In this paper, we propose a new approach to implement a reliable softcore processor on SRAM-based FPGAs, which can mitigate radiation-induced temporary faults (single-event upsets (SEUs)) at moderate cost. A new Enhanced Lockstep scheme built using a pair of MicroBlaze cores is proposed and implemented on Xilinx Virtex-5 FPGA. Unlike the basic lockstep scheme, ours allows to detect and eliminate its internal temporary configuration upsets without interrupting normal functioning. Faults are detected and eliminated using a Configuration Engine built on the basis of the PicoBlaze core which, to avoid a single point of failure, is implemented as fault-tolerant using triple modular redundancy (TMR). A softcore processor can recover from configuration upsets through partial reconfiguration combined with roll-forward recovery. SEUs affecting logic which are significantly less likely than those affecting configuration are handled by checkpointing and rollback. Finally, to handle permanent faults, the tiling technique is also proposed. The new Enhanced Lockstep scheme requires significantly shorter error recovery time compared to conventional lockstep scheme and uses significantly smaller number of slices compared to known TMR-based design (although at the cost of longer error recovery time). The efficiency of the proposed approach was validated through fault injection experiments.
Synchronization is a challenging problem for wireless nodes, especially for applications demanding good synchronization accuracy over wide areas. In that case, the GPS is a valuable solution as the nodes can independently synchronize to UTC. However, the energy consumption of a GPS receiver (over 100 mW when switched on) is not sustainable on a wireless node. Therefore, in this work, we developed a synchronization scheme based on periodic extinctions of the GPS receiver. The goal is to study the GPS power switching effect on the synchronization accuracy. To do so, a node with dedicated timestamping hardware was designed. Two clock models were compared to predict the node time when the GPS is off and the impact of a Kalman filter, to remove the GPS noise, was evaluated. From experimental data, we show that the choice of the clock model depends on the accuracy needed and that the Kalman filter improves the estimation of the clock frequency for both models. In our design, the GPS can be off from 60% up to 95% of the time for mean synchronization errors of 20 ns to 420 ns, respectively. This work demonstrates that GPS power switching is an efficient solution to reduce energy costs while maintaining a high synchronization accuracy.
1 January 2008 in EURASIP Journal on Embedded Systems
Flexibility becomes a major concern for the development of multimedia and mobile communication systems, as well as classical high-performance and low-energy consumption constraints. The use of general-purpose processors solves flexibility problems but fails to cope with the increasing demand for energy efficiency. This paper presents the DART architecture based on the functional-level reconfiguration paradigm which allows a significant improvement in energy efficiency. DART is built around a hierarchical interconnection network allowing high flexibility while keeping the power overhead low. To enable specific optimizations, DART supports two modes of reconfiguration. The compilation framework is built using compilation and high-level synthesis techniques. A 3G mobile communication application has been implemented as a proof of concept. The energy distribution within the architecture and the physical implementation are also discussed. Finally, the VLSI design of a 0.13 m CMOS SoC implementing a specialized DART cluster is presented.
This paper presents the implementation of the error recovery scheme from temporary faults, applicable for datapaths of coarse-grained reconfigurable architectures. We have chosen the DART architecture as a vehicle to study various aspects related to implementation of the instruction retry in a complex highly parallel reconfigurable system. Synthesis results have confirmed the time, hardware, and power consumption efficiency of the proposed approach, which can be applied independently on the concurrent error detection scheme actually used.
This paper introduces adaptive techniques targeted for heterogeneous manycore architectures and introduces the FlexTiles platform, which consists of general purpose processors with some dedicated accelerators. The different components are based on low power DSP cores and an eFPGA on which dedicated IPs can be dynamically configured at run-time. These features enable a breakthrough in term of computing performance while improving the on-line adaptive capabilities brought from smart heuristics. Thus, we propose a virtualisation layer which provides a higher abstraction level to mask the underlying heterogeneity present in such architectures. Given the large variety of possible use cases that these platforms must support and the resulting workload variability, offline approaches are no longer sufficient because they do not allow coping with time changing workloads. The upcoming generation of applications include smart cameras, drones, and cognitive radio. In order to facilitate the architecture adaptation under different scenarios, we propose a programming model that considers both static and dynamic behaviors. This is associated with self adaptive strategies endowed by an operating system kernel that provides a set of functions that guarantee quality of service (QoS) by implementing runtime adaptive policies. Dynamic adaptation will be mainly used to reduce both overall power consumption and temperature and to ease the problem of decreasing yield and reliability that results from submicron CMOS scales.
Managing tasks and resources of reconfigurable system-on-chip is a complex problem which needs specific operating system (OS) functionalities. One of the most important is the task placement which must be done on-line when the application requires flexibility. To ensure an efficient task placement within the reconfigurable resource, OS services must consider the heterogeneity of the reconfigurable resource. While most publications model the reconfigurable resource as homogeneous area, modern reconfigurable circuits are clearly heterogeneous, i.e. there are based on rectangular grid containing logic blocks but also other blocks such as memories, digital signal processing blocks or hard processor cores. In this paper, we tackle the problem of task placement within a reconfigurable resource and we consider a heterogeneous reconfigurable area. Our solution is based on a neural network structure specifically designed to optimize the task placement problem. Our proposition is based on the knowledge of task instantiations within the reconfigurable resource. Compared with other methods, our proposal provides better results in terms of task rejection.