Performance and programmability comparison of the thick control flow architecture and current multicore processors

Commercial multicore central processing units (CPU) integrate a number of processor cores on a single chip to support parallel execution of computational tasks. Multicore CPUs can possibly improve performance over single cores for independent parallel tasks nearly linearly as long as sufficient bandwidth is available. Ideal speedup is, however, difficult to achieve when dense intercommunication between the cores or complex memory access patterns is required. This is caused by expensive synchronization and thread switching, and insufficient latency toleration. These facts guide programmers away from straight-forward parallel processing patterns toward complex and error-prone programming techniques. To address these problems, we have introduced the Thick control flow (TCF) Processor Architecture. TCF is an abstraction of parallel computation that combines self-similar threads into computational entities. In this paper, we compare the performance and programmability of an entry-level TCF processor and two Intel Skylake multicore CPUs on commonly used parallel kernels to find out how well our architecture solves these issues that greatly reduce the productivity of parallel software development. Code examples are given and programming experiences recorded.

Issues of parallel software development for the domain decomposition methods

Рассматриваются различные аспекты разработки параллельного программного обеспечения для метода декомпозиции области: использование технологии MPI-программирования для кластерных систем, точки выбора при проектировании параллельных программ методов декомпозиции области, необходимость реализации действия матрицы без явного ее представления, работа с множествами индексов при программной реализации операторов ограничения и продолжения, а также при обмене данными между подобластями. На ряде численных экспериментов для модельной задачи исследуются вопросы наилучшего выбора конфигурации запуска исполняемой программы на кластере для минимизации времени расчета и предлагается стратегия проведения серии вычислительных экспериментов. Various aspects of parallel software development for the domain decomposition methods are considered: the application of MPI programming technology for cluster systems, the choice points in the design of parallel programs for the domain decomposition methods, the need to implement a matrix action without its explicit representation, the work with index sets in the software implementation of restriction and continuation operators as well as in the data exchange between subdomains. On a series of numerical experiments for a model problem, the questions of the best choice of the configuration of launching an executable program on a cluster are studied to minimize the computation time and a strategy for performing such experiments is proposed.

HIGH-LEVEL PARALLEL SOFTWARE DEVELOPMENT WITH PYTHON AND BSP

One of the main obstacles to a more widespread use of parallel computing in computational science is the difficulty of implementing, testing, and maintaining parallel programs. The combination of a simple parallel computation model, BSP, and a high-level programming language, Python, simplifies these tasks significantly. It allows the rapid development facilities of Python to be applied to parallel programs, providing interactive development as well as interactive debugging of parallel programs.