| W. Shu and M-Y Wu. Solving dynamic and irregular problems on SIMD architectures with runtime support. In 1993 International Conference on Parallel Processing, 1993. |
....Using this set as the interpreted operation set, concurrent interpretation becomes possible. The P kernel runtime system manager is a run time support system that provides a general environment for using interpretation to implement asynchronous applications on SIMD machines [ Shu and Wu, 1995, Shu and Wu, 1993 ] P kernel leaves it to the programmer to define a set of processes for the interpretation set, and then use these processes to express the application. Each process consists of a set of atomic computations that share one common data area. The atomic computations include general code segments as ....
W. Shu and M-Y Wu. Solving dynamic and irregular problems on SIMD architectures with runtime support. In 1993 International Conference on Parallel Processing, 1993.
....also looked into loop transformations that help optimize the node level programs when a BLOCK distribution is used [18] 2.1. 5 MIMD Emulators There has been growing interest in determining if SIMD architectures can successfully be used to handle problems that do not fit the data parallel model [22, 37, 60, 66]. One method is to emulate MIMD execution by writing a SIMD program to interpret a MIMD instruction set. Each PE contains a program to be interpreted and its data. The front end then executes an interpreter loop where the instruction decode step iterates over all possible instruction types ....
W. Shu and M. Wu. Solving dynamic and irregular problems on SIMD architectures with runtime support. submitted to ICPP'93, August 1993.
....serving as a general purpose machine, and to study the feasibility of providing a truly portable parallel programming environment between SIMD and MIMD machines. The first version of the P kernel was written in CM Fortran, running on a 4K processor TMC CM 2 in 1991. The performance was reported in [32]. Fortran is not the best language for system programs. We used the multiple dimension array with indirect addressing to implement queues. Hence, not only was the indirect addressing extremely slow, but accessing different addresses in CM 2 costed much more [9] In 1993, the P kernel was rewritten ....
W. Shu and M.Y. Wu. Solving dynamic and irregular problems on SIMD architectures with runtime support. In Int'l Conf. on Parallel Processing, pages II. 167--174, August 1993.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC