The event was conducted by the Information Technology Department under the guidance of Dr. Duhai Al Shukali, Head of the IT Department, and in coordination with the Staff Development Unit at the University of Technology and Applied Sciences, Nizwa.
Ms. Sonia, coordinator of the SDU-IT Department welcomed Mr. Joji John and briefly introduce his experience in varied technologies and exposure to the current tools being used in the industry.
Mr. Joji took over the session with a comparison of the GPU-accelerated and the CPU-only applications. He highlighted the important facts about the need for GPU-accelerated processing and the use of Compute Unified Device Architecture in developing and implementing the Python Applications. He emphasized the necessity of CUDA with an exam[ple of a real-time application running over CCTV recordings to identify vehicles, face recognition to identify people, motion, actions, etc. wherein multi-processing is very important.
Mr. Joji explained the difference between Vanilla Python Vs CUDA, and the internal processing using the NVIDIA environment, the CUDA toolkit, and Tesla T4 GPU. He demonstrated the performance of the Tesla T4 GPU with an example. He further explained the use of the grid for improving performance and the need for synchronization between the CPU and GPU while using threads.
Mr. Joji John demonstrated a real-time example to recognize faces in motion and the faster processing on CUDA, which leads to a great performance of Face recognition systems in enhancing security using webcams/CCTV.
Ms. Sonia, coordinator, SDU-ITD proposed the vote of thanks.