experiment-scripts.git/parse/sched.py, branch wip-ecrts14-pgm

Compute stats for response time percentiles

2014-01-30T04:25:46+00:00

Compute proportional graph response times.

2014-01-30T03:51:48+00:00

Keep as many records around as needed in parsing.

2014-01-21T23:48:05+00:00

BUG FIX: TimeTracker average computation

2014-01-18T03:16:42+00:00

The formula for computing the running average was incorrect.
This patch fixes that.  It also renames the putput fields
to make grouping similar values on the screen (by lexical order)
easier.

parse_exps.py: Make data analysis more accurate.

2014-01-18T01:18:08+00:00

This patch makes the data analysis of parse_exps.py.
The old code suffered from fragility---logs that were
out of order lead to many dropped records. This is
because it had a look-ahead buffer of 1.

This patch does:
1) Maintains buffers of unmatched records.
2) Preloads 200 records from every sched stream
   and processes them in sorted order.

This patch does increase runtime, but prior results
could be useless for many trace logs.

Support PGM-based sched trace analysis

2014-01-17T21:57:20+00:00

Modified sched.py field padding calculations to account for bitfields.

2013-05-02T17:42:17+00:00

Improved accuracy of sched_trace measurement parsing.

2013-04-23T18:01:35+00:00

* Measurements from tasks missing > 20% of their scheduling records are
ignored. This is configurable in config/config.py.
* Measurements which only have zero values are ignored.
* If either of these 2 situations are encountered print out a message the
first time using the common.log_once() method. See parse_exps.py for how
this is used with multiple threads.
* Measurements from a task's last job are ignored.
* Miss ratio is calculated only as a fraction of the number of jobs whose
matching release and completion records were found, not just release.

Switched sched_trace data to verbose ctypes structs.

2013-04-18T21:42:23+00:00

Fixed calculation of tardiness.

2013-04-17T20:07:35+00:00