Preliminary experiments with Maxeler framework

• Main Author: Choo, Yong Cheng
• Other Authors: School of Computer Engineering
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Performance of systems
• Online Access: http://hdl.handle.net/10356/58989
• Institution: Nanyang Technological University

With an increasing volume and velocity of data in today's applications in many industries such as financial, healthcare and weather predictions, new acceleration tools and methods are coming up rapidly. There is an insatiable demand for more computing power while also achieving accurate and fast results. High Performance Computing is the backbone of these new accelerators. I will be surveying existing techniques on acceleration for complex calculations in the computational finance industry with the 3 platforms CPU, GPU and FPGA. Thus, I will be reporting speedups and conduct detailed experiments to quantify benefits of acceleration in comparison of the 3 platforms. At the end of the day, the deliverable is to use utilize Maxeler framework to optimise and choose the best fit 3D Convolution Array Blocking size.