Estimating parallel performance

Lobachev, Oleg; Guthe, Michael; Loogen, Rita

doi:10.1016/j.jpdc.2013.01.011

Cited by 12 publications

(7 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lobachev et al [26] provides a mechanism to estimate performance of massively parallel code. It also allows exact measurement and prediction with Scalability in terms of problem size and processor members.…”

Section: Gpgpus For Parallelizationmentioning

confidence: 99%

Sentiment and Emotion Analysis for Context Sensitive Information Retrieval of Social Networking Sites: A Survey

GeorgeAmalarethinam¹,

Nirmal²

2014

IJCA

View full text Add to dashboard Cite

Context Sensitive Information Retrieval (CSIR) is quite a challenging issue because of the complexities involved in the process from semantics and ontology to the huge amount of processing capacity required to make it possible in real time. Understanding the semantic gap (where context is neglected) plays a major role in elimination false positives and improving the true positives in the information retrieval process. With big data becoming ubiquitous due to the volume, velocity and variety of data being presented and analysed in almost all the domains today, context sensitive analysis and interpretation of big data becomes important. This paper presents a comprehensive survey of the existing techniques for big data analysis based on massively parallel processing techniques like GPGPUs (CUDA), Hadoop MapReduce and also Data Warehousing. This paper presents a discussion about the datasets that are available for research and also the applications that could be thought of by context sensitive analysis of social media data. Also this paper provides research directions for context sensitive information retrieval and sentiment analysis in big data based on massively parallel processing architecture.

show abstract

Section: Gpgpus For Parallelizationmentioning

confidence: 99%

Sentiment and Emotion Analysis for Context Sensitive Information Retrieval of Social Networking Sites: A Survey

GeorgeAmalarethinam¹,

Nirmal²

2014

IJCA

View full text Add to dashboard Cite

show abstract

“…Similar to Ulrich et al [2014b], we perform an optimisation step over all slices at once to minimise the energy of the non-linear distortions based on detected features ('global' registration). In contrast to the previous work, we 1. Use a multi-resolution approach, Our new method processes not only the features of the scale of the initial rigid transform, but also iteratively much smaller features.…”

Section: Visual and Operational Comparison With Ulrich Et Almentioning

confidence: 99%

Three Dimensional Reconstruction

Ikeuchi¹

2014

Computer Vision

View full text Add to dashboard Cite

GPU implementation with tandem simplification [Ulrich et al., 2014a, Chapter 4] or a module from 3D Slicer [Fedorov et al., 2012, Kikinis et al., 2014]. The meshes are further processed. In most cases we at least heal and remesh the representation using PolyMender [Ju, 2004], then use Taubin smoothing to obtain a nicer surface. After some further processing (e.g., removal of very small unconnected components or colouring) the mesh is ready for quality control, rendering, and evaluation by experts.

show abstract

“…It starts by predicting the execution time of the MDFG. Based on [16,17], this term consists of estimating two parameters whose are the computation time and the communication time. The first one is predicted by multiplying the cycle period and cycle number of the whole application.…”

Section: Technique Choicementioning

confidence: 99%

Execution Time and Code Size Optimization Using Multidimensional Retiming and Loop Striping

Elloumi

Akil

Bedoui

2013

2013 Euromicro Conference on Digital System Design

View full text Add to dashboard Cite

Nested loops present the most critical sections in several embedded real-time applications. To achieve a higher performance, the design process employs an optimization technique in order to increase parallelism. However, the nested loop codes rise greatly in terms of parallelism level. Due to tight execution time constraints, each optimization technique produces implementations with an important code size. This criterion presents a limiting factor to implement the provided results in embedded real-time systems. In this paper, we propose a novel optimization approach that combines the delayed multidimensional retiming and loop striping techniques. It explores the solution space, which is composed by all parallelism cases proposed by both techniques, in order to provide the implementation that achieves the execution time constraint while uses a lower code size. In this context, we present the theory of combining both techniques. Then, we propose efficient algorithms that ensure selecting a set of parallelism transformations, based on their execution time and code size evolutions. The experimental results show that our optimization approach provides optimal solutions compared to those provided by applying only one technique. It achieves average improvements on the code size of 35.21% compared to the delayed multidimensional retiming and 16.38% compared to the loop striping.

show abstract

Estimating parallel performance

Cited by 12 publications

References 27 publications

Sentiment and Emotion Analysis for Context Sensitive Information Retrieval of Social Networking Sites: A Survey

Sentiment and Emotion Analysis for Context Sensitive Information Retrieval of Social Networking Sites: A Survey

Three Dimensional Reconstruction

Execution Time and Code Size Optimization Using Multidimensional Retiming and Loop Striping

Contact Info

Product

Resources

About