PLOTIOTOTALSBYPE(OctoberPablo EnvironmenPLOTIOTOTALSBYPE(October)
NAME
PlotIOtotalsByPE - Reads a trace file containg I/O detail
trace records and generates text files summarizing I/O
data which can be used as input to plotting tools.
SYNOPSIS
PlotIOtotalsByPE [-fileID id] tracefile
DESCRIPTION
PlotIOtotalsByPE reads an input SDDF file containing I/O
detail trace records and produces several output text
files which can be used as input to gnuplot(1) and other
standard plotting programs. The necessary detail trace
event records are produced by the I/O extension to the
Pablo trace library by default whenever the I/O extension
has been initialized and individual I/O calls have been
instrumented.
Each output file contains two values per line. The first
value corresponds to the processor number. The second
value corresponds to the count, duration, or bytes
involved in a particular class of I/O operation on the
processor (as given by the first value).
The names of the individual output files indicate the
meaning of the second value as follows:
o ioclassCountByPE.plot: contains operation count
information
o ioclassDurationByPE.plot: contains operation
duration information
o ioclassByteByPE.plot: contains operation bytes
inforrmation
The ioclass components of the output file names and their
corresponding I/O detail trace reacords are:
o openAndGopen: Open and Global Open
o close: Close
o read: Read
o seek: Seek
o write: Write
o aRead: Async Read
o aWrite: Asynch Write
o other: Flush, Misc I/O, IO Wait, IO Mode, and
Lsize
o allIO: All of the I/O detail record types
As PlotIOtotalsByPE is running, it periodically displays
the number of input trace packets (records) processed to
standard error.
If the input file does not include any detail I/O trace
records, the output files will be generated without error,
but will contain only zeros.
OPTIONS
-fileID id
Only I/O activity associated with the specified
file, as identified by id, is included in the out-
put files. When this option is used, the generated
output file names are of the form ioclassCount-
ByPE_id, ioclassDuratinByPE_id, and ioclassBytes-
ByPE_id where id is the selected processor.
The default behavior is to include I/O activity for
all files in the output.
BEFORE RUNNING THE PROGRAM
The application to be studied must be instrumented with
the I/O extension to the Pablo trace library and should
NOT call the function disableIOdetail(). The instrumented
application is run and one or more trace files are gener-
ated - the number of trace files depends on the number of
processors used to run the application.
If there are multiple trace files, they should be merged
with the MergePabloTraces(1) program to produce a single
trace file for the execution including information from
all processors.
If you intend to use the -fileID option to include only
accesses to individual files, the program SyncIOfileIDs(1)
should be run on the input trace file to synchronize IDs
for files that were opened more than once by the applica-
tion. The file generated by SyncIOfileIDs ending in
".syncFiles" can be used as input to PlotIOtotalsByPE.
The file ending in ".syncFiles.map" can be used to iden-
tify the file ID of the particular file whose I/O you are
interested in plotting.
If I/O detail trace records were not produced for all I/O
operations in the application, either because some calls
were not instrumented or because detail tracing was dis-
abled for part of the execution, the output file totals
will include only those operations for which detail
records were produced and should be interpreted with that
in mind.
SEE ALSO
IOtotalsByPE(1), MergePabloTraces(1), SyncIOfileIDs(1),
gnuplot(1)
Ruth A. Aydt, A User's Guide to Pablo I/O Instrumentation
Ruth A. Aydt, The Pablo Self-Defining Data Format
COPYRIGHT
Copyright 1996, The University of Illinois Board of
Trustees.
Ruth A. Aydt, University of Illinois