https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98680
Bug ID: 98680
Summary: Several 30_threads tests are flaky
Product: gcc
Version: 11.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: libstdc++
Assignee: unassigned at gcc dot gnu.org
Reporter: ro at gcc dot gnu.org
Target Milestone: ---
Target: i386-pc-solaris2.11, sparc-sun-solaris2.11,
i686-apple-darwin*, x86_64-apple-darwin*,
hppa2.0w-hp-hpux11.11
Several 30_threads tests are flaky on Solaris, more often on Solaris/x86 but
also on SPARC:
For the last couple of days, the counts are like this:
11 30_threads/barrier/arrive_and_wait.cc execution test
11 30_threads/barrier/arrive_and_wait.cc execution test program timed out.
1 30_threads/barrier/arrive.cc execution test
1 30_threads/barrier/arrive.cc execution test program timed out.
20 30_threads/barrier/completion.cc execution test
20 30_threads/barrier/completion.cc execution test program timed out.
14 30_threads/latch/3.cc execution test
14 30_threads/latch/3.cc execution test program timed out.
1 30_threads/semaphore/try_acquire_for.cc execution test
1 30_threads/semaphore/try_acquire_for.cc execution test program timed
out.
5 30_threads/semaphore/try_acquire_until.cc execution test
5 30_threads/semaphore/try_acquire_until.cc execution test program timed
out.
At least Darwin and HP-UX are likewise affected; I haven't looked closer for
possible others.
As an example, I ran arrive_and_wait.exe for ca. 100 times with the following
timing:
93 real 0.01
1 real 0.73
1 real 0.95
1 real 1.03
1 real 1:13.00