Joint Crosstalk-Avoidance and Error-Correction Coding for Parallel Data Buses

• Published in: IEEE transactions on information theory Vol. 65; no. 3; pp. 1626 - 1638
• Main Authors: Niesen, Urs, Kudekar, Shrinivas
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-03-2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Avoidance; Codes; Coding; Coupling; Couplings; Crosstalk; Crosstalk-avoidance coding; Data buses; Decoding; Density; density evolution; Encoding; Error correction; Error correction codes; error-correction coding; Evolution; Integrated circuits; irregular repeat-accumulate codes; Iterative decoding; on-chip communication; Wires
• ISSN: 0018-9448; 1557-9654
• DOI: 10.1109/TIT.2018.2859405

Communication in integrated circuits faces two major impediments: inter-wire capacitive coupling and noise. Coding can be used to address both these problems. So-called crosstalk-avoidance codes mitigate capacitive coupling, and traditional error-correction codes introduce resilience against channel errors. Unfortunately, crosstalk-avoidance and error-correction codes cannot be combined in a straightforward manner. On the one hand, crosstalk-avoidance encoding followed by error-correction encoding destroys the crosstalk-avoidance property. On the other hand, error-correction encoding followed by crosstalk-avoidance encoding causes the crosstalk-avoidance decoder to fail in the presence of errors. Existing approaches circumvent this difficulty by using additional bus wires to protect the parities generated from the output of the error-correction encoder, and are therefore inefficient. In this paper, we propose a novel joint crosstalk-avoidance and error-correction coding and decoding scheme that provides higher bus transmission rates compared with existing approaches. Our joint approach carefully embeds the parities such that the crosstalk-avoidance property is preserved. We analyze the rate and minimum distance of the proposed scheme. We also provide a density evolution analysis and predict iterative decoding thresholds for reliable communication under random bus erasures. This density evolution analysis is nonstandard, since the crosstalk-avoidance constraints are inherently nonlinear.