arrow
Return

Streaming I/O for scientific workflow engine acceleration

delete2025-07-22
delete0
delete
OA
AI
S
Simone Perrotta
C
Ciro Giuseppe De Vita *
G
Gennaro Mellone
M
Marco Edoardo Santimaria
M
Massimo Torquati
J
Javier Garcia‐Blas
R
Raffaele Montella
DOI:10.1016/j.future.2025.107978delete
deleteOriginal
deleteShare
deleteSave
View PDF
Abstract

Abstract

En 中文
• Integration of memory-based streaming I/O in workflow engines. • Auto-generation of sync rules via DAGonStar dependency analysis. • Faster pipeline task execution via CAPIO system call interception. • Benchmark shows up to 33% execution time reduction with DAGonCAPIO. • Support for local batch and SLURM-based distributed runs.
Keywords:
High-performance computing
Workflow optimization
Scientific workflows
Streaming I/O
In-memory file systems
File-based pipeline acceleration
AI Summary

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.

Journal

F
Future Generation Computer Systems
IF:
0
Papers:
642
Citations:
0

Organization

U
University of Turin
Scholars:
3.7W
Papers: 2.8W
Citations: 3.2W
U
University of Pisa
Scholars:
3.1W
Papers: 2.4W
Citations: 2.4W