Performance evaluation of apriori with memory mapped files
The concept of memory mapped files reduces the I/O data movement by mapping file data directly to the process address space. This is best suitable for the data mining applications which involve accessing large data files. The recent improvement in parallel processor architectures is the multi-core architectures. To get the real benefit from these architectures we have to redesign the existing serial algorithms so that they can be parallelized on multi-core architectures. OpenMP is an API for parallel programming which make a serial program to run in parallel without much redesigning job. Our main concern in this paper is to evaluate the performance of apriori using linux mmap() function compared to fread() function in both the serial and parallel environments. Experiments are conducted with both simulated and standard datasets on multi-core architectures using openMP threads. Our experiments show that mmap() function gives better results than fread() function with both serial as well as parallel implementations of apriori on dual core.
Keywords: apriori, fread(), mmap(), multi-core, OpenMP
Download Full-Text
ABOUT THE AUTHORS
T.Anuradha
Mrs.T.Anuradha is working as an associate professor in the department of ECM in KLUniversity,Guntur,A.P.,India.She is currently doing her Ph.d.under the supervision of Dr.R.Satya Prasad. Her research interests include datamining and web technologies.
R.Satya Prasad
Dr.R.Satya Prasad is working as an associate Professor in the department of CS&Engineering in Nagarjuna University,Guntur,A.P., India. He got his Ph.d., degree from Nagarjuna University and presently he is guiding Ph.d., scholars in the areas of data mining and Software Engineering.
S.N.Tirumala Rao
Dr.S.N.TirumalaRao is working as Professor in Narasaraopeta Engineering College, Guntur, A.P., India.He got his Ph.d., degree from JNTU, Kakinada. He is interested in data mining and parallel processing on multicore architectures.
T.Anuradha
Mrs.T.Anuradha is working as an associate professor in the department of ECM in KLUniversity,Guntur,A.P.,India.She is currently doing her Ph.d.under the supervision of Dr.R.Satya Prasad. Her research interests include datamining and web technologies.
R.Satya Prasad
Dr.R.Satya Prasad is working as an associate Professor in the department of CS&Engineering in Nagarjuna University,Guntur,A.P., India. He got his Ph.d., degree from Nagarjuna University and presently he is guiding Ph.d., scholars in the areas of data mining and Software Engineering.
S.N.Tirumala Rao
Dr.S.N.TirumalaRao is working as Professor in Narasaraopeta Engineering College, Guntur, A.P., India.He got his Ph.d., degree from JNTU, Kakinada. He is interested in data mining and parallel processing on multicore architectures.