文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > An Efficient Constraint-Based Test Selection Method for Embedded Processor Cores
侵权投诉

An Efficient Constraint-Based Test Selection Method for Embedded Processor Cores

Abstract. Software-based self-test has been proposed as a low-cost strategy for on-line periodic testing of embedded processors. However, for complex processors, the number of possible combinations of self-test routines may be huge, and the selection of a

propose a software implementation of a Linear Feedback Shift Register (LFSR) for each target component, taking into consideration the constraints imposed by the processor instruction set. Deterministic methodologies have been proposed by Paschalis and Gizopoulos [4]. If gate-level information of the processor core is available, deterministic patterns can be generated by an Automatic Test Pattern Generator (ATPG). The application of the generated patterns can be implemented in two ways: with test patterns fetched from the data memory or applied as immediate operands. Paschalis et. al. [5] and Krainitis et. al. [6] improve this approach by proposing high level structural self-test methodologies. The self-test routines generate test patterns using a regular function that is independent of the gate-level implementation of the components, but are based on the knowledge of the processor's Instruction Set Architecture (ISA) and its RTL description.

The main concern of the mentioned works is the optimization of the self-test routines generation, and the resulting fault coverage of the target components. However, those methods do not address other important aspects of the resulting test programs, such as memories size, power consumption during test, and resulting test time. Such costs may have a non-negligible impact in safety-critical real-time embedded systems. Hence, the definition of a suitable self-test program, considering the system and test constraints, is an additional challenge for the system designer. Moreover, as the number of possible test programs (combinations of self-test routines) is very large, an efficient test program generation method is crucial for the system design optimization.

In this paper, we firstly show that the selection of a set of self-test routines to compose a test program must consider other system aspects, in addition to the test aspects. We further show that the generation of the test program requires an efficient search method when complex processors are considered. Then we propose a heuristic to reduce the searching space and test generation time. Our method ensures that the chosen test program meets the system requirements while minimizing a test cost function defined by the designer. The proposed solution is supported by experimental results performed for the test of a Java processor, where memory size, energy consumption, and test period constraints are considered.

The paper is organized as follows: Section 2 details the complexity of combining self-test routines into a test program aiming at real-time embedded systems. In Section 3, the test program selection problem is defined and the constraint-based search method is presented. Section 4 introduces the target platform, the real-time application, and the self-test library used in the experiments. Section 5 provides experimental results while conclusions and future directions are pointed out in Section 6.

2. The Problem of Test Program Definition

Let us consider an embedded processor with c target components. As mentioned in Section 1, a number of self-test routines can be defined for each target component (pseudo-random, deterministic, etc). Thus, let us consider also a library of self-test routines composed of s routines for each component. For instance, the Java processor,

上一页第2页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email