Legacy fortran software: applying syntactic metrics to global climate models
It is di cult to maintain legacy Fortran programs that use outdated programming constructs, especially when this maintenance requires a detailed understanding of the code (e.g., for parallelization). Initially, we want to gauge the prevalence of such constructs by applying straightforward syntactic...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Objeto de conferencia |
| Lenguaje: | Inglés |
| Publicado: |
2012
|
| Materias: | |
| Acceso en línea: | http://sedici.unlp.edu.ar/handle/10915/23718 |
| Aporte de: |
| Sumario: | It is di cult to maintain legacy Fortran programs that use outdated programming constructs, especially when this maintenance requires a detailed understanding of the code (e.g., for parallelization).
Initially, we want to gauge the prevalence of such constructs by applying straightforward syntactic metrics to some well-known global climate models. Detailed information regarding les, subroutines, and loops has been collected from each model by applying a lightweight source code static analysis based on ASTs (Abstract Syntax Tree) for a posterior analysis. Modernizing Fortran Legacy programs is still a challenge. Our objective has been to collect relevant information on these programs to help us approach parallelizing legacy scienti c programs in a shared memory environment (e.g. using multi-core processors). The data we collected indicate that old Fortran features are still being used on these models in these days. Furthermore, we propose some metrics to be used as a guide to determine how many changes a program needs in order to be modernized, optimized, and eventually, parallelized. |
|---|