Porting MassiveThreads Thread Library to FX10 Supercomputer System
スポンサーリンク
概要
- 論文の詳細を見る
We ported MassiveThreads, a light-weight thread library, to FX10 supercomputer system. We illustrate the technical challenges of implementing context switching for SPARC architecture, and identify the performance bottlenecks in context switch implementations in current Linux kernel and GNU C library. The evaluation shows that our user-level implementation in assembly is over 30x faster than using library context switch routines, e. g., ucontext series. Our implementation also enables the Chapel programming language to run more efficiently on FX10 when using MassiveThreads than other tasking layers such as Pthreads, Nanos++, and Qthreads.
- 2012-12-06
著者
-
Nan Dun
The University Of Tokyo
-
Kenjiro Taura
The University of Tokyo
-
Jun Nakashima
The University of Tokyo
関連論文
- Cynk: A Hybrid Rsync and SSH Filesystem for Cloud Computing
- A Preliminary Evaluation of Chapel with Molecular Dynamics Simulation
- A Preliminary Evaluation of Chapel with Molecular Dynamics Simulation
- A Preliminary Evaluation of Chapel with Molecular Dynamics Simulation
- A Preliminary Evaluation of Chapel with Molecular Dynamics Simulation
- Porting MassiveThreads Thread Library to FX10 Supercomputer System
- Porting MassiveThreads Thread Library to FX10 Supercomputer System
- A Power State Transition Algorithm for Power Controller of Energy Efficient Computer Clusters
- A Performance Analyzer for Task Parallel Applications based on Execution Time Stretches