No abstract

No abstract

No abstract

This paper addresses the problem that multiple DSP system does not support OpenCL programming. With the compiler, runtime, and the kernel scheduler proposed, an OpenCL application becomes portable not only between multiple CPU and GPU, but also between embedded multiple DSP systems. Firstly, the LLVM compiler was imported for source-to-source translation in which the translated source was supported by CCS. Secondly, two-level schedulers were proposed to support efficient OpenCL kernel execution. The DSP/BIOS is used to schedule system level tasks such as interrupts and drivers; however, the synchronization mechanism resulted in heavy overhead during task switching. So we designed an efficient second level scheduler especially for OpenCL kernel work-item scheduling. The context switch process utilizes the 8 functional units and cross path links which was superior to DSP/BIOS in the aspect of task switching. Finally, dynamic loading and software managed CACHE were redesigned for OpenCL running on multiple DSP system. We evaluated the performance using some common OpenCL kernels from NVIDIA, AMD, NAS, and Parboil benchmarks. Experimental results show that the DSP OpenCL can efficiently exploit the computing resource of multiple cores.

A fast training support vector machine using parallel sequential minimal optimization is presented in this paper. Up to now, sequential minimal optimization (SMO) is one of the major algorithms for training SVM, but it still requires a large amount of computation time for the large sample problems. Unlike the traditional SMO, the parallel SMO partitions the entire training data set into small subsets first and then runs multiple CPU processors to seal with each of the partitioned data set. Experiments show that the new algorithm has great advantage in terms of speediness when applied to problems with large training sets and high dimensional spaces without reducing generalization performance of SVM.