A method for compressing test data based on Burrows-Wheeler transformation

The overall throughput of automatic test equipment (ATE) is affected by the download time of test data. An effective approach to the reduction of the download time is to compress test data before the download. A compression algorithm for test data should meet two requirements: lossless and simple decompression. In this paper, we propose a new test data compression method that aims to fully utilize the unique characteristics of test data compression. The key idea of the proposed method is to perform the Burrows-Wheeler transformation on the sequence of test patterns and then to apply run-length coding. Experimental results show that our compression method performs better than six other methods for compressing test data. The average compression ratio of the proposed method performed on 15 test data sets is 94.6, while that for the next best one, Gzip, is 65.0. The experimental results also show that our method indeed reduces the download time significantly, provided a dedicated hardware decompressor is employed