Embeddings of Task Mappings to Multicore Systems

The problem of finding good mappings is central to designing and executing applications efficiently in embedded systems. In heterogeneous multicores, which are ubiquitous today, this yields an intractably large design space possible mappings. Most methods explore using heuristics, many implicitly use geometric notions paper we the geometry mapping explicitly, for embeddings that capture its structure. This allows us formulate new strategies by leveraging space, as well improving existing heuristics do so implicitly. We evaluate our approach on a novel heuristic based gradient descent, multiple meta-heuristics. For complex architectures, improved results established exploration meta-heuristics about order magnitude average.