• Reconfigurable Computing: The Theory and Practice of FPGA-Based Computation (Systems on Silicon) | | | |
• Synthesis of Arithmetic Circuits: FPGA, ASIC and Embedded Systems | | | |
• Introduction to Reconfigurable Computing: Architectures, Algorithms, and Applications | | | |
• Advanced FPGA Design: Architecture, Implementation, and Optimization | | | |
• Rapid System Prototyping with FPGAs: Accelerating the design process (Embedded Technology) | | | |

First book in a new field
Review written by: wiredweird From Earth, or somewhere nearby
FPGA-based reconfigurable computing may be the first fundamental innovation in computing since von Neumann's stored programs. Gokhale, one of the founders of the field, has written the first book I know of that summarizes current state of the art.

The book is brief and dense. It starts with introductory material about what FPGAs are and why they're good for some problems. The later chapters might not have full impact if you aren't already familair with FPGAs, though. Next, there's a short discussion of compilers. To tell the truth, state of the art right now is pretty grim, about where software tools were in the 1960s. Things like VHDL have tamed logic design, but computation is a very different beast. We need something different, something that isn't there yet - no matter what the inflated claims of the tool-builders.

The remainder of the book discusses application areas. The first two chapters in this section cover the classic FPGA computing applications - signal processing and image processing, both emerging from their niches in the defense market. The last three chapters cover network security, often an FPGA-based super-Snort, bioinformatics, and applications in heat transfer and road traffic analysis. Even though FPGA computing has largely been a research field rather than a production technology, it's still too big to cover in one book. Gokhale missed some interesting bioinformatics applications, such as microarray and regulatory network analysis, and other applications including electromagnetics, molecular dynamics, and astrophysical models. Still, shes has picked good representatives of the FPGA computing field.

Now that Silicon Graphics's RASC and Cray's XD1 add-on are on the market, we'll see a lot about FGPA-based computing. In a few years, when there are more tools and tool users in the field, this may look dated. For now, if you want to know about FPGA-based computing, you'll read the proceedings of the research conferences or you'll read this.

//wiredweird