Date of Award


Publication Type

Master Thesis

Degree Name



Electrical and Computer Engineering


Expectation Maximization, FPGA, High-Level Synthesis, High Performance Computing, Intel FPGA SDK for OpenCL, Reconfigurable Computing


Khalid, Mohammed




Expectation Maximization (EM) is a soft clustering algorithm which partitions data iteratively into M clusters. It is one of the most popular data mining algorithms that uses Gaussian Mixture Models (GMM) for probability density modeling and is widely used in applications such as signal processing and Machine Learning (ML). EM requires high computation time and large amount of memory when dealing with large data sets. Conventionally, the HDL-based design methodology is used to program FPGAs for accelerating computationally intensive algorithms. In many real world applications, FPGA provide great speedup along with lower power consumption compared to multicore CPUs and GPUs. Intel FPGA SDK for OpenCL enables developers with no hardware knowledge to program the FPGAs with short development time. This thesis presents an optimized implementation of EM algorithm on Stratix V and Arria 10 FPGAs using Intel FPGA SDK for OpenCL. Comparison of performance and power consumption between CPU, GPU and FPGA is presented for various dimension and cluster sizes. Compared to an Intel(R) Xeon(R) CPU E5-2637 our fully optimized OpenCL model for EM targeting Arria 10 FPGA achieved up to 1000X speedup in terms of throughput (Tspeedup) and 5395X speedup in terms of throughput per unit of power consumed (T/Pspeedup). Compared to previous research on EM-GMM implementation on GPUs, Arria 10 FPGA obtained up to 64.74X Tspeedup and 486.78X T/Pspeedup.