基于分组共享种子和位翻转的测试数据压缩方法

Test data compression based on grouping of share seeds and bit-flipping

文章提出了一种基于分组共享种子和位翻转的测试数据压缩方法。该方法根据测试集生成的特点,将确定位处于相同位置的测试向量分为一类,并将每类确定位不同处无关位化,合并得到每类的首模式,这就减少了首模式确定位的个数;同时每类中向量共享一个种子,减少了编码种子个数,提高种子的利用率,达到压缩测试数据的目的;解压时通过记录与LFSR重新播种展开序列确定位不同处的地址信息和翻转信号还原即可。实验表明,该方法与基于部分向量切分的LFSR重新播种方法和混合码相比,在压缩效率上具有明显的优势。

A new method of test data compression based on grouping of share seeds and bit-flipping is proposed. The test set is classified according to the rule that the special bits between the vectors are in the same position on the basis of characteristics generated by the test set. The bit in the position where the special bits are different in the same group is changed to the bit without position. Then the first pattern is gained by the combination. So the number of special bits is reduced. At the same time, as one group shares one seed, the number of the seeds is also reduced, and the efficiency of seeds is improved. In decompression, by the address and signal of flipping of the position that is different on the list using LFSR, the original test can be obtained. The experiment indicates that the scheme has prominent advantages compared with the hybrid code and the method which is LFSR reseeding based on syncopation of some test patterns.