My mainline testing gets intermittent failures for test libgomp/testsuite/libgomp.g++/pr27337.c. Some of the failures are due to wrong counts from the constructor and destructor after the call to foo, and others are for wrong values of x.i (with values in the range 19-27) after the call to bar. The percentage of failures depends on which system I test on, but I haven't figured out the significant factors: CPU type, number of processors, binutils version, kernel version, load on the system
I haven't studied omp programs; the failures could be due to errors in the test itself. The failures occur at all levels of optimization. I get this failure on powerpc*-linux (with both -m32 and -m64) and on x86_64-linux. -- Summary: libgomp test pr27337.C fails intermittently Product: gcc Version: 4.2.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: libgomp AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: janis at gcc dot gnu dot org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28046