The DeSyRe project builds on-demand adaptive and reliable Systems-on-Chips (SoCs). As fabrication technology scales down, chips are becoming less reliable, thereby incurring increased power and performance costs for fault tolerance. To make matters worse, power density is becoming a significant limiting factor in SoC design, in general. In the face of such changes in the technological landscape, current solutions for fault tolerance are expected to introduce excessive overheads in future systems. Moreover, attempting to design and manufacture a totally defect-/fault-free system, would impact heavily, even prohibitively, the design, manufacturing, and testing costs, as well as the system performance and power consumption. In this context, DeSyRe delivers a new generation of systems that are reliable by design at well-balanced power, performance, and design costs. In our attempt to reduce the overheads of fault-tolerance, only a small fraction of the chip is built to be fault-free. This fault-free part is then employed to manage the remaining fault-prone resources of the SoC. The DeSyRe framework is applied to two medical systems with high safety requirements (measured using the IEC 61508 functional safety standard) and tight power and performance constraints.

In this article, new heuristic-search methods and algorithms are presented for enabling highly efficient and adaptive, defect-tolerant multiprocessor arrays. We consider systems where a homogeneous multiprocessor array lies on top of reconfigurable interconnects which allow the pipeline stages of the processors to be connected in all possible configurations. Considering the multiprocessor array partitioned in substitutable units at the granularity of pipeline stages, we employ a variety of heuristic-search methods and algorithms to isolate and replace defective units. The proposed heuristics are designed for off-line execution and aim at minimizing the performance overhead necessarily introduced to the array by the interconnects' latency. An empirical evaluation of the designed algorithms is then carried out, in order to assess the targeted problem and the efficacy of our approach. Our findings indicate this to be a NP-complete computational problem, however, our heuristic-search methods can achieve, for the problem sizes we exhaustively searched, 100% accuracy in finding the optimal solution among 1019 possible candidates within 2.5 seconds. Alternatively, they can provide near-optimal solutions at an accuracy which consistently exceeds 70% (compared to the optimal solution) in only 10-4 seconds.

The DeSyRe project builds on-demand adaptive and reliable Systems-on-Chips (SoCs). As fabrication technology scales down, chips are becoming less reliable, thereby incurring increased power and performance costs for fault tolerance. To make matters worse, power density is becoming a significant limiting factor in SoC design, in general. In the face of such changes in the technological landscape, current solutions for fault tolerance are expected to introduce excessive overheads in future systems. Moreover, attempting to design and manufacture a totally defect-/fault-free system, would impact heavily, even prohibitively, the design, manufacturing, and testing costs, as well as the system performance and power consumption. In this context, DeSyRe will deliver a new generation of systems that are reliable by design at well-balanced power, performance, and design costs.

In this paper, we describe the design and implementation of a new fault-tolerant RISC-processor architecture suitable for a design framework targeting biomedical implants. The design targets both soft and hard faults and is original in efficiently combining as well as enhancing classic fault-tolerance techniques. The proposed architecture allows run-time tradeoffs between performance and fault tolerance by means of instruction-level configurability. The system design is synthesized for UMC 90nm CMOS standard-process and is evaluated in terms of fault coverage, area, average power consumption, total energy consumption and performance for various duplication policies and test-sequence schedules. It is shown that area and power overheads of approximately 25% and 32%, respectively, are required to implement our techniques on the baseline processor. The major overheads of the proposed architecture are performance (up to 107%) and energy consumption (up to 157%). It is observed that the average power consumption is often reduced when a higher degree of fault tolerance is targeted. It is shown that test sequences can effectively be scheduled during the available program stalls and that nearly all soft faults are tolerated by using instruction duplication. The main advantages of the proposed architecture are the high portability of the introduced architecture-level fault-tolerance techniques, the flexibility in trading processor overheads for required fault-tolerance degree as well as affordable area and power consumption overheads.