summaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)AuthorLines
2021-01-13pelt: Allow overrides via environment variablesDaniel Friesel-4/+23
2021-01-13EnergyTrace loader: remove bogus tail sync data as wellDaniel Friesel-7/+14
2021-01-12add a TODO entrDaniel Friesel-1/+4
2021-01-12add z-score based outlier detection and optional removalDaniel Friesel-1/+41
Note that outliers are removed from both training and evaluation data
2021-01-12plotter: print notice when exporting plotsDaniel Friesel-0/+3
2021-01-11evaluate E = P*t model. Seems slightly better in some casesDaniel Friesel-0/+19
TODO: Fix send duration model generation for cc1200, should improve eval
2021-01-08work around missing W_mean_delta_prev in first energytrace+barcode transitionDaniel Friesel-0/+5
2021-01-08model transition power, not energyDaniel Friesel-21/+54
in contrast to duration/energy, duration/power are independent
2021-01-08gracefully handle zbar unavailabilityDaniel Friesel-1/+8
2020-12-16loader: fix off-by-one in energytrace++ state changesDaniel Friesel-1/+2
2020-12-15add proof of concept for EnergyTrace++ based drift compensationDaniel Friesel-4/+61
2020-12-15support benchmarks with EnergyTrace++Daniel Friesel-3/+9
2020-12-15dataprocessor: export drift compensation dataDaniel Friesel-0/+14
2020-12-14energytrace: add pelt-based drift compensation experiment.Daniel Friesel-6/+140
Enable with DFATOOL_COMPENSATE_DRIFT=1 so far it's pretty unreliable.
2020-12-14runner: energytrace: wait for benchmark end synchronization pulsesDaniel Friesel-0/+2
2020-12-11generate-dfa-benchmark: include gpio.hDaniel Friesel-0/+8
2020-12-03use common energytrace loader codeDaniel Friesel-102/+73
this also fixes a nasty off-by-one in the barcode loader (now _load_energytrace): wrong: interval_start_timestamp = data[:-1, 0] * 1e-6 correct: interval_start_timestamp = data[1:, 0] * 1e-6
2020-12-03loader: energytrace: add support for legacy file formatDaniel Friesel-0/+3
2020-12-01loader: gracefully handle exceptions in energytrace sub-loaderDaniel Friesel-3/+7
2020-12-01runner, harness: add logging support and debug outputDaniel Friesel-2/+14
2020-12-01loader: fix typoDaniel Friesel-1/+1
2020-11-30pelt: allow disabling multiprocessingDaniel Friesel-2/+6
2020-11-27clean up energytrace drift compensation and raw model generation codeDaniel Friesel-147/+133
2020-11-26some commentsDaniel Friesel-7/+3
2020-11-19DataProcessor: Handle broken LA measurementsDaniel Friesel-0/+9
2020-11-16do not parse energytrace logs with large sync offsetDaniel Friesel-3/+7
2020-11-09DataProcessor: Warn for positive and negative offset errorsDaniel Friesel-1/+1
2020-11-04wipDaniel Friesel-0/+44
2020-11-04harness: show message when encountering unparseable lineDaniel Friesel-0/+8
2020-11-03analyze number of substates per stateDaniel Friesel-92/+181
2020-11-02Do not pass the entirety of by_nane/by_param to subprocessesDaniel Friesel-105/+76
This should speed up analysis quite a bit and also reduce memory usage significantly
2020-11-02PELT: Do not set jump when len(signal) < num_samplesDaniel Friesel-1/+2
2020-11-02EnergyTrace: Correctly set "power_traces"Daniel Friesel-0/+6
2020-11-02Simplify PELT usage. remove kneedle, refactor codeDaniel Friesel-298/+100
2020-11-02Merge branch 'master' into merge-prep/janisDaniel Friesel-97/+1420
2020-11-02harness: redo benchmark when running out of syncDaniel Friesel-4/+4
2020-10-30Add DFATOOL_EXPORT_LASYNC variable for ET+LA / ET+Timer sync evalDaniel Friesel-1/+36
2020-10-29Set DFATOOL_PLOT_LASYNC to plot ET+LA/ET+Timer sync dataDaniel Friesel-2/+2
2020-10-29EnergyTraceWithTimer: Gracefully handle errors in a single measurement runDaniel Friesel-3/+10
2020-10-26blacken code; do not run lint check on external Python librariesDaniel Friesel-14/+29
2020-10-23assorted minor improvementsmerge-prep/lennartDaniel Friesel-4/+5
2020-10-23sync=timer: fix end of measurement offset errorDaniel Friesel-11/+11
2020-10-23ET+Timer: handle start offset between sync pulse and first traceDaniel Friesel-7/+17
looking pretty good now!
2020-10-23minor refactoringDaniel Friesel-3/+3
2020-10-22EnergyTraceWithTimer: turns out multipass was the issue.Daniel Friesel-1/+3
2020-10-22Improve sync=la timing restoration. There's still something fishy though...Daniel Friesel-53/+42
2020-10-22unfuck LA<->ET drift calculationDaniel Friesel-2/+8
2020-10-21loader: Support new "files" format in MIMOSA filesDaniel Friesel-0/+6
2020-10-21--plot-traces: use the correct time base for each backendDaniel Friesel-17/+36
2020-10-20DataProcessor: commentsDaniel Friesel-10/+13
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-30 09:31:13 +0000