Install otf2 with Homebrew

Project links

• Homepage

Commands and aliases

• otf2-config
• otf2-estimator
• otf2-marker
• otf2-print
• otf2-snapshots

Project history

OTF2 was created as the common successor to the original Open Trace Format and the Epilog trace format. The 2012 OTF2 paper described it as a redesign informed by those predecessor formats, with a new file encoding, close integration with Score-P, and explicit scalability improvements for large trace data.

The OTF2 manual identifies the library as a Score-P project component and describes the package as an interface for reading and writing trace data, distributed under a BSD license suitable for academic and commercial use. The Score-P documentation groups OTF2 with Cube4 and OPARI2 as part of the wider VI-HPS performance instrumentation and analysis stack.

Adoption history

VI-HPS documentation describes OTF2 as the standard trace format for Scalasca, Vampir, and TAU, while OLCF user documentation explains that Score-P generates OTF2 traces for analysis by those tools. That combination made OTF2 a package-manager-facing format library rather than just an internal file format.

The original OTF project page at TU Dresden points users from the retired Open Trace Format to OTF2 in the Score-P infrastructure, which marks the transition from first-generation OTF tooling to the Score-P-era trace stack.

How it is used

Developers and performance engineers use OTF2 indirectly through Score-P measurement runs and directly through its reader and writer APIs and command-line tools. A typical Score-P workflow instruments C, C++, or Fortran code, runs a measurement, enables tracing, and then analyzes the resulting OTF2 trace with tools such as Vampir.

Package users care about executables such as otf2-print, otf2-config, otf2-estimator, and otf2-snapshots because they make trace files inspectable and scriptable outside a full GUI performance-analysis session.

Why package nerds care

OTF2 is significant because it is a format package with ecosystem gravity: installing it supplies the shared library and utilities behind several HPC profiling and tracing workflows. It is also a good example of a scientific package whose importance comes from interoperability between instrumentation, storage format, and visualization tools.

Timeline

• 2012: The OTF2 paper introduced Open Trace Format 2 as a redesigned successor to OTF version 1 and Epilog, focused on scalable trace encoding and Score-P integration.
• 2014: TU Dresden listed OTF 1.12.5 as the last stable OTF release and directed users to OTF2 through Score-P.
• 2021: The OTF2 2.3 software record on Zenodo described HIP accelerator and Kokkos paradigm additions, showing the format continuing to track HPC programming models.

Related projects

• Score-P produces OTF2 traces; Scalasca, Vampir, and TAU consume or work with OTF2 trace data; Cube4 occupies the corresponding profile-data role in the same performance-analysis family.

Sources

• https://docs.olcf.ornl.gov/software/profiling/Scorep.html
• https://perftools.pages.jsc.fz-juelich.de/cicd/otf2/tags/otf2-3.0/pdf/otf2.pdf
• https://tu-dresden.de/zih/forschung/projekte/otf/?set_language=en
• https://www.researchgate.net/publication/259229252_Open_Trace_Format_2_The_Next_Generation_of_Scalable_Trace_Formats_and_Support_Libraries
• https://www.vi-hps.org/projects/score-p/documentation/documentation.html
• https://zenodo.org/records/4682684