ONGOING PROJECT:
1. Image Segmentation Using SNET Compiler (Description about SNET http://snet-home.org/)
This project deals with the development of parallel image segmentation algorithms using SNET compiler. To find out more details about the SNET compiler, please follow the above mentioned SNET description link.
2. Parallel adaptive Arbitration Using Hadoop Framework for Image Processing Applications
Arbitration employing distributed and parallel computing has become a requirement in industries which handle tasks with big data and various high-end computing applications comprising bulk of huge images and video data banks. In order to reduce the task execution time of bulk image processing, researchers have come up with various arbitration methodologies which uses parallel computation of tasks. At the same time, the researchers have to maximize the performance of tasks using high-end multi-core CPUs with advanced buses and interconnections to attain high bandwidth with low system latency. Nonetheless, earlier implemented arbitration techniques have not been able to exploit maximum CPU core usage due to inability to implement an acceptable degree of task parallelism. This prolongs task execution time of applications. Hence, this research develops a parallel arbitration scheme called Parallel Adaptive Arbitration (PAA) to overcome the issues with bulk image data processing. Firstly, the PAA was designed with parallel implementation of masters with real-time parameter which works as per the data traffic behaviour. Then, an intense loop benchmark called STREAM (Sustainable Memory Bandwidth in High Performance Computers) was used to evaluate the performance of PAA in terms of task execution time and CPU cores usage. Once the performance benchmarking was done based on task parallelization, PAA was implemented on Hadoop distributed framework to perform image segmentation task, which is a prominent step in image processing. The performance of designed PAA for bulk image segmentation on Hadoop framework has been compared with segmentation process using sequential programming method in visual studio integrated development environment platform. The assessments aimed to investigate the optimal threshold of data size at which PAA is able to outperform sequential programming method in terms of task execution time using multiple CPU cores and threads. The results have shown improved performance with PAA when compared to sequential programming method particularly when data capacity is greater than 250 MB. This is because PAA has been able to utilize the CPU usage with octacore up to 96%, hence reducing the task execution time up to approximately 1.7 times compared with the sequential style of programming. The proposed parallel arbitration design has shown to be useful for researchers at performing bulk image segmentation in parallel, which can save tremendous execution time.
INTERNATIONAL PUBLICATION :
[1] M. Nishat Akhtar, J.M Saleh and O. Sidek, “Design and simulation of a parallel adaptive arbiter for maximum CPU utilization using multi-core processors”, in Journal of Computer and Electrical Engineering- pp-51-68, vol. (47), 2015.
[2] M. Nishat Akhtar and J.M Saleh, “Efficient CPU core usage and balanced bandwidth distribution using smart adaptive arbitration”, in Indian Journal of Science and Technology- vol. 10 (6), pp-1-10, 2017.
[3] M. Nishat Akhtar and J.M Saleh, “Parallel Adaptive Arbiter for Improved CPU Utilization and Fair Bandwidth Allocation”, in INASE 2015- Austria, pp-241-252.
[4] M. Nishat Akhtar and O. Sidek, “ An Intelligent Adaptive Arbiter for Maximum CPU Utilization, Fair Bandwidth Allocation and Low Latency”, in IETE Journal of Research- pp-48-54, Jan-Feb-issue 2013.
[5] M.Nishat Akhtar and O.Sidek, “An Adaptive Arbitration Algorithm for Fair Bandwidth Allocation, Low Latency and Maximum CPU Utilization”, Part I, CCIS 293, pp. 330–343, 2012. © Springer-Verlag Berlin Heidelberg 2012.
[6] Danish M, Khanday WA, Hashim R, Sulaiman NS, M.Nishat Akhtar , Nizami M., “Application of optimized large surface area date stone (Phoenix dactylifera ) activated carbon for rhodamin B removal from aqueous solution: Box-Behnken design approach”, in Journal of Ecotoxicology and Environment Safety- vol. (139), pp-280-290, 2017.
[7] Muhammad Rafiq Kakar, Meor Othman Hamzah, M.Nishat Akhtar, David Woodward, (2016), Surface free energy and moisture susceptibility evaluation of asphalt binders modified with surfactant-based chemical additive, Journal of Cleaner Production, 112 (2016) pp. 2342-2353.
[8] M. Nishat Akhtar and O. Sidek, “An arbiter with fair bandwidth allocation and low latency for real time computing system,” in 3rd International Conference on Computer Technology and Development, 2011, pp. 189-195.
[9] M.Nishat Akhtar and O.Sidek, “An intelligent arbiter for fair bandwidth allocation,” presented at the 9th IEEE SCORED 2011, pp. 322-327.
[10] M.Nishat Akhtar and O.Sidek, “An Intelligent Arbiter for Maximum CPU Utilization, Fair Bandwidth Allocation and Low Latency:Survey”, in IEEE-CSPA-2012.