Adjust top level script

[jeffmurphy]

Create a top level python script that will descend into each folder and run the given benchmark. The python script should

• accept the same arguments as the individual benchmarks (see Adjust command line arguments #1).
• capture output
• parse out the compute:start and compute:stop records
• separate individual threads by ID
• create CSV files showing threadid,startdelta,runlength where runlength is the difference between the stop and start times for the given threadid, and startdelta is the difference between the start time of the given threadid and the start time of the previous threadid (defined to be 0 for the first record)
• emit graphs using matplotlib showing periodicity of iterations as well as run times.

So the repo should look something like:

./runthemall.py
./erlang/benchmark.erl
./haskell/benchmark.hs

etc

and should, post run, leave the raw data and graphs in each folder