Sign In
 

|

Date: Monday, November 22, 2021

Time: 01:15 PM – 02:15 PM

Location: ONLINE by Using the Link: (https://kfupm.zoom.us/j/98499651720)

Abstract:

Massively Parallel Computing is gaining ground in high-performance computing using various accelerator architectures. Recent advances in accelerated computing for proper utilization of these architectures opened a new opportunity in harnessing their computing power for general-purpose computing. The massive data-parallel computing power provided by inexpensive commodity Graphics Processing Units (GPUs) makes large-scale data processing on GPUs and GPU-accelerated clusters attractive from both a research and practical perspective. Both commercial and technical systems today embrace accelerated computing to handle jobs such as machine learning, data analytics, simulations, and visualizations. It's a modern style of computing that delivers high performance and energy efficiency. The scalable computing power of GPUs is highly essential for scientific simulations, especially for the class of Structured Grid Computing (SGC). In order to achieve higher simulation accuracy, large-scale simulations of the problem with highly efficient code are required. We identified GPU architectural optimizations and techniques used in research and industry compilers to produce optimized code. Optimizing SGCs is found to be complex, error-prone, and involve a variety of heterogeneous tools and techniques, which can be envisioned only from a research perspective. However, spreading the use of SGC on GPUs requires a deliberate effort for identifying the required automatic techniques for alleviating the complexity and the integration within a well-engineered framework. We presented the details of these techniques and described an integrated library with the required essential functionalities to ease the process of developing efficient storage, optimized code by using a high-level interactive interface and intelligent domain-specific annotations. In addition to that, a software API for the fast development of large-scale deep learning models for big data analytics has been proposed to analyze the datasets with a higher amount of volume, velocity, and variety.

Biography:

Dr. Ayaz Ul Hassan Khan is a skilled techno master and an accomplished computer scientist. With a diverse experience of 15 years in industry and academics, he contributed positively to several universities and professional organizations with his excellent problem-solving skills. As a technology professional, he is experienced to deal with many international clients including Dubai Financial Market, Dubai Municipality, Emirates, Comets Services, Doha Securities Market, Tensator UK, CreditOne Bank of USA, Quadrem Global Supply Chain Solutions, American Honda Motor Company, BMW, Nautilus Incorporation, Bentley Motors, AT&T, and Volkswagen to provide the cutting edge solutions for their day-to-day business processes. Some examples of his services include the development of Queuing Management Systems, Online Ticker Systems, Self-Service KIOSKs, ETL processes, and reporting, etc. As a computer scientist with an interest in high-performance computing, parallel programming, and deep learning, he is an author of 20+ publications in reputed journals and conferences along with a book on parallel processing. He has successfully completed several research-funded projects and a few more are in process of the accumulated worth of about half a million dollars. Dr. Ayaz Ul Hassan Khan holds a Ph.D. degree in Computer Science and Engineering from King Fahd University (Saudi Arabia), Bachelor of Computer Science and Information Technology from NED University (Pakistan) with secured 2nd position (90%) and MS Computer Science from Lahore University of Management Sciences with 3.61 CGPA. In addition to these, he possesses dozens of professional certifications for various technical and soft skills. Dr. Ayaz is an occasional traveler at both local and international levels. He has traveled to 60+ cities of 7 countries on 3 different continents. He has achieved a Local Guide – Level 6 on Google maps with a 5-star rating and accumulated 1953 points for his travel contributions.


31 Mar 2022