Graph Input-Aware Matrix Multiplication for Pruned Graph Neural Network Acceleration

Graphs-based neural networks are powerful tools for analyzing intricate non-euclidean data structures to solve complex real-world problems bounded by latency and throughput constraints. However, the computation structures operate on ultra-sparse and unstructured matrices resulting in load balancing...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	Proceedings - IEEE International Parallel and Distributed Processing Symposium S. 913 - 925
Hauptverfasser:	Khan, Hanan, Gurevin, Deniz, Khan, Omer
Format:	Tagungsbericht
Sprache:	Englisch
Veröffentlicht:	IEEE 03.06.2025
Schlagworte:	Computational modeling GPU Graph neural networks graph processing Graphics processing units Instruction sets Load modeling parallel algorithm Parallel processing pruning Runtime Sparse matrices Sparse matrix multiplication Throughput Vectors
ISSN:	1530-2075
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!