To Parallelize or Not to Parallelize, Speed Up Issue
نویسنده
چکیده
Running parallel applications requires special and expensive processing resources to obtain the required results within a reasonable time. Before parallelizing serial applications, some analysis is recommended to be carried out to decide whether it will benefit from parallelization or not. In this paper we discuss the issue of speed up gained from parallelization using Message Passing Interface (MPI) to compromise between the overhead of parallelization cost and the gained parallel speed up. We also propose an experimental method to predict the speed up of MPI applications.
منابع مشابه
To parallelize or not to parallelize, control and data flow issue
New trends towards multiple core processors imply using standard programming models to develop efficient, reliable and portable programs for distributed memory multiprocessors and workstation PC clusters. Message passing using MPI is widely used to write efficient , reliable and portable applications. Control and data flow analysis concepts, techniques and tools are needed to understand and ana...
متن کاملSpeeding up Index Construction with Gpu for Dna Data Sequences
The advancement of technology in scientific community has produced terabytes of biological data. This datum includes DNA sequences. String matching algorithm which is traditionally used to match DNA sequences now takes much longer time to execute because of the large size of DNA data and also the small number of alphabets. To overcome this problem, the indexing methods such as suffix arrays or ...
متن کاملA Parallel Saturation Algorithm on Shared Memory Architectures
Symbolic state-space generators are notoriously hard to parallelize. However, the Saturation algorithm implemented in the SMART verification tool differs from other sequential symbolic state-space generators in that it exploits the locality of firing events in asynchronous system models. This paper explores whether event locality can be utilized to efficiently parallelize Saturation on shared-m...
متن کاملReasons not to Parallelize TCP Connections for Long Fat Networks
For utilizing long fat networks effectively, parallel TCP was proposed, and has been employed. However, as high-speed transport-layer protocols appear, it is necessary to reinvestigate the performance of parallel TCP. In this paper, we use mathematical analysis to explore its performance. Analysis results show that the open issue of choosing the number of TCP connections is difficult to be solv...
متن کاملParallel Computations for Hierarchical Agglomerative Clustering using CUDA Fast and Scalable Computations on Graphics Processors
Graphics Processing Units (GPU) in today’s desktops can well be thought of as a high performance parallel processor. Traditionally, parallel computing is the usage of multiple computing resources to execute computational problems simultaneously. Such computations are possible using multi-core CPUs or computers with multiple CPUs or by using a network of computers in parallel. Today’s GPUs are c...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- CoRR
دوره abs/1103.5616 شماره
صفحات -
تاریخ انتشار 2011