The Darshan source tree is divided into two parts:
- darshan-runtime: to be installed on systems where you intend to instrument MPI applications.
- darshan-util: to be installed on systems where you intend to analyze log files produced by darshan-runtime.
This document gives details on the design of the new modularized version of Darshan (3.x) and how new instrumentation modules may be developed within Darshan.
Documentation on PyDarshan, a Python package providing interfaces to Darshan log files can be found here.
Site-specific documentation for facilities that deploy Darshan in production:
- Argonne Leadership Computing Facility (ALCF): Theta, Cooley
- National Energy Research Scientific Computing Center (NERSC)
- National Center for Supercomputing Applications (NCSA)
- Oak Ridge Leadership Computing Facility (OLCF): darshan-runtime, darshan-util
- King Abdullah University of Science and Technology (KAUST)
- European Centre for Medium-Range Weather Forecasts (ECMWF)
- Ohio Supercomputer Center (OSC)
- Jülich Supercomputing Centre (JSC)