UMPa: A multi-objective, multi-level partitioner for communication minimization

MaRDI QIDQ5298933zbMATH OpenFDO

Authors Mehmet Deveci, Kamer Kaya, Bora Uçar, Ümit V. Çatalyürek

Publication date 25 June 2013

zbMATH Keywords

processing unitsdirected hypergraph modeltask distributions

Mathematics Subject Classification ID

Programming involving graphs or networks (90C35) Directed graphs (digraphs), tournaments (05C20) Graph theory (including graph drawing) in computer science (68R10) Hypergraphs (05C65) Edge subsets with special properties (factorization, matching, partitioning, covering and packing, etc.) (05C70) Network design and communication in computer systems (68M10)

Recommendations

• Optimal partitioning and granularity of uniform task graphs
• Encapsulating Multiple Communication-Cost Metrics in Partitioning Sparse Rectangular Matrices for Parallel Matrix-Vector Multiplies
• A computation + communication load balanced loop partitioning method for distributed memory systems
• Minimum nonuniform graph partitioning with unrelated weights
• scientific article; zbMATH DE number 3885290

Cited in

(7)

• Simultaneous Computational and Data Load Balancing in Distributed-Memory Setting
• Complex network partitioning using label propagation
• Improving unstructured mesh partitions for multiple criteria using mesh adjacencies
• Scalable high-quality hypergraph partitioning
• UMPa
• Multi-objective k-way parallel hypergraph partitioning with proximal gradient algorithm
• Optimal partitioning and granularity of uniform task graphs

Describes a project that uses

Uses Software

• Parkway
• PaToH

This page was built for publication: UMPa: A multi-objective, multi-level partitioner for communication minimization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5298933)