Hardware Accelerated Computing (ELEN90096)

Hardware acceleration for computationally intensive applications is of growing importance for improving workload performance in cloud data centres, the network edge, and IoT embedded devices. This subject introduces students to the basics of hardware design for field programmable gate arrays (FPGAs) which are widely used to accelerate algorithms in applications areas such as machine learning, artificial intelligence, networking, cryptography, and multimedia signal processing. In addition to covering FPGA fundamentals, the subject will take a systems-based approach to analysing algorithms for suitability of acceleration and mapping to heterogeneous computing resources. 

Topics covered in this subject may include:  

• Review of combinational and sequential digital logic 
• FPGA architectures and fundamentals 
• Hardware description languages (Verilog/VHDL) and hardware design flows 
• High-level synthesis and OpenCL 
• The use of parallelism, locality, and precision in hardware accelerators 
• Host-accelerator interactions and hardware-software co-design 
• Optimisation of hardware designs with respect to throughput, latency, energy, and area 
• Accelerator design for selected applications such as machine learning, artificial intelligence, networking, cryptography, and multimedia signal processing 

As part of this subject, students will complete a significant design project in which they design, implement, verify, and benchmark a hardware accelerator for a selected application

Intended learning outcomes

On completion of this subject, students should be able to:

• Explain hardware accelerator architectures and computational models
• Implement digital logic functions in a variety of industrial tools, such as hardware description languages, high-level synthesis tools, and OpenCL
• Analyse advanced computational algorithms for amenability to hardware acceleration, mapping to heterogeneous computing systems to exploit parallel computation
• Design hardware to accelerate computationally intensive algorithms using a systems-based approach with consideration given to trade-offs in speed, energy efficiency, and area
• Articulate the importance of hardware acceleration for computational systems as a continuing trend in industry

Generic skills

• Capacity for independent critical thought, rational inquiry and self-directed learning
• Ability to undertake problem identification, formulation, and solution
• Ability to utilise a systems approach to design and operational performance
• Ability to work effectively in a team environment in order to produce a satisfactory project outcome