A Software Pipelining Method Based on a Hierarchical Social Algorithm
نویسندگان
چکیده
Software pipelining is a compile-time scheduling technique that overlaps successive loop iterations to achieve instruction-level parallelism. It allows us to hide memory latency by overlapping the prefetches for a future iteration with the computation of the current iteration. This paper presents an efficient algorithm for determining the iteration bound of cyclic data flow graphs and the optimal loop scheduling. To solve this problem, a new general metaheuristic has been developed. This metaheuristic is inspired in some hierarchical social structures and processes. Several variant of the proposed algorithm has been implemented and tested on some benchmark programs.
منابع مشابه
A partition-based algorithm for clustering large-scale software systems
Clustering techniques are used to extract the structure of software for understanding, maintaining, and refactoring. In the literature, most of the proposed approaches for software clustering are divided into hierarchical algorithms and search-based techniques. In the former, clustering is a process of merging (splitting) similar (non-similar) clusters. These techniques suffered from the drawba...
متن کاملEquivalence Checking of Loops before and after Pipelining by Applying Symbolic Simulation and Induction
When applications contain large loops, high level synthesis often takes advantage of software pipelining technique in order to improve the performance. High level synthesis with pipelining utilization needs complicated algorithms. So it is desired to check its correctness. In this paper, we propose a novel approach for equivalence checking of loops before and after pipelining. The proposed appr...
متن کاملHigh-Dimensional Unsupervised Active Learning Method
In this work, a hierarchical ensemble of projected clustering algorithm for high-dimensional data is proposed. The basic concept of the algorithm is based on the active learning method (ALM) which is a fuzzy learning scheme, inspired by some behavioral features of human brain functionality. High-dimensional unsupervised active learning method (HUALM) is a clustering algorithm which blurs the da...
متن کاملTechnical Report an Enhanced Co-scheduling Method Using Reduced Ms-state Diagrams 1
Instruction scheduling methods based on the construction of state diagrams (or automatons) have been used for architectures involving deeply pipelined function units. However, the size of the state diagram is prohibitively large, resulting in high execution time and space requirement, which in turn, restrict the use of these methods. In this paper, we develop the underlying theory for reducing ...
متن کاملSoftware Pipelining for the Pegasus IR
Modern processors, especially VLIW processors, often have the ability to execute multiple instructions simultaneously. Taking advantage of this capability is crucial for high performance software applications. Software pipelining is a technique designed to increase the level of parallelism in loops. We propose a new approach to software pipelining based on direct manipulations of control flow g...
متن کامل