener 2022-05-27 09:37:50 UTC
GCC 9 branch is being closed
Comment 19 Jakub Jelinek 2022-06-28 10:33:59 UTC
GCC 10.4 is being released, retargeting bugs to GCC 10.5.
Comment 20 Richard Biener 2023-07-07 10:32:44 UTC
GCC 10 branch is being closed.