Now, there is logic for generating collective/indpendent i/o events
from darshan logs. There are a couple minor timing and i/o parameter
bugs that still need to be worked out.

Update the darshan workload generator to use new collective open/close
generation algorithm. Previously, generating these operations was
done sequentially on one process, so the algorithm has to be updated
to give consistent results in a distributed environment.

The codes-workload-test and codes-workload-mpi-replay both support
the new paramters of the bgp-io language workload generator. Also,
support has been added to the config file and codes-workload-mpi-replay
program for loading the parameters of a darshan config file.

Now, the darshan generator can generate independent file events, as
well as basic collective file events. Still working on rewriting
collective events.

Pull all darshan workload generator logic into the workload generator
API. Nondeterminism (i.e. using rand()) still needs to be removed from
generating collective I/O events, since they will now be generated in
a collective manner -- not independently as before. Also, still need
to sort out collective timing issues.

Now, specify the darshan install directory with --with-darshan=/dir/
to build the darshan workload generator and link it with test
applications.

Modifications to autotools to include necessary info for building
the darshan workload API. This API compiles fine, but the mpi replay
test is not linking properly with darshan library.

Pull all darshan workload generator logic into the workload generator
API. Nondeterminism (i.e. using rand()) still needs to be removed from
generating collective I/O events, since they will now be generated in
a collective manner -- not independently as before. Also, still need
to sort out collective timing issues.

Now, specify the darshan install directory with --with-darshan=/dir/
to build the darshan workload generator and link it with test
applications.

Modifications to autotools to include necessary info for building
the darshan workload API. This API compiles fine, but the mpi replay
test is not linking properly with darshan library.

updates to codes workloads: switched to hash tables in the workload--i/o lang plug-in, removed codes_workload_get_info from i/o lang and darshan workloads as this is no longer needed in the bg/p model

Improve the performance of the darshan i/o workload by converting the
linked list of per-rank contexts to a hash table.

Update the output directory argument of the mpi-replay tool to just
specify an already existing directory to perform the workload I/O in.

Modify the mpi-replay tool (located at tests/workload) to spit out
verbose information on each event it replays, if user specifies
-v option. Also, modify the mpi-replay tool to perform no-ops for
each replayed event, if user specifies --noop option. The replay
tool should function properly for all workload types now.

TODO: The tool should offer functionality for initializing the context
of the output replay directory. That is, the tool should first scan
the given workload and create all files and preallocate space for all
I/O ops in the workload. Currently, the replay tool will fail on any
files that are not explicitly created by the workload.

adding the mpi-replay tool to the tests/workload directory for replaying workloads from the codes workload api

(1) adding sample codes example and (2) updating codes best practices document with the CODES example