IN-MEMORY PARALLEL PROCESSING FOR MATRIX TRANSPOSE.

Efficient implementation of parallel algorithms for matrix transposition requires careful consideration of data distribution, synchronization, communication overheads, and load balancing to achieve optimal performance in-memory parallel processing. We propose an algorithm and its implementation for...

Descripción completa

Detalles Bibliográficos
Publicado en:InSight: Rivier Academic Journal Vol. 18; no. 1; pp. 1 - 15
Autores principales: Malita, Mihaela, Ştefan, Gheorghe M.
Formato: Artículo
Publicado: Rivier College (InSight: Rivier Academic Journal) Fall2023
Materias:
Acceso en línea:Ver este registro en EBSCOhost
Descripción
Sumario:Efficient implementation of parallel algorithms for matrix transposition requires careful consideration of data distribution, synchronization, communication overheads, and load balancing to achieve optimal performance in-memory parallel processing. We propose an algorithm and its implementation for execution on a p-cell parallel in-memory cellular processor (CP). Two versions of CP are considered, one with size in O(p) and another with size in O(p×log p).