Experiences with implementing common mathematical operations using field programmable gate arrays

Many computationally intensive scientific applications involve repetitive floating point operations other than addition and multiplication which may present a significant performance bottleneck due to the relatively large latency or low throughput involved in executing such arithmetic primitives on commod- ity processors. A promising alternative is to execute such primitives on Field Programmable Gate Array (FPGA) hardware acting as an application-specific custom co-processor in a high performance reconfig- urable computing platform. The use of FPGAs can provide advantages such as fine-grain parallelism but issues relating to code development in a hardware description language and efficient data transfer to and from the FPGA chip can present significant application development challenges. In this paper, we discuss our practical experiences in developing a selection of floating point hardware designs to be implemented using FPGAs. Our designs include some basic mathemati cal library functions which can be implemented for user defined precisions suitable for novel applications requiring non-standard floating point represen- tation. We discuss the details of our designs along with results from performance and accuracy analysis tests.

Solving tri-diagonal linear systems using field programmable gate arrays

In this paper, we present the outcomes of a project on the exploration of the use of Field Programmable Gate Arrays(FPGAs) as co-processors for scientific computation. We designed a custom circuit for the pipelined solving of multiple tri-diagonal linear systems. The design is well suited for applications that require many independent tri diagonal system solves, such as finite difference methods for solving PDEs or applications utilising cubic spline interpolation. The selected solver algorithm was the Tri Diagonal Matrix Algorithm (TDMA or Thomas Algorithm). Our solver supports user specified precision thought the use of a custom floating point VHDL library supporting addition, subtraction, multiplication and division. The variable precision TDMA solver was tested for correctness in simulation mode. The TDMA pipeline was tested successfully in hardware using a simplified solver model. The details of implementation, the limitations, and future work are also discussed.

Frame-rate stereopsis using non-parametric transforms and programmable logic

A frame-rate stereo vision system, based on non-parametric matching metrics, is described. Traditional metrics, such as normalized cross-correlation, are expensive in terms of logic. Non-parametric measures require only simple, parallelizable, functions such as comparators, counters and exclusive-or, and are thus very well suited to implementation in reprogrammable logic.

Solving tri-diagonal linear systems using field programmable gate arrays

In this paper, we present the outcomes of a project on the exploration of the use of Field Programmable Gate Arrays (FPGAs) as co-processors for scientific computation. We designed a custom circuit for the pipelined solving of multiple tri-diagonal linear systems. The design is well suited for applications that require many independent tri-diagonal system solves, such as finite difference methods for solving PDEs or applications utilising cubic spline interpolation. The selected solver algorithm was the Tri-Diagonal Matrix Algorithm (TDMA or Thomas Algorithm). Our solver supports user specified precision thought the use of a custom floating point VHDL library supporting addition, subtraction, multiplication and division. The variable precision TDMA solver was tested for correctness in simulation mode. The TDMA pipeline was tested successfully in hardware using a simplified solver model. The details of implementation, the limitations, and future work are also discussed.

Energy efficient home with price sensitive stochastically programmable TCAs

Introduction of dynamic pricing in present retail market, considerably affects customers with an increased cost of energy consumption. Therefore, customers are enforced to control their loads according to price variation. This paper proposes a new technique of Home Energy Management, which helps customers to minimize their cost of energy consumption by appropriately controlling their loads. Thermostatically Controllable Appliances (TCAs) such as air conditioner and water heater are focused in this study, as they consume more than 50% of the total household energy consumption. The control process includes stochastic dynamic programming, which incorporated uncertainties in price and demand variation. It leads to an accurate selection of appliance settings. It is followed by a real time control of selected appliances with its optimal settings. Temperature set points of TCAs are adjusted based on price droop which is a reflection of actual cost of energy consumption. Customer satisfaction is maintained within limits using constraint optimization. It is showed that considerable energy savings is achieved.

Programmable digital signal averager.

A high speed digital signal averager with programmable features for the sampling period, for the number of channels and for the number of sweeps is described. The system implements a stable averaging algorithm (Deadroff and Trimble 1968) to provide a stable, calibrated display. The performance of the instrument has been evaluated for the reduction of random noise and for comb-filter action. Special uses of the instrument as a box-car integrator and as a transient recorder are also indicated.

A Design for Testability of Undetectable Crosspoint Faults in Programmable Logic Arrays

In this paper, the validity of'single fault assumption in deriving diagnostic test sets is examined with respect to crosspoint faults in programmable logic arrays (PLA's). The control input procedure developed here can be used to convert PLA's having undetectable crosspoint faults to crosspoint-irredundant PLA's for testing purposes. All crosspoints will be testable in crosspoint-irredundant PLA's. The control inputs are used as extra variables during testing. They are maintained at logic I during normal operation. A useful heuristic for obtaining a near-minimal number of control inputs is suggested. Expressions for calculating bounds on the number of control inputs have also been obtained.

Comments and Correspondence on a programmable digital signal averager

In a recent paper, Srinivasan et al (1980) have described a programmable digital signal averager with facility for programming the sampling period, number of channels and number of sweeps. We have examined this paper in some detail and find that some points need clarification.

An on-line algorithm for the location of cross point faults in programmable logic arrays

An on-line algorithm is developed for the location of single cross point faults in a PLA (FPLA). The main feature of the algorithm is the determination of a fault set corresponding to the response obtained for a failed test. For the apparently small number of faults in this set, all other tests are generated and a fault table is formed. Subsequently, an adaptive procedure is used to diagnose the fault. Functional equivalence test is carried out to determine the actual fault class if the adaptive testing results in a set of faults with identical tests. The large amount of computation time and storage required in the determination, a priori, of all the fault equivalence classes or in the construction of a fault dictionary are not needed here. A brief study of functional equivalence among the cross point faults is also made.

An On-Line Algorithm for the Location of Cross Point Faults in Programmable Logic Arrays

An on-line algorithm is developed for the location of single cross point faults in a PLA (FPLA). The main feature of the valgorithm is the determination of a fault set corresponding to the response obtained for a failed test. For the apparently small number of faults in this set, all other tests are generated and a fault table is formed. Subsequently, an adaptive procedure is used to diagnose the fault. Functional equivalence test is carried out to determine the actual fault class if the adaptive testing results in a set of faults with identical tests. The large amount of computation time and storage required in the determination, a priori, of all the fault equivalence classes or in the construction of a fault dictionary are not needed here. A brief study of functional equivalence among the cross point faults is also made.

Future switching technology: bacterial protein-based programmable all-optical switch for integrated optics applications?

We report on the bacterial protein-based all-optical switches which operate at low laser power, high speed and fulfil most of the requirements to be an ideal all-optical switch without any moving parts involved. This consists of conventional optical waveguides coated with bacteriorhodopsin films at switching locations. The principle of operation of the switch is based on the light-induced refractive index change of bacteriorhodopsin. This approach opens the possibility of realizing proteinbased all-optical switches for communication network, integrated optics and optical computers.

A microprocessor-controlled programmable pulse generator

The hardware and the software details of a user-friendly, simple, flexible and inexpensive pulse programmer using programmable counters interfaced to a microprocessor are described. The control of the various parameters that are required for NMR applications is implemented using the microprocessor. The basic hardware is extendable to other applications which require programmable pulse trains.

Foldable compatibility matrix for the folding of programmable logic arrays

This paper describes a switching theoretic algorithm for the folding of programmable logic arrays (PLA). The algorithm is valid for both column and row folding, although it has been presented considering only the simple column folding. The pairwise compatibility relations among all the pairs of the columns of the PLA are mapped into a square matrix, called the compatibility matrix of the PLA. A foldable compatibility matrix (FCM), a new concept introduced by the author, is then derived from the compatibility matrix. A new theorem called the folding theorem is then proved. The theorem states that the existence of an m by 2m FCM is both necessary and sufficient to fold 2m columns of the n column PLA (2m ≤ n). Once an FCM is obtained, the ordered pairs of foldable columns and the re-ordering of the rows are readily determined.

'A programmable digital signal averager' Reply

In a recent paper, Srinivasan et al (1980) have described a programmable digital signal averager with facility for programming the sampling period, number of channels and number of sweeps. We have examined this paper in some detail and find that some points need clarification.

A Simple Programmable 2-Pulse Generator for Pulsed NMR NQR Applications

Programmable pulse generator (PPG) circuits using programmable interval timer chips are normally based on a PC or a microprocessor. We describe here a simple low cost programmable two-pulse generator using Intel 8253s in a stand-alone mode, eliminating the need for a PC or a microprocessor, though our design also can be operated via a PC or a microprocessor.