Fixing bug which caused enzo to hang when running with enzo-performance module and restarting with more processors than used to write out restart file.

#296 Merged at 1ae8e42
Repository
chummels
Branch
week-of-code
Repository
enzo
Branch
week-of-code
Author
  1. chummels
Reviewers
Description

This addresses an issue recently discussed on the mailing list involving the enzo performance module and restarting enzo using N processors on an output generated with M processors where N>M. When that happens, the enzo timing module for SolveHydroEquations isn't yet initialized since the N-M processors without a root grid don't have any data to solve until LoadBalancing kicks in on the second cycle. This error results in a hang.

This PR avoids that issue by initializing all of the default timers in enzo.C before any of them can be called to start or stop.

Comments (2)