Modular Construction and Optimization of the UZP Sparse Format for SpMV on CPUs

Sparse data structures are ubiquitous in modern computing, and numerous formats have been designed to represent them. These formats may exploit specific sparsity patterns, aiming to achieve higher performance for key numerical computations than more general-purpose formats such as CSR and COO. In th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings of ACM on programming languages Jg. 9; H. PLDI; S. 2106 - 2130
Hauptverfasser: Rodríguez-Iglesias, Alonso, Tongli, Santoshkumar T., Tucker, Emily, Pouchet, Louis-Noël, Rodríguez, Gabriel, Touriño, Juan
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York, NY, USA ACM 10.06.2025
Schlagworte:
General and reference
Theory of computation
Vector / streaming algorithms
SIMD vectorization
code generation
polyhedral compilation
sparse format
sparse linear algebra
ISSN:2475-1421, 2475-1421
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Sparse data structures are ubiquitous in modern computing, and numerous formats have been designed to represent them. These formats may exploit specific sparsity patterns, aiming to achieve higher performance for key numerical computations than more general-purpose formats such as CSR and COO. In this work we present UZP, a new sparse format based on polyhedral sets of integer points. UZP is a flexible format that subsumes CSR, COO, DIA, BCSR, etc., by raising them to a common mathematical abstraction: a union of integer polyhedra, each intersected with an affine lattice. We present a modular approach to building and optimizing UZP: it captures equivalence classes for the sparse structure, enabling the tuning of the representation for target-specific and application-specific performance considerations. UZP is built from any input sparse structure using integer coordinates, and is interoperable with existing software using CSR and COO data layout. We provide detailed performance evaluation of UZP on 200+ matrices from SuiteSparse, demonstrating how simple and mostly unoptimized generic executors for UZP can already achieve solid performance by exploiting Z-polyhedra structures.
ISSN:2475-1421
2475-1421
DOI:10.1145/3729335