Stanford Seminar - Computing with FPGAs - Oskar Mencer

Explore the world of computing with FPGAs in this Stanford seminar presented by Oskar Mencer. Delve into the reasons for using FPGAs, examining their performance advantages through examples like Boolean Satisfiability and Content Addressable Memory. Discover how FPGAs compare to microprocessors and DSP processors, and learn about stream architecture and its power-saving benefits. Investigate the challenges of programming FPGAs, including VLSI synthesis and the productivity gap in VLSI CAD. Gain insights into arithmetic for FPGAs, covering number representation, precision optimization, and compound arithmetic examples. Conclude with thoughts on syntax and semantics for FPGA programming, providing a comprehensive overview of this powerful computing technology.

Introduction.
Why Compute with FPGAS.
A Programmer's Technology Perspective.
Generic System Architecture.
Performance of FPGAS Examples.
Boolean Satisfiability.
BSAT FPGA versus Microprocessor.
Content Addressable Memory (CAM).
Memory-like Computation Example: Dynamic Graph Accelerator based on the Adjacency Matrix.
FPGAs versus DSP Processor International Data Encryption Algorithm (IDEA).
Stream Architecture Mapping a dataflow graph directly to the hardware.
Why Stream architectures save power.
Classification into Application Domains.
The VLSI CAD Productivity Gap.
Programming FPGAS #VLSI Synthesis.
The FPGA Programmers Task.
Adding Levels of Abstraction.
IDEA - Encryption.
What if there is a loop?.
Arithmetic for FPGAS.
Number Representation.
Precision / Range Optimization.
Arithmetic Styles.
Compound Arithmetic Example: Elementary Functions- undC math.
thoughts on syntax and semantics for programming FPGAS Syntax.
Conclusions.