Data-parallel model is attractive in a point that data-parallelism is easily expressed in loops and the efficiency is likely to be the highest if processors perform computations on data elements stored in local memory. However, communication is necessary in many cases and single-threaded data-parallel programs have a limitation in efficient communication. In this paper, we introduce multithreading concept to overcome the limitation. We present a new representation, PCFG(Parallel Control Flow Graph) and show how to construct it. We also show multithreaded codes can be easily transformed from single-threaded data-parallel codes using PCFGs. Keywords: data-parallel, multithreading, control dependence, data dependence, parallel control flow gr...
This article contains a brief description of existing graphical methods for presenting multithreaded...
High Performance Fortran (HPF) has emerged as a standard dialect of Fortran for data-parallel comput...
We present a comprehensive approach to performing data flow analysis in parallel. We identify three ...
Data parallel programming languages, such as HPF, are the easiest way to program Distributed Memory ...
Research on programming distributed memory multiprocessors has resulted in a well-understood program...
Over the past few decades, scientific research has grown to rely increasingly on simulation and othe...
[[abstract]]The data dependence graph (DDG) is a useful tool for the parallelism detection which is ...
The data-parallel language High Performance Fortran (HPF) does not allow efficient expression of mix...
We present a comprehensive approach to performing data flow analysis in parallel. We identify three ...
Current parallelizing compilers cannot identify a significant fraction of parallelizable loops becau...
Current parallelizing compilers cannot identify a significant fraction of parallelizable loops becau...
Loop distribution is an integral part of transforming a sequential program into a parallel one. It i...
. Data-parallel languages, in particular HPF, provide a highlevel view of operators overs parallel d...
Abstract. Multipartitioning is a skewed-cyclic block distribution that yields better parallel effici...
High Performance Fortran (HPF) does not allow ecient expression of mixed task/data-parallel computat...
This article contains a brief description of existing graphical methods for presenting multithreaded...
High Performance Fortran (HPF) has emerged as a standard dialect of Fortran for data-parallel comput...
We present a comprehensive approach to performing data flow analysis in parallel. We identify three ...
Data parallel programming languages, such as HPF, are the easiest way to program Distributed Memory ...
Research on programming distributed memory multiprocessors has resulted in a well-understood program...
Over the past few decades, scientific research has grown to rely increasingly on simulation and othe...
[[abstract]]The data dependence graph (DDG) is a useful tool for the parallelism detection which is ...
The data-parallel language High Performance Fortran (HPF) does not allow efficient expression of mix...
We present a comprehensive approach to performing data flow analysis in parallel. We identify three ...
Current parallelizing compilers cannot identify a significant fraction of parallelizable loops becau...
Current parallelizing compilers cannot identify a significant fraction of parallelizable loops becau...
Loop distribution is an integral part of transforming a sequential program into a parallel one. It i...
. Data-parallel languages, in particular HPF, provide a highlevel view of operators overs parallel d...
Abstract. Multipartitioning is a skewed-cyclic block distribution that yields better parallel effici...
High Performance Fortran (HPF) does not allow ecient expression of mixed task/data-parallel computat...
This article contains a brief description of existing graphical methods for presenting multithreaded...
High Performance Fortran (HPF) has emerged as a standard dialect of Fortran for data-parallel comput...
We present a comprehensive approach to performing data flow analysis in parallel. We identify three ...