Parallel triangular matrix system solving on CPU-GPU system
Ryma Mahfoudhi, Sami Achour, Zaher Mahjoub · 2016
GPU-accelerated computing consists in using a graphics processing unit (GPU) together with a CPU in order to enhance the performance of scientific and engineering applications. With the increasing spread of GPUs as hardware accelerators for scientific applications, several optimized linear algebra libraries have emerged to make use of this additional computing power. In this paper we present an implementation of a recursive algorithm for triangular matrix system solving targeting a hybrid multicore + GPU system.