A Parallel Categorical Abstract Machine for Bulk Synchronous Parallel ML

We have designed a functional data-parallel language called BSML for programming bulksynchronous parallel (BSP) algorithms in so-called direct mode. In a direct-mode BSP algorithm, the physical structure of processes is made explicit. The execution time can then be estimated and dead-locks and indeterminism are avoided. The BSMLlib library has been implemented for the Objective Caml language. But there is currently no full implementation of such a language and an abstract machine is needed to validate such an implementation. Our approach is based on a natural semantics and a bytecode compilation to a parallel abstract machine performing exchange of data and synchronous requests derived from the abstract machine of the Caml language.