ParseHPMC.py giving incorrect results

[jordancarlin]

#1367 got all of the benchmarking and performance validation scripts running again, but ParseHPMC.py seems to be giving incorrect results. The results it computes from a Wally log diverge significantly from the reference model, and in some cases trend in the opposite direction from the expected results. The problem is most likely somewhere in the parsing script and not in the branch predictor itself, but this needs to be confirmed.