Comprehensive online defect diagnosis in on-chip networks

Abstract

Abstract-We propose a comprehensive yet low-cost solution for online detection and diagnosis of permanent faults in on-chip networks. Using error syndrome collection and packet/flit-counting techniques, highresolution defect diagnosis is feasible in both datapath and control logic of the on-chip network without injecting any test traffic or incurring significant performance overhead.

Transaction-based online debug for NoC-based multiprocessor SoCs

“…In the diagnosis unit there are different comparators to compare data from all the possible pairs of switch input ports. A comprehensive defect diagnosis for NoCs is proposed in [22]. The approach uses an end-to-end error symptom collection mechanism [23] to localize datapath faults and a distributed counting and timeout-based technique to localize faulty control components [22].…”

“…A comprehensive defect diagnosis for NoCs is proposed in [22]. The approach uses an end-to-end error symptom collection mechanism [23] to localize datapath faults and a distributed counting and timeout-based technique to localize faulty control components [22]. The work in [24] diagnoses the NoC switch faults using hardware redundancy in each switch and a high level fault model.…”

Variable-width datapath for on-chip network static power reduction

“…Mechanisms to detect silicon defects during network operation have been proposed [15]. Faulty channels can be disabled which forces packets to use alternate and often longer paths [29,40].…”

Structural Software-Based Self-Test of Network-on-Chip

“…For functional testing of NoC, references [5][6][7] use data encoding in order to detect faults in the communication links and the switch datapath elements. Test repetition classifies the transient or permanent nature of the fault.…”

“…In [10], based on a system level fault model, an online fault detection for the NoC switches has been proposed. In [6], a fault detection and diagnosis mechanism targeting permanent faults is presented that uses error syndrome collection and packet/flit counting. The method uses error detecting codes in order to deal with data faults, and introduces dropped flits/packets, spurious flits/packets and misrouted packets to describe control faults in the switch.…”