1. 1/20 A Novel Technique for Input Vector Compression in System-on-Chip Testing Student: Chien Nan Lin Satyendra Biswas, Sunil Das, and Altaf Hossain,” Information Technology, 2008. ICIT '08. International Conference on ”, Bhubaneswar, pp. 53 - 58, 17-20 Dec. 2008.

2. 2/24 Outline Introduction Frame of Compression Technique Theoretical Background Proposed Technique Experimental Results

3. 3/24 Introduction In this paper, a new test vector compression method for VLSI circuit testing is presented. To reduce the on-chip: Storage area Testing time Simulation experiments on ISCAS 85 benchmark.

4. 4/24 Introduction

5. 5/24 Outline Introduction Frame of Compression Technique Theoretical Background Proposed Technique Experimental Results

6. 6/24 Frame of Compression Technique Original Test Vectors Block Matching Lzw Coding Compressed Test Vectors Low Frequency Data Sets High Frequency Data Sets Burrows-Wheeler Transformation +

7. 7/24 Output Outline Introduction Frame of Compression Technique Theoretical Background Proposed Technique Experimental Results

8. 8/24 Theoretical Background Original Test Vectors Block Matching Lzw Coding Compressed Test Vectors Low Frequency Data Sets High Frequency Data Sets Burrows-Wheeler Transformation + Frame

9. 9/24 Theoretical Background Burrows-Wheeler Transform The Burrows-Wheeler transformation algorithm is described in the following: Step 1:Create a list of possible rotation of string. Step 2:Let each rotation be one row in a large, sequare table. Step 3:Sort the rows of the alphabetically, treating each row as a string. Step 4:Return the last column of the table.

10. 10/24 Theoretical Background Burrows Wheeler Transform

11. 11/24 Theoretical Background Burrows-Wheeler Transform For example: Input All Rotations Sort the Rows Output ^BANANA@ @^BANANA A@^BANAN NA@^BANA ANA@^BAN NANA@^BA ANANA@^B BANANA@^ ANANA@^B ANA@^BAN A@^BANAN BANANA@^ NANA@^BA NA@^BANA ^ BANANA@ @ ^ BANANA BNN^AA@A (the red @ character indicates the 'EOF' pointer)EOF

12. 12/24 Theoretical Background Burrows-Wheeler Transform Compressing test data using run-length coding and Burrows-Wheeler transformation. For example: BNN^AA@A ─> 1B2N1^2A1@1A AAABBBBBBBBBAA ─> 3A9B2A A is 「 Run 」 3 is 「 Length 」

13. 13/24 Theoretical Background Burrows-Wheeler Transform Reversing the example above is done like this:

14. 14/24 Reversing the example above is done like this:

15. 15/24 Theoretical Background Lzw Coding

16. 16/24 Theoretical Background The initial dictionary # = 00000 = 0 A = 00001 = 1 B = 00010 C = 00011. Z = 11010 = 26 Example: TOBEORNOTTOBEORTO BEORNOT# Lzw Coding

17. 17/24 Theoretical Background Encoding

18. 18/24 Output Outline Introduction Frame of Compression Technique Theoretical Background Proposed Technique Experimental Results

19. 19/24 Frame Original Test Vectors Block Matching Lzw Coding Compressed Test Vectors Low Frequency Data Sets High Frequency Data Sets Burrows-Wheeler Transformation +

20. 20/24 All the test vectors are divided into several block of equal size. Proposed Technique BlOCK Number 1 2 3 4 5 6 7 Test Vector- 0100110000011000011010000111 Test Vector- 0100101101100111 Test Vector- 0100100101100111

21. 21/24 Proposed Technique T K,where K=1,2,3,…,n, as a matrix of M N, M > 2, N= block size of data.

22. 22/24 Output Outline Introduction Frame of Compression Technique Theoretical Background Proposed Technique Experimental Results

23. 23/24 Experimental Results

24. 24/24 Experimental Results

25. 25/24 Thanks