独立于测试数据的字典编码方法

Dictionary encoding method based on independent of test data

为减少测试数据量,提出一种新的独立于测试数据的字典编码方法。首先设计基于图论寻找团的启发式算法,合并相容扫描链,压缩测试数据宽度,然后利用最简分数对应浮点数的二进制形式生成字典,使用字典编码进一步压缩测试数据长度。该编码方法无需填充测试集中的无关位且字典的生成独立于测试数据,可以预先生成字典,缩短了测试时间;解压时无需存储字典,通过简单计算即可还原测试数据,减小了硬件开销。实验表明,该方法在ISCAS 89标准电路Mintest集上的平均压缩率达到65.57%,能够有效地压缩测试数据。

A new dictionary encoding method based on independent of test data is presented to reduce test data vol-ume.First,a heuristic algorithm is proposed,which can find clique based on graph theory to compress test data width by merging compatible scan chains.Then,the dictionary is generated by floating-point number’s binary form of fraction in lowest term,which is used to compress test data length further by dictionary encoding.Don’t care bits needn’t to be filled and the dictionary can be pre-generated because it is independent of test data.So the scheme can shorten test time.There is no requirement to store dictionary when decompressing and test data can be restored by a simple calculation.So the scheme can also reduce the hardware cost.The experiment shows that the average compression ratio of the proposed scheme in ISCAS 89 benchmark circuits can reach 65 .57%.The scheme can efficiently compress test data.