kahip ☆
« Back to VersTracker
Description:
Karlsruhe High Quality Partitioning
Type: Formula  |  Latest Version: 3.22@0  |  Tracked Since: Nov 3, 2025
Links: Homepage  |  formulae.brew.sh
Category: Developer tools
Tags: graph-partitioning hpc parallel-computing algorithm clustering
Install: brew install kahip
About:
KaHIP (Karlsruhe High Quality Partitioning) is a state-of-the-art library for graph partitioning. It provides parallel multi-level algorithms to solve the NP-hard graph partitioning problem, optimizing edge cuts to minimize communication overhead. The tool is designed to deliver high-quality partitions that significantly improve performance in parallel simulations and data processing.
Key Features:
  • Parallel multi-level partitioning algorithms
  • Supports various quality presets (fast, strong, etc.)
  • Handles weighted graphs and different partition counts
  • Includes tools for graph generation and quality evaluation
Use Cases:
  • Optimizing communication in parallel High-Performance Computing (HPC) applications
  • Load balancing for sparse matrix operations
  • Preprocessing graphs for efficient storage or traversal
Alternatives:
  • METIS – METIS is the industry standard but proprietary for commercial use; KaHIP is open-source and often provides better partition quality at the cost of speed.
  • scotch – SCOTCH focuses on minimizing communication volume and is widely used in scientific computing; KaHIP is often compared for its pure edge-cut minimization.
Version History
Detected Version Rev Change Commit
Nov 3, 2025 8:09pm 0 VERSION_BUMP de069b5d
Sep 12, 2025 3:06pm 0 VERSION_BUMP 2a217a24
Nov 27, 2024 7:58am 0 VERSION_BUMP 0bd615c7
Sep 14, 2024 10:32am 0 VERSION_BUMP a72178b2
« Back to VersTracker  |  Browse Packages  |  About