Asynchronous design is a promising alternative for emerging technologies facing extreme parameter variation, severe timing/clock skew and power consumption issues. However, the complexity in design and test is one of the major obstacles for the widespread use of asynchronous circuits in digital design. Circuits utilizing templates are often implemented to mitigate the design complexity of an asynchronous circuit. One of the most commonly used pre-designed templates is the so-called Pre-Charged Full Buffer (PCFB), however, when testing template-based designs, most of the faults are undetectable by using conventional methods. In this paper, the PCFB template is designed such that faults always result in three scenarios (deadlock, token generation and dropping) for ease of detection, its operation and the new design of the hardware required for testability are described in detail. It is analytically shown that under a model that includes all single stuck-at faults, the new template (as characterized by novel features in its design) accomplishes ease of testability as well as online detection and fail-safe circuit operation. 100% coverage of single faults is accomplished. Simulation results for benchmark circuits are provided.
[Show abstract][Hide abstract] ABSTRACT: The design of an asynchronous clone of a MIPS R3000 microprocessor is presented. In 0:6m CMOS, we expect performance close to 280 MIPS, for a power consumption of 7 W. The paper describes the structure of a high-performance asynchronous pipeline, in particular precise exceptions, pipelined caches, arithmetic, and registers, and the circuit techniques developed to achieve high throughput. 1 Introduction This paper describes the architectural algorithms and circuit techniques used in the design of an asynchronous MIPS R3000 microprocessor. The project has two main goals. First, we are investigating issues in asynchronous processor architecture that we have not tackled in the Caltech Asynchronous Microprocessor : caches, precise exceptions, register bypassing, branch-delay slot and branch prediction. Secondly, we are developing new techniques for asynchronous digital VLSI---based on very fine pipelining---that can meet high throughput requirements without sacrificing the low-power ad...
[Show abstract][Hide abstract] ABSTRACT: A new technique to test asynchronous circuits obtained by direct mapping technique from I-safe Petri nets is proposed. Low-level physical faults in the cells implementing Petri net places are analysed and mapped into high-level specification, a Petri net. A "pseudo clock" is used to handle hazards and activate faults which exhibit themselves only under particular arrangements. Asynchronous circuit obtained by direct mapping technique can be made 100% testable for stuck-at-faults by implementing testability features. An algorithm to insert testability features and generate test sequences is presented using a benchmark.
VLSI Design, 2005. 18th International Conference on; 02/2005
[Show abstract][Hide abstract] ABSTRACT: In this paper, we relate the property of semi-modularity to the testability of speed-independent circuits. We show that, under the pure chaos delay model, live speed-independent circuits that are strongly connected and composed of ANDs, ORs, and C-elements (with a possible inverter on each gate input) can be decomposed into a set of semi-modular circuits and therefore fully testable for certain classes of output stuck-at-faults (OSAFs). In addition, we show that a subclass of such speed-independent circuits are fully testable for all multiple OSAFs and for certain input SAFs (ISAFs) as well. Specifically, we qualify the kind of SAFs that are detectable during the normal operation of speed-independent circuits regardless of individual gate delays. These results demonstrate the inherent self-checking property of speed-independent circuits and indicate the kind of faults for which speed-independent circuits can be easily tested. We also present a CAD tool that checks the testability of a speed-independent circuit.
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.