[Trace analysis] Timestamps are not monotonically increasing
- Michael Kobert
Owned by Michael Kobert
Issue
In my trace analysis I receive the following error:
2024-04-18 10:06:13.284 8152 ManagedWorker-7*PE.Run() ERROR:
Traceback (most recent call last):
File "tts\traceAnalysis\common\steps\episode\EpisodeExec.py", line 631, in Execute
File "tts\traceAnalysis\common\steps\episode\EpisodeExec.py", line 1024, in _ExecuteSession
File "src-gen\ttTraceCheckLib\session.py", line 62, in processWithExecState
File "src-gen\ttTraceCheckLib\session.py", line 653, in performProcessing
File "src-gen\ttTraceCheckLib\session.py", line 462, in _ReadNumpySources
File "tts\traceRecording\parser\NumpyRecordingEventSource.py", line 351, in ReadNumpyArrays
File "tts\traceRecording\parser\NumpyRecordingEventSource.py", line 199, in _CheckMonotony
tt._Exceptions_.TTError: Error reading <TraceSignalName>: TThe provided timestamps are not monotonically increasing. Violation: time <timestamp1>-><timestamp2>. File: <pathToFile>How can I fix it?
Solution
The behavior is part of tool adapter Vector: XL API and the issue is highly related to the settings in your Vector Hardware Manager where a time synchronization other than Legacy is set.
If you take a look at your recorded trace file you will probably notice, that the timestamps typically start at approximately 1.844*1010 seconds and then overflow.
If nothing speaks against it, then change the time synchronization to Legacy.