Custom Computing

Ovewview

Custom computers are special-purpose systems customised for specific applications such as signal processing and database operations, when general-purpose computers are too slow, too bulky or consume too much power. Traditionally the development of custom computers is an expensive, time-consuming and error-prone activity. This course introduces an approach which enables the rapid and systematic design of custom computers.

Learning Outcomes

On successful completion of the module, students should be able to:

• develop parametric descriptions of custom computers,
• develop alternative designs for custom computers that meet specified requirements,
• analyse the performance of a custom computer in terms of time and space,
• evaluate space/time trade-offs between competing custom computing designs in order to determine optimal solutions,
• use simulation to compare the intended and actual behaviour of custom computers.

• Course plan, lccture notes and other materials: Scientia link
• Ruby notes: introduction (Section 1.1 - 1.7), state machines, priority queues, binomial filters, integer divider, median filters, rank evaluators, serial convolvers
• Ruby tools: Rebecca setup, examples, prelude, simulation example, simulator primitives

References

• motivation: Microsoft AI cloud, Amazon F1, floating-point FPGA, Sony reconfigurable circuits, Makimoto's wave, survey1, survey2, DSP applications, Harmonic, roadmap, FPGA architecture, FPGA tools, system-on-chip.
• applications: monitor self-adaptive systems, run-time verification, medical imaging, financial modelling, financial option evaluation, financial risk analysis, telescope data processing, particle physics, video processing, posture analysis, computer graphics, network flow analysis, software defined radio, multithreaded processor, machine learning
• custom computing systems: Axel, Convey, Cube, web search engines, crypto system, Sonic, UltraSonic, PAM, Splash2, Pilchard, Riley2, video engine, Harp1
• background: introduction to computer architecture, architecture at Stanford, architecture at UC Berkeley, advanced architecture, Modern VLSI Design

Links

• the future of microprocessor, ASIC vs FPGA, processor overview, Leon, Nano FPGA, quantum FPGA, IRAM
• Symposium on Field-Programmable Custom Computing Machines, Conference on Field-Programmable Technology, Conference on Field-Programmable Logic and Applications