Document Details

Document Type : Thesis 
Document Title :
IMPROVING THE PERFORMANCE OF JACOBI ITERATIVE METHOD ON INTEL MIC ARCHITECTURE
تحسين أداء طريقة جاكوبي بإستخدام معمارية إنتل متعددة الأنوية
 
Subject : Faculty of Computing and Information Technology 
Document Language : Arabic 
Abstract : Numerous important scientific and engineering applications rely on and are hindered by, the intensive computational and storage requirements of solving large sparse linear equation systems. The numerical methods for solving linear equations can be categorised into direct methods and iterative methods. The Jacobi method is a widely used iterative solver due to its simplicity and efficiency. Its performance is affected by factors including the storage format, the specific computational algorithms and its implementation. While the performance of Jacobi has been studied extensively on conventional CPU architectures, research on its performance on emerging architectures, such as the Intel Many Integrated Core (MIC) Architecture, is still in its infancy. The aim of this thesis is to efficiently implement the Jacobi method on Intel MIC. A performance analysis of the parallel implementation of Sparse Matrix Vector Multiplication (SpMV) and the Jacobi method on the first-generation of Intel MIC, namely Knights Corner (KNC), is provided. Four versions of Jacobi in addition to the sequential implementation have been implemented. Jacobi was implemented with two storage formats: Compressed Sparse Row (CSR) and Modified Sparse Row (MSR). Offloading mode and OpenMP was used. The performance was measured in terms of the execution time and the offloading time. The results of the sparse matrices were reported, with over 28 million rows and 640 million nonzero elements acquired from thirteen diverse application domains.The experimental results showed that this Jacobi parallel implementation on MIC achieved speedups of up to 27.75x compared to sequential implementation. It delivered a speedup of up to 3.81x compared to a powerful Multi-Core node comprising 24 cores (two Intel Xeon E5-2695v2 processors,12 cores each).Future work will look into further analysis of the parallel Jacobi method for sparse linear equations systems of larger sizes with the aim of further improving performance 
Supervisor : Prof. Rashid Mehmood 
Thesis Type : Master Thesis 
Publishing Year : 1439 AH
2018 AD 
Co-Supervisor : Dr. Iyad Katib 
Added Date : Monday, January 29, 2018 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
هناء محمد اليحياءAl-Yahya, Hana MohamedResearcherMaster 

Files

File NameTypeDescription
 43034.pdf pdf 

Back To Researches Page