Mapping Computations in Heterogeneous Multicore Systems with Statistical Regression on Program Inputs

A hardware configuration is a set of processors and their frequency levels in multicore heterogeneous system. This article presents compiler-based technique to match functions with configurations. Such consists using multivariate linear regression associate function arguments particular By showing that this classification space tends be convex practice, demonstrates not only an efficient tool map computations hardware, but also effective one. To demonstrate the viability as way perform adaptive compilation for architectures, we have implemented our ideas onto Soot Java bytecode analyzer. Code produce can predict best large class Scala benchmarks running on Odroid XU4 big.LITTLE board; hence, outperforming prior techniques such ARM’s GTS CHOAMP, recently released static program scheduler.