Test Data Compression Method Based On Adjacent Run Length Feature

From the daily life of electronic products,to the recent hot Chat GPT,are inseparable from the shadow of the chip.In the modern industrial system,chip has a pivotal position,chip manufacturing competition is increasingly fierce.In the past,the emergence of System on Chip(So C)has greatly improved the integration of chips,but also brought many problems to integrated circuit testing,one of which is the huge scale of test data.A large amount of test data inevitably increases the test time.When the size of test data exceeds the memory limit of the automatic test device,the test cannot be completed properly.Lossless compression of test data is one of the effective methods to solve this problem.Therefore,this thesis mainly studies So C test data compression,introduces the basic principle and technology of integrated circuit test and several classic test data compression methods,and puts forward two new test data compression methods aiming at the problem of large test data scale.The main contents are as follows:(1)When traditional test data compression schemes encode for the difference between adjacent runs,they use run type flags and positive/negative difference flags.The use of these flags increases the length of the codeword after compression.To address this issue,this thesis proposes Test Data Compression Scheme for Adjacent Run-length Incremental Coding.First,get the difference between the current run and the previous run.Then,during the encoding process,compare the length of the codeword after encoding with the encoding table and the difference between adjacent runs.When the difference is a positive number and 1 is added,the obtained value is still not greater than the length of the code word corresponding to the coding table,then the special code consisting of 0 and the length of the difference plus 1 is used to represent the current run.Otherwise,use the codeword corresponding to the run length on the encoding table to represent the current run.During decompression,you only need to get the first character to select the corresponding decompression method.During decompression,you only need to obtain the first character to select the corresponding decompression method.Experimental results show that the method is feasible,and the average compression rate is62.99%.(2)Test Data Compression Scheme for Adjacent Run-length Incremental Coding only considers individual differences when using the difference in adjacent run lengths and does not take advantage of the correlation between adjacent run length differences in the entire test data.Based on this,this thesis proposes an adaptive test data compression scheme for relative encoding of adjacent run lengths.Firstly,after don’t-care bits are filled,the difference between adjacent run lengths is counted to find out the difference with the highest frequency as a special value.Then,don’t-care bits are filled on the original test data again to improve the frequency of occurrence of special values.As the code word corresponding to the special value,it is coded at last.When the value of the last run length minus the previous run length is a special value,its corresponding code word is coded;otherwise,it is coded according to the coding table.The experimental results show that this method has good compression effect,and the average compression rate can reach 67.61%.In this thesis,two test data compression schemes are proposed by analyzing the relationship between adjacent run lengths.In the compression process,the relative length encoding is dynamically selected to replace the absolute length encoding to improve the compression efficiency.During decompression,the compression method can be automatically identified according to the character of the code word,which does not add much extra overhead.