This report complements and accompanies the M12 releases of the M A X lighthouse codes. The releases represent a consolidation of the results achieved in phase 2 of M A X regarding performance, portability, and reliability and also contain the more mature re- sults of the development work that is going on in phase 3 of the CoE. We complete this report by presenting and discussing the progress of the M A X codes for what concerns the more advanced and non-production-ready developments that have been going on in M A X codes during this first year. The most mature achievements for portability (T1.1) have been obtained for CUDA architectures and are, for the most part, already integrated into the production releases. Other significant results have been achieved for different types of heterogeneous archi- tectures, such as the support of the AMD accelerators and, more in general, consider- able work for supporting OpenMP offloading (e.g. Q UANTUM ESPRESSO, YAMBO, DeviceXlib to name a few) or using SYCL based platforms (e.g. B IG DFT) Moreover, parallel performance has been largely consolidated, eliminating known issues and inefficiencies. The code base at this point also represents a starting point for the more advanced and exascale-oriented parallelization strategies that we are experimenting with. The effort to expand and streamline the continuous integration and deployment of these advanced developments is crucial for monitoring the project and for achieving an effective integration within the EuroHPC environment. For this reason, we have worked with WP3 and WP4 to increase the code coverage with CI tests, and to experiment with the different deployment tools (CD) within the collaboration with CASTIEL-2 regarding CI/CD on EuroHPC machines. We also report about the development activities aimed at enabling the implemen- tation of exascale workflows and generally targeted to deploying the M A X lighthouse applications. These include the necessary interoperability traits planned in T1.4 and the introduction of new features and property calculators in the codes.

D1.2 First release of MAX software: report on performed and planned refactoring

Stefano Baroni;Ivan Carnimeo;Pietro Delugas;Fabrizio Ferrari Ruffino;Andrea Ferretti;Paolo Giannozzi;Davide Sangalli;Nicola Spallanzani;Daniele Varsano
2023

Abstract

This report complements and accompanies the M12 releases of the M A X lighthouse codes. The releases represent a consolidation of the results achieved in phase 2 of M A X regarding performance, portability, and reliability and also contain the more mature re- sults of the development work that is going on in phase 3 of the CoE. We complete this report by presenting and discussing the progress of the M A X codes for what concerns the more advanced and non-production-ready developments that have been going on in M A X codes during this first year. The most mature achievements for portability (T1.1) have been obtained for CUDA architectures and are, for the most part, already integrated into the production releases. Other significant results have been achieved for different types of heterogeneous archi- tectures, such as the support of the AMD accelerators and, more in general, consider- able work for supporting OpenMP offloading (e.g. Q UANTUM ESPRESSO, YAMBO, DeviceXlib to name a few) or using SYCL based platforms (e.g. B IG DFT) Moreover, parallel performance has been largely consolidated, eliminating known issues and inefficiencies. The code base at this point also represents a starting point for the more advanced and exascale-oriented parallelization strategies that we are experimenting with. The effort to expand and streamline the continuous integration and deployment of these advanced developments is crucial for monitoring the project and for achieving an effective integration within the EuroHPC environment. For this reason, we have worked with WP3 and WP4 to increase the code coverage with CI tests, and to experiment with the different deployment tools (CD) within the collaboration with CASTIEL-2 regarding CI/CD on EuroHPC machines. We also report about the development activities aimed at enabling the implemen- tation of exascale workflows and generally targeted to deploying the M A X lighthouse applications. These include the necessary interoperability traits planned in T1.4 and the introduction of new features and property calculators in the codes.
2023
Istituto Nanoscienze - NANO - Sede Secondaria Modena
Istituto di Struttura della Materia - ISM - Sede Secondaria Montelibretti
Istituto Officina dei Materiali - IOM -
Rapporto intermedio di progetto
MaX
File in questo prodotto:
File Dimensione Formato  
D1.2_First release of MAX software_report on performed and planned refactoring.pdf

accesso aperto

Licenza: Dominio pubblico
Dimensione 1.9 MB
Formato Adobe PDF
1.9 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14243/516021
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact