soctrace-inria / framesoc

Framesoc is a generic trace management and analysis infrastructure.
12 stars 0 forks source link

PJdump export makes framesoc crash #190

Closed dbeniamine closed 8 years ago

dbeniamine commented 8 years ago

Hi,

I'm trying to export this trace (Import each file with Moca importer, with both trim lone producer and keep only accesses in data structure checked) to pjdump, but it makes framesoc crash after a few seconds.

I have the following log in the console:

Settings values:

Cache activated: true
Cache directory: /home/david/Work/soctrace/runtime-EclipseApplication/ocelotlCache
Cache size: -1
Snapshot directory: /home/david/Work/soctrace/runtime-EclipseApplication/ocelotlSnapshot
Cache Policy: CACHEPOLICY_FAST
Cache Time slices: 900
4 Metrics detected:
    Event Occurences [ALL] [PARTS, PROPORTIONS, MODE, EVENTMODE]
    State Total Duration [ALL] [PARTS, PROPORTIONS, MODE, STATEMODE]
    State Average Duration [ALL] [PARTS, PROPORTIONS, MODE, STATEMODE]
    Variable Average Value [ALL] [PARTS, PROPORTIONS, VARIABLEMODE]
2 Data aggregation operators detected:
    Temporal Aggregation [PARTS, PROPORTIONS, STATEMODE, EVENTMODE, VARIABLEMODE][OLP_3D]
    Spatiotemporal Aggregation [PARTS, STATEMODE, EVENTMODE, VARIABLEMODE][DLP_2D_Leaf]
8 Visualizations detected:
    Temporal Partition [PARTS] [OLP_3D]
    Temporal Stacked Bar Chart [PROPORTIONS] [OLP_3D]
    Spatiotemporal Partition [PARTS] [DLP_2D, DLP_2D_Leaf]
    Spatiotemporal State Mode [STATEMODE] [DLP_2D, DLP_2D_Leaf]
    Spatiotemporal Event Mode [EVENTMODE] [DLP_2D, DLP_2D_Leaf]
    Spatiotemporal Variable Mode [VARIABLEMODE] [DLP_2D, DLP_2D_Leaf]
    Temporal Mode Event [EVENTMODE] [OLP_3D]
    Temporal Mode State [STATEMODE] [OLP_3D]
5 Statistics aggregation operators detected:
    Quantity of Events [EVENT_SUM] [OLP_3D, DLP_2D, DLP_2D_Leaf]
    State Total Duration [STATE_SUM] [OLP_3D, DLP_2D]
    State Total Duration (only Leaves) - Temporal [STATE_SUM] [OLP_3D]
    State Total Duration (only Leaves) - Spatiotemporal) [STATE_SUM] [DLP_2D, DLP_2D_Leaf]
    State Total Duration (only Active Producers) [STATE_SUM] [OLP_3D, DLP_2D]
Found Moca_matrix_bloc_replacemalloc_20160224_183355_CET_Virtual in /home/david/Work/soctrace/runtime-EclipseApplication/ocelotlCache/Moca_matrix_bloc_replacemalloc_20160224_183355_CET_Virtual_8/Moca_matrix_bloc_replacemalloc_20160224_183355_CET_Virtual_8_Variable Average Value_24-02-2016 183840 CET.otlcache, Variable Average Value, Spatiotemporal Variable Mode, 1, 170297
Size of the current cache is: 193468803 bytes (193 MB).
[DICHOTOMY CACHE] Size of the current cache is: 0 bytes (0 MB).
Size of the current cache is: 193468803 bytes (193 MB).
[DICHOTOMY CACHE] Size of the current cache is: 0 bytes (0 MB).
Size of the current cache is: 193468803 bytes (193 MB).
[DICHOTOMY CACHE] Size of the current cache is: 0 bytes (0 MB).
Size of the current cache is: 193468803 bytes (193 MB).
[DICHOTOMY CACHE] Size of the current cache is: 0 bytes (0 MB).
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007f292af2d58a, pid=3508, tid=139814800213760
#
# JRE version: Java(TM) SE Runtime Environment (7.0_80-b15) (build 1.7.0_80-b15)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (24.80-b11 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [sqlite-3.7.2-libsqlitejdbc.so+0xf58a]  sqlite3_os_end+0x661
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# /home/david/hs_err_pid3508.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.

And there is the report file create by eclipse: hs_err_pid3508.log.txt.

David

ycorre commented 8 years ago

I did not managed to import the moca trace you provided: It fails with the following error "For input string: "@Virt"" and the following log:

java.lang.NumberFormatException: For input string: "@Virt"
    at java.lang.NumberFormatException.forInputString(NumberFormatException.java:65)
    at java.lang.Integer.parseInt(Integer.java:580)
    at java.lang.Integer.valueOf(Integer.java:766)
    at fr.inria.soctrace.tools.importer.moca.core.MocaParser.parseEventProd(MocaParser.java:606)
    at fr.inria.soctrace.tools.importer.moca.core.MocaParser.parseTrace(MocaParser.java:241)
    at fr.inria.soctrace.tools.importer.moca.MocaImporter$MocaImporterPluginJobBody.run(MocaImporter.java:118)
    at fr.inria.soctrace.framesoc.core.tools.management.PluginToolJob.run(PluginToolJob.java:76)
    at org.eclipse.core.internal.jobs.Worker.run(Worker.java:54)
dbeniamine commented 8 years ago

Sorry I put the wrong file in the archive, there is the right trace.zip.

Still the error is weird, it looks like the parser tried actually parse the header line while it should be dismissed, can you check that you have the last version of the importer ? commit: 0cb8b3e659a91d3429ad8178a403e19ac4116762

ycorre commented 8 years ago

I have pushed a new version of framesoc (and of framesoc.importer, although it was not at fault here) that should fix the issue.

(And the error I had with the moca importer was beacause the file containing the events had the wrong name and was thus processed as a file containing event producers).

dbeniamine commented 8 years ago

Hi,

Thanks for your answers, it seems to work here :)

David