1 / 62

Improving NoC-based Testing Through Compression Schemes

1. 2. Improving NoC-based Testing Through Compression Schemes. Érika Cota 1 Julien Dalmasso 2 Marie-Lise Flottes 2 Bruno Rouzeyre 2 WNOC 2007. Introduction. SoC design communication among cores Network-on-chip. SoC testing access to cores

cody-byers
Download Presentation

Improving NoC-based Testing Through Compression Schemes

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. 1 2 Improving NoC-based Testing Through Compression Schemes Érika Cota1 Julien Dalmasso2 Marie-Lise Flottes2 Bruno Rouzeyre2 WNOC 2007

  2. Introduction SoC design • communication among cores • Network-on-chip SoC testing • access to cores • 1500 standard NoC-based SoC • available access to each embedded core • efficient communication mechanism • reuse the NoC as Test Access Mechanism (TAM)

  3. NoC-based system Wrapper SoC core core core Router core core core

  4. NoC-based testing ATE SoC core core core core core core

  5. Reuse Model ATE SoC core core core core core core

  6. Reuse Model ATE SoC core core core core core core

  7. Reuse Model Wrapper: - test mode - basic 1500 modes SoC core core core core core core

  8. NoC-based Testing • Minimize total test time: • Maximize the use of network resources during test • Test scheduling techniques • Preemptive • Non-preemptive • Wrappers design • NoC protocol • 1500-compliant

  9. NoC-based Testing Approaches • Preemptive testing • One vector per message, non-reserved paths • Non-preemptive testing • All vectors in one message, dedicated paths 1 0 1 1 packet header 0 1 0 0 0 0 0 1 test header flit flit CUT 1 0 0 1 1 01 flit tail

  10. NoC-based Testing Approaches • Preemptive testing • One vector per message, non-reserved paths • Non-preemptive testing • All vectors in one message, dedicated paths • More than 1 scan slice per flit 1 0 11 packet header 0 1 00 0 0 01 test header Slice 1 Slice 2 flit CUT 1 0 011 01 flit tail

  11. NoC-based Testing Approaches • Preemptive testing • One vector per message, non-reserved paths • Non-preemptive testing • All vectors in one message, dedicated paths • More than 1 scan slice per flit 1 0 11 packet header 0 1 00 0 0 01 test header Slice 1 Slice 2 flit CUT 1 0 011 01 flit tail

  12. Reuse Model: number of Test Ports ATE = 2W channels SoC core core core w w w w w core core core • 1 core under test

  13. Reuse Model: number of Test Ports ATE = 4W channels SoC w w w w core core core w w w w w core core core • 2 cores under test

  14. Test Time Number of Extra Pins DfT Costs • p93791 • 103 Inputs • 79 Outputs • 66 Bidirs • 32 Cores

  15. Test Time Number of ATE Channels ATE Costs • p93791 • 103 Inputs • 79 Outputs • 66 Bidirs • 32 Cores

  16. Problem How to increase the number of test ports Increase test parallelism Maximize NoC channels usage without increasing the ATE cost?

  17. Goal Goal: • Use Horizontal compression to reduce the number of ATE channels required for each NoC test port, i.e. to increase the number of test ports How ? • Combine a horizontal compression scheme with a non–preemptive test scheduling approach to reduce test time

  18. Outline • Compression applied to NoC-based testing • Horizontal compression method • Test scheduling algorithm • Experimental results • Final remarks

  19. Core Core router router router router wrapper wrapper W W Core Core N W wrapper wrapper W W Compression Applied to NoC-based Test W input

  20. router Compression Applied to NoC-based Test Functional input pins NoC Fi Core i M decompressor wrapper W W W Communication channels W M≤ Fi ≤ W

  21. router Compression Applied to NoC-based Test Functional output pins NoC Fi Core i M compressor wrapper W W W Communication channels W M≤ Fi ≤ W

  22. router Compression Applied to NoC-based Test • Each test port needs less than W bits • Less ATE channels per port • Increase the number of possible test ports • Increase test parallelism Functional input pins NoC Fi Core i Wk decompressor wrapper W W W Communication channels W Wk≤ Fi ≤ W

  23. Compression Applied to NoC-based Test • Horizontal compression • Test width reduction is the primary goal • Test vectors compression • Implies extra hardware at NoC-level (decompressor sharing) • May increase cores test time • Test responses compression • Implies extra hardware at NoC-level • Does not affect core test time

  24. On-chip Decompressor Compression W Scan Chains Compressed Test Data Test Data W W Core M M ATE Outline Horizontal Compression Techniques • Compression applied to NoC-based testing • Horizontal compression method • Test scheduling algorithm • Experimental results • Final remarks M < W

  25. Horizontal Compression: Requirements • Features • Circuit netlist independent (suitable for IPs) • Test data independent (additional test patterns) • Specific tools independent • Low cost hardware decompressor • No impact on fault coverage • Allow Shared decompressor for several cores

  26.         Horizontal Compression • Many published methods • Take advantage of Don’t Care bits (X’s) in test sequence • May increase core Test Time

  27. Decompressor architecture From ATE M 0 0 0 0 Add Cells Output Shift Register W To scan chains [1] Julien Dalmasso, Marie-Lise Flottes, Bruno Rouzeyre: Fitting ATE Channels with Scan Chains: a Comparison between a Test Data Compression Technique and Serial Loading of Scan Chains - DELTA 2006: 295-300

  28. 0 0 1 0 1 1 1 1 0 0 1 0 1 1 1 1 1 1 1 1 1 1 0 0 X 1 1 1 0 1 1 0 1 0 1 0 0 0 1 0 1 X 0 0 1 0 0 1 1 X X 1 1 1 1 0 1 X 0 X 0 X X 1 0 1 0 0 X 0 X 1 X 1 X X 1 1 0 X X 0 0 X X 0 0 0 X X X 1 0 X X 1 X 1 X 1 X X X X X 0 X 0 X 0 X X X 0 X 0 X X X 0 Decompression synchronization Original test Sequence Scan enable Control 0 0 S1 -> S2 : 0 1 S2 -> S3 : 1 1 S3 -> S4 : - - Compressed test Sequence Sc1 1 0 Sc2 0 0 FSM Sc3 0 1 Sc4 1 1 CLK Sc5 0 1 Sc6 1 0

  29. Compression Applied to NoC Data Test pattern: 011010110101110 Compressed test packet (M = 2) Original test packet (W= 5) Uncompressed test packet (W = 5) packet header packet header packet header packet header test header test header packet header 01101 01101 test header 01101 01101 test header 01110 01110 test header tail tail 01 10 compressor decompressor 1X 00 01 tail

  30. Compression Applied to NoC-based Test • Example for d695 ITC’02 benchmark • uncompressed and compressed data (#flits) Compression may increase test time of individual cores

  31. 32 ATE channels Compression Applied to NoC-based Test • Conclusion: • Local increase in test time • Increase test parallelism • Global test time reduction • d695 • 32 Inputs • 32 Outputs • 10 Cores 33%

  32. Outline • Compression applied to NoC-based testing • Horizontal compression method • Test scheduling algorithm • Experimental results • Final remarks

  33. Test Scheduling Using Dedicated Paths • Each core is associated with a routing path • Includes input and output ports • All resources are reserved until test completed • Test pipeline maintained • No complex logic • Similar to a circuit switching • Efficiently assign I/Os and paths to core • Each input port leads to a different core test time [2] C. Liu, E. Cota, H. Sharif, D.K. Pradhan: Test Scheduling for Network-on-Chip with BIST and Precedence Constraints - ITC2004: pp. 1369-1378.

  34. Test Scheduling with Compression • Two test packets per core • Compressed test vectors • Uncompressed test responses • Pre-defined number of test interfaces • Number of inputs = number of outputs • List of I/O pairs • For each core and for each I/O pair • Input width changes (compression ratio changes) • size of input test packets pre computed

  35. Test Scheduling with Compression Define test packets Define access paths for each core Select a core Find available access path Schedule packet

  36. Test Scheduling with Compression Define test packets Packets sorted by probable test time Define access paths for each core Select a packet Find available access path Schedule packet

  37. Test Scheduling with Compression Define test packets Packets sorted by probable test time Define access paths for each core Select I/O pair that leads to minimal total test time Select a packet Find available access path Schedule packet

  38. Test Scheduling with Compression Define test packets Packets sorted by probable test time Define access paths for each core Select I/O pair that leads to minimal total test time Select a packet Find available access path If no path is found, try next core Schedule packet

  39. 6 5 4 8 10 7 3 9 2 1 Test Scheduling Using Dedicated Paths • d695 from ITC02 benchmark • Channel width=32 • 3 inputs • 10, 10, 12bits • 3 outputs • I/O pairs • 3/9 • 6/7 • 8/4 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  40. 6 5 4 8 10 7 3 9 2 1 10856 10856 6 6 9869 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  41. 5 4 8 10 7 3 9 2 1 15459 6826 6850 5 5 5 9869 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  42. 4 8 10 7 3 9 2 1 15115 5829 12655 4 4 4 6826 5 9869 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  43. 8 10 7 3 9 2 1 5829 11431 14013 10434 4 8 8 8 6826 5 9869 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  44. 10 7 3 9 2 1 5829 10434 4 8 6206 10069 5 10 9869 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  45. 7 3 9 2 1 5829 10434 4 8 6206 10069 5 10 9869 13328 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  46. 3 9 2 1 5829 10434 4 8 6206 10069 12576 5 10 3 9869 13328 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  47. 9 2 1 5829 10434 11022 4 8 2 6206 10069 12576 5 10 3 9869 13328 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  48. 9 1 5829 10434 11022 11047 4 8 2 1 6206 10069 12576 5 10 3 9869 13328 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out Out 9 8 7 Out In 10

  49. 9 5829 10434 11022 11047 4 8 2 1 6206 10069 12576 5 10 3 9869 13328 9 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out 13412 Out 9 8 7 7 Out In 10

  50. 9 1 5829 10434 11022 11047 4 8 2 1 6206 10069 12576 5 10 3 9869 13328 7 6 Test Scheduling Using Dedicated Paths 1 5 10 2 12 In 10 In 3 6 4 Out 9 Out 9 8 7 Out In 10

More Related