benchmark.logbase Module

This module has utilities to benchmark paraview.

You can set up arbitrary pipelines and this module helps you obtain, interpret and report the information recorded by ParaView’s logs. Do that like so:

1. Optionally, call maximize logs first

2. Setup and run your visualization pipeline (via GUI or script as you prefer)

3. Call print_logs() to print out the logs in raw format

class paraview.benchmark.logbase.OneLog(runmode='batch', servertype='unified', component=0, rank=0)

Bases: object

print_log(showlines=False)

toString(showlines=False)

paraview.benchmark.logbase.component_to_string(comp)

paraview.benchmark.logbase.dump_logs(filename)

This saves off the logs we’ve gathered. Ot allows you to run a benchmark somewhere, save off all of the details in raw format, then load them somewhere else. You can then do a detailed analysis and you always have the raw data to go back to.

paraview.benchmark.logbase.get_logs()

This is for bringing in logs at run time to parse while running.

paraview.benchmark.logbase.get_memuse()

paraview.benchmark.logbase.import_logs(filename)

This is for bringing in a saved log files and parse it after the fact. TODO: add an option to load in raw paraview logs in text format

paraview.benchmark.logbase.maximize_logs()

Convenience method to ask paraview to produce logs with lots of space and highest resolution.

paraview.benchmark.logbase.print_logs()

Print logs on the root node by gathering logs across all the nodes regardless if the process was started in symmetric mode or not.

paraview.benchmark.logbase.test_module()

Simply exercises a few components of the module.