Scan Grammars: Parallel Attribute Evaluation Via Data-Parallelism
Thomas Reps
This paper concerns the problem of how to exploit parallelism during
the phases of compilation involving syntax-directed analysis and
translation. In particular, we address the problem of how to exploit
parallelism during the evaluation of the attributes of a derivation
tree of a non-circular attribute grammar. What distinguishes the
ideas presented in this paper from earlier work on parallel attribute
evaluation is the use of a data-parallel model: We define a
new variant of attribute grammars, called scan grammars, that
incorporates a data-parallel attribution construct.
(Click here to access the paper:
PostScript,
PDF.)
University of Wisconsin