Field programmable gate arrays (FPGAs) are an increasingly popular technology for implementing digital signal processing (DSP) systems. By allowing designers to create circuit architectures developed for the specific applications, high levels of performance can be achieved for many DSP applications providing considerable improvements over conventional microprocessor and dedicated DSP processor solutions. The book addresses the key issue in this process specifically, the methods and tools needed for the design, optimization and implementation of DSP systems in programmable FPGA hardware. It presents a review of the leading-edge techniques in this field, analyzing advanced DSP-based design flows for both signal flow graph- (SFG-) based and dataflow-based implementation, system on chip (SoC) aspects, and future trends and challenges for FPGAs. The automation of the techniques for component architectural synthesis, computational models, and the reduction of energy consumption to help improve FPGA performance, are given in detail. Written from a system level design perspective and with a DSP focus, the authors present many practical application examples of complex DSP implementation, involving: high-performance computing e.g. matrix operations such as matrix multiplication; high-speed filtering including finite impulse response (FIR) filters and wave digital filters (WDFs); adaptive filtering e.g. recursive least squares (RLS) filtering; transforms such as the fast Fourier transform (FFT). FPGA-based Implementation of Signal Processing Systems is an important reference for practising engineers and researchers working on the design and development of DSP systems for radio, telecommunication, information, audio-visual and security applications. Senior level electrical and computer engineering graduates taking courses in signal processing or digital signal processing shall also find this volume of interest.

/pdf/fpga-based-implementation-of-signal-processing-systems-2dq7no5xii.pdf

FPGA-based Implementation of Signal Processing Systems

Secure multi-party computation (MPC) is a technique well suited for privacy-preserving data mining. Even with the recent progress in two-party computation techniques such as fully homomorphic encryption, general MPC remains relevant as it has shown promising performance metrics in real-world benchmarks. Sharemind is a secure multi-party computation framework designed with real-life efficiency in mind. It has been applied in several practical scenarios, and from these experiments, new requirements have been identified. Firstly, large datasets require more efficient protocols for standard operations such as multiplication and comparison. Secondly, the confidential processing of financial data requires the use of more complex primitives, including a secure division operation. This paper describes new protocols in the Sharemind model for secure multiplication, share conversion, equality, bit shift, bit extraction, and division. All the protocols are implemented and benchmarked, showing that the current approach provides remarkable speed improvements over the previous work. This is verified using real-world benchmarks for both operations and algorithms.

High-performance secure multi-party computation for data mining applications

Preface. About the Authors. 1. Introduction. 1.1 Number Representation. 1.2 Algorithms. 1.3 Hardware Platforms. 1.4 Hardware-Software Partitioning. 1.5 Software Generation. 1.6 Synthesis. 1.7 A First Example. 1.7.1 Specification. 1.7.2 Number Representation. 1.7.3 Algorithms. 1.7.4 Hardware Platform. 1.7.5 Hardware-Software Partitioning. 1.7.6 Program Generation. 1.7.7 Synthesis. 1.7.8 Prototype. 1.8 Bibliography. 2. Mathematical Background. 2.1 Number Theory. 2.1.1 Basic Definitions. 2.1.2 Euclidean Algorithms. 2.1.3 Congruences. 2.2 Algebra. 2.2.1 Groups. 2.2.2 Rings. 2.2.3 Fields. 2.2.4 Polynomial Rings. 2.2.5 Congruences of Polynomial. 2.3 Function Approximation. 2.4 Bibliography. 3. Number Representation. 3.1 Natural Numbers. 3.1.1 Weighted Systems. 3.1.2 Residue Number System. 3.2 Integers. 3.2.1 Sign-Magnitude Representation. 3.2.2 Excess-E Representation. 3.2.3 B's Complement Representation. 3.2.4 Booth's Encoding. 3.3 Real Numbers. 3.4 Bibliography. 4. Arithmetic Operations: Addition and Subtraction. 4.1 Addition of Natural Numbers. 4.1.1 Basic Algorithm. 4.1.2 Faster Algorithms. 4.1.3 Long-Operand Addition. 4.1.4 Multioperand Addition. 4.1.5 Long-Multioperand Addition. 4.2 Subtraction of Natural Numbers. 4.3 Integers. 4.3.1 B's Complement Addition. 4.3.2 B's Complement Sign Change. 4.3.3 B's Complement Subtraction. 4.3.4 B's Complement Overflow Detection. 4.3.5 Excess-E Addition and Subtraction. 4.3.6 Sign-Magnitude Addition and Subtraction. 4.4 Bibliography. 5. Arithmetic Operations: Multiplication. 5.1 Natural Numbers Multiplication. 5.1.1 Introduction. 5.1.2 Shift and Add Algorithms. 5.1.2.1 Shift and Add 1. 5.1.2.2 Shift and Add 2. 5.1.2.3 Extended Shift and Add Algorithm: XY t C t D. 5.1.2.4 Cellular Shift and Add. 5.1.3 Long-Operand Algorithm. 5.2 Integers. 5.2.1 B's Complement Multiplication. 5.2.1.1 Mod Bntm B's Complement Multiplication. 5.2.1.2 Signed Shift and Add. 5.2.1.3 Postcorrection B's Complement Multiplication. 5.2.2 Postcorrection 2's Complement Multiplication. 5.2.3 Booth Multiplication for Binary Numbers. 5.2.3.1 Booth-r Algorithms. 5.2.3.2 Per Gelosia Signed-Digit Algorithm. 5.2.4 Booth Multiplication for Base-B Numbers (Booth-r Algorithm in Base B). 5.3 Squaring. 5.3.1 Base-B Squaring. 5.3.1.1 Cellular Carry-Save Squaring Algorithm. 5.3.2 Base-2 Squaring. 5.4 Bibliography. 6 Arithmetic Operations: Division. 6.1 Natural Numbers. 6.2 Integers. 6.2.1 General Algorithm. 6.2.2 Restoring Division Algorithm. 6.2.3 Base-2 Nonrestoring Division Algorithm. 6.2.4 SRT Radix-2 Division. 6.2.5 SRT Radix-2 Division with Stored-Carry Encoding. 6.2.6 P-D Diagram. 6.2.7 SRT-4 Division. 6.2.8 Base-B Nonrestoring Division Algorithm. 6.3 Convergence (Functional Iteration) Algorithms. 6.3.1 Introduction. 6.3.2 Newton-Raphson Iteration Technique. 6.3.3 MacLaurin Expansion-Goldschmidt's Algorithm. 6.4 Bibliography. 7. Other Arithmetic Operations. 7.1 Base Conversion. 7.2 Residue Number System Conversion. 7.2.1 Introduction. 7.2.2 Base-B to RNS Conversion. 7.2.3 RNS to Base-B Conversion. 7.3 Logarithmic, Exponential, and Trigonometric Functions. 7.3.1 Taylor-MacLaurin Series. 7.3.2 Polynomial Approximation. 7.3.3 Logarithm and Exponential Functions Approximation by Convergence Methods. 7.3.3.1 Logarithm Function Approximation by Multiplicative Normalization. 7.3.3.2 Exponential Function Approximation by Additive Normalization. 7.3.4 Trigonometric Functions-CORDIC Algorithms. 7.4 Square Rooting. 7.4.1 Digit Recurrence Algorithm-Base-B Integers. 7.4.2 Restoring Binary Shift-and-Subtract Square Rooting Algorithm. 7.4.3 Nonrestoring Binary Add-and-Subtract Square Rooting Algorithm. 7.4.4 Convergence Method-Newton-Raphson. 7.5 Bibliography. 8. Finite Field Operations. 8.1 Operations in Zm. 8.1.1 Addition. 8.1.2 Subtraction. 8.1.3 Multiplication. 8.1.3.1 Multiply and Reduce. 8.1.3.2 Modified Shift-and-Add Algorithm. 8.1.3.3 Montgomery Multiplication. 8.1.3.4 Specific Ring. 8.1.4 Exponentiation. 8.2 Operations in GF(p). 8.3 Operations in Zp[x]/f (x). 8.3.1 Addition and Subtraction. 8.3.2 Multiplication. 8.4 Operations in GF(pn). 8.5 Bibliography. Appendix 8.1 Computation of fki. 9 Hardware Platforms. 9.1 Design Methods for Electronic Systems. 9.1.1 Basic Blocks of Integrated Systems. 9.1.2 Recurring Topics in Electronic Design. 9.1.2.1 Design Challenge: Optimizing Design Metrics. 9.1.2.2 Cost in Integrated Circuits. 9.1.2.3 Moore's Law. 9.1.2.4 Time-to-Market. 9.1.2.5 Performance Metric. 9.1.2.6 The Power Dimension. 9.2 Instruction Set Processors. 9.2.1 Microprocessors. 9.2.2 Microcontrollers. 9.2.3 Embedded Processors Everywhere. 9.2.4 Digital Signal Processors. 9.2.5 Application-Specific Instruction Set Processors. 9.2.6 Programming Instruction Set Processors. 9.3 ASIC Designs. 9.3.1 Full-Custom ASIC. 9.3.2 Semicustom ASIC. 9.3.2.1 Gate-Array ASIC. 9.3.2.2 Standard-Cell-Based ASIC. 9.3.3 Design Flow in ASIC. 9.4 Programmable Logic. 9.4.1 Programmable Logic Devices (PLDs). 9.4.2 Field Programmable Gate Array (FPGA). 9.4.2.1 Why FPGA? A Short Historical Survey. 9.4.2.2 Basic FPGA Concepts. 9.4.3 XilinxTM Specifics. 9.4.3.1 Configurable Logic Blocks (CLBs). 9.4.3.2 Input/Output Blocks (IOBs). 9.4.3.3 RAM Blocks. 9.4.3.4 Programmable Routing. 9.4.3.5 Arithmetic Resources in Xilinx FPGAs. 9.4.4 FPGA Generic Design Flow. 9.5 Hardware Description Languages (HDLs). 9.5.1 Today's and Tomorrow's HDLs. 9.6 Further Readings. 9.7 Bibliography. 10. Circuit Synthesis: General Principles. 10.1 Resources. 10.2 Precedence Relation and Scheduling. 10.3 Pipeline. 10.4 Self-Timed Circuits. 10.5 Bibliography. 11 Adders and Subtractors. 11.1 Natural Numbers. 11.1.1 Basic Adder (Ripple-Carry Adder). 11.1.2 Carry-Chain Adder. 11.1.3 Carry-Skip Adder. 11.1.4 Optimization of Carry-Skip Adders. 11.1.5 Base-Bs Adder. 11.1.6 Carry-Select Adder. 11.1.7 Optimization of Carry-Select Adders. 11.1.8 Carry-Lookahead Adders (CLAs). 11.1.9 Prefix Adders. 11.1.10 FPGA Implementation of Adders. 11.1.10.1 Carry-Chain Adders. 11.1.10.2 Carry-Skip Adders. 11.1.10.3 Experimental Results. 11.1.11 Long-Operand Adders. 11.1.12 Multioperand Adders. 11.1.12.1 Sequential Multioperand Adders. 11.1.12.2 Combinational Multioperand Adders. 11.1.12.3 Carry-Save Adders. 11.1.12.4 Parallel Counters. 11.1.13 Subtractors and Adder-Subtractors. 11.1.14 Termination Detection. 11.1.15 FPGA Implementation of the Termination Detection. 11.2 Integers. 11.2.1 B's Complement Adders and Subtractors. 11.2.2 Excess-E Adders and Subtractors. 11.2.3 Sign-Magnitude Adders and Subtractors. 11.3 Bibliography. 12 Multipliers. 12.1 Natural Numbers. 12.1.1 Basic Multiplier. 12.1.2 Sequential Multipliers. 12.1.3 Cellular Multiplier Arrays. 12.1.3.1 Ripple-Carry Multiplier. 12.1.3.2 Carry-Save Multiplier. 12.1.3.3 Figures of Merit. 12.1.4 Multipliers Based on Dissymmetric Br Bs Cells. 12.1.5 Multipliers Based on Multioperand Adders. 12.1.6 Per Gelosia Multiplication Arrays. 12.1.6.1 Introduction. 12.1.6.2 Adding Tree for Base-B Partial Products. 12.1.7 FPGA Implementation of Multipliers. 12.2 Integers. 12.2.1 B's Complement Multipliers. 12.2.2 Booth Multipliers. 12.2.2.1 Booth-1 Multiplier. 12.2.2.2 Booth-2 Multiplier. 12.2.2.3 Signed-Digit Multiplier. 12.2.3 FPGA Implementation of the Booth-1 Multiplier. 12.3 Bibliography. 13. Dividers. 13.1 Natural Numbers. 13.2 Integers. 13.2.1 Base-2 Nonrestoring Divider. 13.2.2 Base-B Nonrestoring Divider. 13.2.3 SRT Dividers. 13.2.3.1 SRT-2 Divider. 13.2.3.2 SRT-2 Divider with Carry-Save Computation of the Remainder. 13.2.3.3 FPGA Implementation of the Carry-Save SRT-2 Divider. 13.2.4 SRT-4 Divider. 13.2.5 Convergence Dividers. 13.2.5.1 Newton-Raphson Divider. 13.2.5.2 Goldschmidt Divider. 13.2.5.3 Comparative Data Between Newton-Raphson (NR) and Goldschmidt (G) Implementations. 13.3 Bibliography. 14 Other Arithmetic Operators. 14.1 Base Conversion. 14.1.1 General Base Conversion. 14.1.2 BCD to Binary Converter. 14.1.2.1 Nonrestoring 2p Subtracting Implementation. 14.1.2.2 Shift-and-Add BCD to Binary Converter. 14.1.3 Binary to BCD Converter. 14.1.4 Base-B to RNS Converter. 14.1.5 CRT RNS to Base-B Converter. 14.1.6 RNS to Mixed-Radix System Converter. 14.2 Polynomial Computation Circuits. 14.3 Logarithm Operator. 14.4 Exponential Operator. 14.5 Sine and Cosine Operators. 14.6 Square Rooters. 14.6.1 Restoring Shift-and-Subtract Square Rooter (Naturals). 14.6.2 Nonrestoring Shift-and-Subtract Square Rooter (Naturals). 14.6.3 Newton-Raphson Square Rooter (Naturals). 14.7 Bibliography. 15. Circuits for Finite Field Operations. 15.1 Operations in Zm. 15.1.1 Adders and Subtractors. 15.1.2 Multiplication. 15.1.2.1 Multiply and Reduce. 15.1.2.2 Shift and Add. 15.1.2.3 Montgomery Multiplication. 15.1.2.4 Modulo (Bk2c) Reduction. 15.1.2.5 Exponentiation. 15.2 Inversion in GF(p). 15.3 Operations in Zp[x]/f (x). 15.4 Inversion in GF(pn). 15.5 Bibliography. 16. Floating-Point Unit. 16.1 Floating-Point System Definition. 16.2 Arithmetic Operations. 16.2.1 Addition of Positive Numbers. 16.2.2 Difference of Positive Numbers. 16.2.3 Addition and Subtraction. 16.2.4 Multiplication. 16.2.5 Division. 16.2.6 Square Root. 16.3 Rounding Schemes. 16.4 Guard Digits. 16.5 Adder-Subtractor. 16.5.1 Alignment. 16.5.2 Additions. 16.5.3 Normalization. 16.5.4 Rounding. 16.6 Multiplier. 16.7 Divider. 16.8 Square Root. 16.9 Comments. 16.10 Bibliography. Index.

Synthesis of Arithmetic Circuits: FPGA, ASIC and Embedded Systems

Goldschmidt's Algorithms for division and square root are often characterized as being useful for hardware implementation, and lacking self-correction. A reexam- ination of these algorithms show that there are good software counterparts that retain the speed advantage of Goldschmidt's Algorithm over the Newton-Raphson iteration. A final step is needed, however, to get the last bit rounded correctly.

https://www.informatik.uni-trier.de/Reports/TR-08-2004/rnc6_12_markstein.pdf

Software Division and Square Root Using Goldschmidt's Algorithms

We propose optimizations for the IEEE floating-point fused multiply-add operation by considering multiple exclusive parallel computation paths in the implementation. For the proposed design we can show a significant performance improvement over conventional implementations. Considering a variable latency implementation allows for further reduction of the average latency.

Multiple path IEEE floating-point fused multiply-add

Back in the 60's Goldschmidt presented a variation of Newton-Raphson iterations for division that is well suited for pipelining. The problem in using Goldschmidt's division algorithm is to present an error analysis that enables one to save hardware by using just the right amount of precision for intermediate calculations while still providing correct rounding. Previous implementations relied on combining formal proof methods (that span thousands of lines) with millions of test vectors. These techniques yield correct designs but the analysis is hard to follow and is not quite tight. We present a simple parametric error analysis of Goldschmidt's division algorithm. This analysis sheds more light on the effect of the different parameters on the error. In addition, we derive closed error formulae that allow to determine optimal parameter choices in four practical settings. We apply our analysis to show that a few bits of precision can be saved in the floating-point division (FP-DIV) microarchitecture of the AMD-K7/spl trade/ microprocessor. These reductions in precision apply to the initial approximation and to the lengths of the multiplicands in the multiplier. When translated to cost, the reductions reflect a savings of 10.6% in the overall cost of the FP-DIV microarchitecture.

/pdf/a-parametric-error-analysis-of-goldschmidt-s-division-4uz8pvtwc0.pdf

Warren E. Ferguson

Papers

A parametric error analysis of Goldschmidt's division algorithm

A parametric error analysis of Goldschmidt's division algorithm