For data processing and graphing, i.e., process_data.py (previously called plot.py)
fixed type 1 and type 2 utilization calculation bug that didn't
weight by vm mem size correctly
add calculation for idle time
add scheduler latency calculation
utilization graph now uses a rolling window average instead of
sampling
For controller
added scheduler latency measurement
Also added a simple script as a base for building more automated
experiments.
I created this PR in case you want some details on what was changed, particularly on how data is now processed. I'll update the basecamp post soon to show current results which will be more tangible and easier to interpret than code changes in this PR =)
But on a high level, I'm much more comfortable with the numbers and calculations we're getting now. and I believe this PR is ready to merge.
For data processing and graphing, i.e.,
process_data.py(previously calledplot.py)For controller
Also added a simple script as a base for building more automated experiments.
I created this PR in case you want some details on what was changed, particularly on how data is now processed. I'll update the basecamp post soon to show current results which will be more tangible and easier to interpret than code changes in this PR =) But on a high level, I'm much more comfortable with the numbers and calculations we're getting now. and I believe this PR is ready to merge.