D4.1 Working Python API to schedule MLC routines

R. Amela, R. Badia, R. Rossi

Author information
S. Zaghia,d, X. Martineza,c, R. Rossia,b
a CIMNE – Centre Internacional de Metodes Numerics en Enginyeria, Universitat Politécnica de Catalunya (UPC-BarcelonaTech), Barcelona 08034, Spain / Department of Civil and Environmental Engineering, Universitat Politécnica de Catalunya (UPC-BarcelonaTech), Barcelona 08034, Spain

, S. Böhm, J. Beranek, C. Soriano

Author information
Contact: csoriano@cimne.upc.edu
International Centre for Numerical Methods in Engineering

Document data

D4.1 Working Python API to schedule MLC routines

R. Amela, R. Badia, R. Rossi

, S. Böhm, J. Beranek, C. Soriano

Author information
Contact: csoriano@cimne.upc.edu
International Centre for Numerical Methods in Engineering

Abstract

This deliverable focuses on the denition of a common API for the PyCOMPSs programming

model and HyperLoom scheduler provided respectively by BSC and IT4I. The

objective of the work is to hide the details of the actual task scheduling technology, so that

the Multi Level Monte Carlo Python engine is agnostic of the backend being employed.

It includes the description of:

Common API for calls

Examples of usage

Basic Documentation

The document also contains an initial description on how MPI-distributed data shall

be treated from the scheduling point of view.

How to cite

R. Amela, R. Badia, R. Rossi, S. Böhm, J. Beranek and C. Soriano, D4.1 Working Python API to schedule MLC routines, Open Access Repository of the ExaQUte project: Deliverables (2021). 10 URL https://www.scipedia.com/public/table_Soriano_2018b

Abstract

Full document

Document information

Document Score

Share this document

claim authorship