An Efficient Algorithm for Spare Allocation Problems

Abstract

The spare allocation problem in redundant RAM is to replace faulty rows/columns of memory cells with spare rows/columns. To solve the problem, comparison-based search tree structures were used in traditional exact algorithms. These algorithms are not efficient for large problems because significant amounts of data have to be retained and copied in order to generate new partial solutions. Many data may need to be compared for the removal of each redundant partial solution. To overcome these drawbacks, an efficient algorithm is proposed in this paper. The algorithm transforms a spare allocation problem into Boolean functions, and the renowned BDD is used to manipulate them. Experimental results indicate that the proposed algorithm is very efficient in terms of speed and memory requirements. It may also be useful for problems which can be modeled as constraint bipartite vertex cover problems.

A Fast Built-in Redundancy Analysis for Memories With Optimal Repair Rate Using a Line-Based Search Tree

“…To achieve optimal repair rate for a memory block using 2-D spare architecture with a line replacement policy, three basic observations can be made for spare assignment during the RA. Some of these observations were also introduced in many previous studies [1], [6], [7], [12], [14]. The three observations are as follow.…”

“…1(a). Single fault and must-repair faulty lines were introduced in previous studies [1], [6], [7], [14], [16]. Sparse faulty line faults were handled like other faults, but identification of a sparse faulty line is very important for the proposed BIRA.…”

“…However, the repair rates of both algorithms are not optimal because of their excessive omission of faulty information. There has been much research that has attempted to solve the spare allocation problem for reconfigurable memory arrays based on the branch-and-bound (B&B) algorithm [6], [7], [15]- [18]. The B&B algorithm is a simple, fault-driven approach.…”

Optimal cutting design and analysis for fault-tolerant scheme of redundant memories

“…A solution-finding algorithm (modification of most-repair algorithm) There are a lot of algorithms for solving the spare allocation problem Fuchs 1986, 1992;Hasan and Liu 1988;Lombardi and Huang 1988;Huang, Shen, and Lombardi 1990;Funabiki and Takefuji, 1991;Haddad, Dahura, and Sharma 1991;Hadas and Liu 1991;Kuo and Chen 1991;Che and Koren 1992;Shi and Fuchs 1992;Blough and Pelc 1993;Chen and Upadhyaya 1993;Blough 1996;Low and Leong 1996a, b;Bhavsar 1999;Chor 2000;Fernau and Niedermeier 2001;Chen and Kanj 2003;Huang, Wu, Li, and Wu 2003;Lin, Chou, Yeh, Chen, and Kuo 2004a;Lin, Yeh, Chen, and Kuo 2004b;Liang et al 2005;Yu et al 2005;Lin et al 2006). These algorithms can be categorised into two types (Kuo and Fuchs 1992): exact algorithms and heuristic algorithms.…”

“…The second step is to test redundant columns or rows to replace the faulty ones. The second step is called the spare allocation problem Fuchs 1986, 1992;Che and Koren 1992;Chen and Upadhyaya 1993;Liang, Ho, and Cheng 2005;Yu et al 2005;Lin, Yeh, and Kuo 2006) or the memory reconfiguration problem (Kuo and Fuchs 1986;Yu et al 2005), in which we search for a solution to replace all faulty cells with redundant cells. As the trend of chip size becomes smaller for higher speeds, the quantity of spare lines should be restricted.…”