The goal of this paper is to present ANDES, a language for modelling q
uantitative behaviour of parallel programs. ANDES is employed in the A
LPES environment. From an ANDES quantitative model, a parallel synthet
ic program is generated. This synthetic program is executed on a real
distributed memory parallel computer. Some performance indices are obt
ained from this execution. Currently, mapping strategies are being eva
luated using this approach. The ALPES environment is a research domain
in the APACHE project.