Opened 10 years ago
Closed 7 years ago
#1235 closed Bug (wontfix)
Division by zero in timing routines
Reported by: | ufla | Owned by: | nemo |
---|---|---|---|
Priority: | low | Milestone: | |
Component: | OCE | Version: | trunk |
Severity: | Keywords: | ||
Cc: |
Description
Subroutine wcurrent_info in timing.F90 contains this code:
DO WHILE ( ASSOCIATED(s_timer) ) WRITE(numtime,TRIM(clfmt)) s_timer%cname, & & s_timer%tsum_clock,s_timer%tsum_clock*100./t_elaps(2), & & s_timer%tsum_cpu ,s_timer%tsum_cpu*100./t_cpu(2) , & & s_timer%tsum_cpu/s_timer%tsum_clock, s_timer%niter s_timer => s_timer%next END DO
However, s_timer%tsum_clock can be zero, which causes the code to crash if floating point exceptions are enabled. This was tested with TotalView? finding that tsum_clock was zero for "dia_fwb" in a test run.
Not sure if tsum_clock can actually be zero for some routines or if this is an issue of timer granularity/accuracy.
Commit History (0)
(No commits)
Change History (2)
comment:1 Changed 9 years ago by ufla
- Type changed from Development branch to Bug
- Version changed from v3.6 to trunk
comment:2 Changed 7 years ago by clevy
- Resolution set to wontfix
- Status changed from new to closed
Note: See
TracTickets for help on using
tickets.
Suppressing tickets opened on the trunk since more than 2 years... cleaning started!