Search

Modern graphics accelerators (GPUs) can significantly speed up the execution of numerical problems. However, porting programs to graphics accelerators is not an easy task, sometimes requiring their almost complete rewriting. CUDA graphics accelerators, thanks to technology developed by NVIDIA, allow one to have a single source code for both conventional processors (CPUs) and CUDA. However, parallelization on shared memory is still done differently and should be specified explicitly. The use of a functional programming library developed by the authors makes it possible to hide the use of one or another parallelization mechanism on shared memory within the library and make the user's source code completely independent of the computing device used (CPU or CUDA). This article shows how this can be done. [ABSTRACT FROM AUTHOR]

Modern functional programming languages, such as Haskell, Scala, ML, and F#, have properties that make it possible to implement logically complicated algorithms relatively easily. Among such properties is the composition of functions, currying, metafunctions (functions over functions), and some others. This makes it possible to obtain complex functions by combining simpler functions. An example of a complicated algorithm is a parser. In Haskell, there is a library called Parsec, which is a set of elementary parsers; by combining these parsers, one can create more complicated parsers. This fact makes the library relatively simple, yet powerful. The majority of these tools are not explicitly included in C++. However, C++ is sufficiently powerful for implementing many properties of functional programming languages. In this paper, an attempt is made to develop such a library. [ABSTRACT FROM AUTHOR]