cubelib
« Back to VersTracker
Description:
Performance report explorer for Scalasca and Score-P
Type: Formula  |  Latest Version: 4.9@0  |  Tracked Since: Dec 17, 2025
Links: Homepage  |  formulae.brew.sh
Category: Developer tools
Tags: hpc performance profiling analysis scientific-computing
Install: brew install cubelib
About:
CubeLib is the core library and command-line interface for the Cube performance report explorer. It provides a robust API and tools for parsing, processing, and navigating large-scale performance data generated by Scalasca and Score-P. Its main value is enabling developers to programmatically analyze and manage complex performance metrics from HPC applications.
Key Features:
  • Command-line interface for report inspection and manipulation
  • C++ and Python APIs for programmatic data access
  • Support for standard Cube performance report formats
  • Efficient handling of large, multidimensional performance datasets
Use Cases:
  • Automating the extraction of performance metrics from HPC job runs
  • Integrating performance data analysis into custom scripting pipelines
  • Programmatically generating custom performance summaries and reports
Alternatives:
  • paraver – Paraver offers detailed timeline visualization, whereas CubeLib focuses on hierarchical data exploration and programmatic access.
  • hatchet – Hatchet provides similar programmatic analysis for graph-based profiles, often used with PAPI or Caliper data.
License: BSD-3-Clause
Bottles available for: arm64_tahoe, arm64_sequoia, arm64_sonoma, arm64_ventura, sonoma, ventura, arm64_linux, x86_64_linux
Version History
Detected Version Rev Change Commit
Oct 15, 2025 5:02pm 0 VERSION_BUMP 32d65b1f
Nov 17, 2024 8:37pm 0 VERSION_BUMP 0ca09ab1