See http://hmmer.janelia.org/. Problem reported by Chris Daley. Example execution reads gigabytes of data and writes 19 MiB of data, but Darshan reports 8 bytes read and roughly 1 MiB written. Application is using fread and fwrite for I/O, and everything shows up as expected in strace as read and write operations.

This indicates that there may be a bug in how Darshan records the sizes of fread and fwrite operations.

Removing from 2.2.9 milestone for now, but we should still try to produce a prototype to test with hmmer.