Browse Prior Art Database

System and Method for detecting syndrome based ECC in a chip design RTL

IP.com Disclosure Number: IPCOM000237354D
Publication Date: 2014-Jun-15
Document File: 2 page(s) / 31K

Publishing Venue

The IP.com Prior Art Database

Abstract

Automatically identifying syndrome based ECC logic in a chip design RTL based merely on the netlist. Given an internal or external design it enables assessment of how protected and safe the design is for an external or internal examiner; it also enables debugging for the designer. It can also serve as an basis for detecting which part of the design is ECC protected.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

System and Method for detecting syndrome based ECC in a chip design RTL

Syndrome based ECC detection Our ultimate objective is to identify automatically the generation of an error correction code (ECC) in a netlist. As a first step we identify syndrome based ECC Shubu Mukherjee: Architecture design for soft errors, chapter
5.2.3, which is the method in which the majority of the ECC is done in IBM.

Methodology, taken fromi: A parity check matrix, P, consists of r rowsand n columns

where k is the number of data bits and r is the number of check bits (n=r+k) . Each column in the parity check matrix corresponds to either a data bit d1,…,dk or a code bit c1,…,cr . The positions of the 1s in a row in the parity check matrix indicate the bit position which is involved in the parity check equation for this row. Define the corresponding code word , E as E=c1...crd1...dk. The syndrome, S, is defined as S=P\dot ET. Finally, let the check bits word, C, be C=c1…cr.

    The challenge of our algorithm is to recognize syndrome calculations and check bits word calculation in the netlist without having any prior knowledge: given a vector on chip (e.g,. X(0 to 10), YY(0 to 100)), identifies if it a calculation as above and if so of

what type and what is the underlying parity check matrix.

Our algorithm is able to:
1. Find all vectors that are suspected of being syndrome or check bits word calculation.

2. Identify each such vector as a syndrome vector or a check bits vector.

3. In both cases, identify the underlying parity check matrix.

Novelties of the algorithm: 1. Look for syndrome/check bits vector only in vectors of the netlist, e,g, X(0 to 7). For any practical purpose this is suffice, and it is obvious that this reduction of the search space is necessary.

2. The main observation is that each bit of the check bits/syndrome is a parity (xor) calculation of some support set. Finding a vector which all of its bits are parity calculations is an indication of that vector being a check bits/syndrome vector. The union of the supports of the parity calculations is the support set. Our technology enables us to recognize such parity calculations in the netlist. The challenge here is to find the "correct" support set,...