GFAST crashes during processing after running for 10-20 hours.
-> I cannot reproduce this behavior in on-demand fake earthquake tests even with identical source parameters.
-> I'm growing less confident this is the fault of a memory leak.
-> My current working hypothesis is this is in the data structure copying in H5.
-> GFAST seems much more reliable when using HDF5 v 1.10.1. or 1.10.0_patch.
If GFAST crashes on you then email Ben Baker the output.
GFAST crashes during processing after running for 10-20 hours.
-> I cannot reproduce this behavior in on-demand fake earthquake tests even with identical source parameters.
-> I'm growing less confident this is the fault of a memory leak.
-> My current working hypothesis is this is in the data structure copying in H5.
-> GFAST seems much more reliable when using HDF5 v 1.10.1. or 1.10.0_patch.
If GFAST crashes on you then email Ben Baker the output.