Optimizing explicit data transfers for data parallel applications on the cell architecture
نویسندگان
چکیده
منابع مشابه
Optimizing Data Partitioning for Data-Parallel Computing
Performance of data-parallel computing (e.g., MapReduce, DryadLINQ) heavily depends on its data partitions. Solutions implemented by the current state of the art systems are far from optimal. Techniques proposed by the database community to find optimal data partitions are not directly applicable when complex user-defined functions and data models are involved. We outline our solution, which dr...
متن کاملOptimizing Data Alignment for Data Parallel Programs
Data decomposition across processors is critical to the performance of data parallel programs on distributed-memory machines. The data decomposition problem involves data alignment and data distribution. This paper addresses the data alignment phase which can be classiied into slope alignment and ooset alignment. We propose a data reference graph (DRG) model. Based on the DRG model, a slope ali...
متن کاملOptimizing Data Decomposition for Data Parallel Programs
A critical issue in achieving the performance of data parallel programs is how to eeciently decompose data across processors. On distributed-memory machines, a good data decomposition should increase processor workload balance and reduce interprocessor communication. Data decomposition consists of data distribution and data alignment. In this paper, we propose a trapezoid data distribution patt...
متن کاملOptimizing Data - Parallel Stencil
We have developed a communication optimizer that concentrates on stencil communication patterns. This optimizer has been done in the context of the UNH C* compiler that targets distributed-memory MIMD computers. Our work has two distinguishing features: The compiler/optimizer is designed to be highly portable. We achieve this goal by providing eecient support for the optimizations in the run-ti...
متن کاملLADS: Optimizing Data Transfers Using Layout-Aware Data Scheduling
While future terabit networks hold the promise of significantly improving big-data motion among geographically distributed data centers, significant challenges must be overcome even on today’s 100 gigabit networks to realize end-to-end performance. Multiple bottlenecks exist along the end-to-end path from source to sink. Data storage infrastructure at both the source and sink and its interplay ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: ACM Transactions on Architecture and Code Optimization
سال: 2012
ISSN: 1544-3566,1544-3973
DOI: 10.1145/2086696.2086716