PetscLogView#

Prints a summary of the logging.

Synopsis#

#include "petscsys.h"   
PetscErrorCode PetscLogView(PetscViewer viewer)

Collective over MPI_Comm

Input Parameter#

viewer - an ASCII viewer

Options Database Keys#

-log_view [:filename] - Prints summary of log information
-log_view :filename.py:ascii_info_detail - Saves logging information from each process as a Python file
-log_view :filename.xml:ascii_xml - Saves a summary of the logging information in a nested format (see below for how to view it)
-log_view :filename.txt:ascii_flamegraph - Saves logging information in a format suitable for visualising as a Flame Graph (see below for how to view it)
-log_view_memory - Also display memory usage in each event
-log_view_gpu_time - Also display time in each event for GPU kernels (Note this may slow the computation)
-log_all - Saves a file Log.rank for each MPI rank with details of each step of the computation
-log_trace [filename] - Displays a trace of what each process is doing

Notes#

It is possible to control the logging programmatically but we recommend using the options database approach whenever possible By default the summary is printed to stdout.

Before calling this routine you must have called either PetscLogDefaultBegin() or PetscLogNestedBegin()

If PETSc is configured with –with-logging=0 then this functionality is not available

To view the nested XML format filename.xml first copy ${PETSC_DIR}/share/petsc/xml/performance_xml2html.xsl to the current directory then open filename.xml with your browser. Specific notes for certain browsers

Firefox and Internet explorer - simply open the file

Google Chrome - you must start up Chrome with the option --allow-file-access-from-files

Safari - see https://ccm.net/faq/36342-safari-how-to-enable-local-file-access

or one can use the package http://xmlsoft.org/XSLT/xsltproc2.html to translate the xml file to html and then open it with your browser. Alternatively, use the script ${PETSC_DIR}/lib/petsc/bin/petsc-performance-view to automatically open a new browser window and render the XML log file contents.

The nested XML format was kindly donated by Koos Huijssen and Christiaan M. Klaij MARITIME RESEARCH INSTITUTE NETHERLANDS

The Flame Graph output can be visualised using either the original Flame Graph script (https://github.com/brendangregg/FlameGraph) or using speedscope (https://www.speedscope.app). Old XML profiles may be converted into this format using the script ${PETSC_DIR}/lib/petsc/bin/xml2flamegraph.py.