Tau 2014 Contest: Common Path Pessimism Removal

TAU 2014 Contest: Common Path Pessimism Removal (CPPR)

TAU 2014 would like to invite students to participate in this year's contest focused on Common Path Pessimism Removal (CPPR) during static timing analysis.

With the increasing number and significance of manufacturing and environmental sources of variability in the design and use of modern VLSI designs, timing analysis tools must somehow model and account for uncertainty. The predominant approach is to create early and late bounds, where each signal's delay is bounded by a lower- and upper-bound. However, this early/late timing split can cause the analysis to be artificially pessimistic, thereby leading to unnecessary increases for design turnaround time and cost. This contest will focus on removing this artificially-introduced pessimism in the fastest amount of time possible.

Update [10/01/14]: Some TAU 2014 Contest participants are also presenting their work at ICCAD 2014 in the Special Session: Pessimism Removal during Timing Analysis.

Participation

The contest registration is now closed.

For more information on the contest, please visit the Resources page. Please check the resources page often for updates.

Important Dates The contest has officially completed. We look forward to seeing you next year!

Contest Winners [contest presentation slides available under Resources] Update [03/18/15]: Additional TAU 2014 Pictures

Updates / Announcements

-- March 18, 2015: [Main page]: Added link to TAU 2014 pictures.

-- October 1, 2014: [Main page]: Uploaded winners' pictures and linked to the ICCAD's special session presentation.

-- March 14, 2014 [Resources]: Uploaded results of the finalists; uploaded output files for Phase3Benchmarks; uploaded output file evaluator.

-- March 12, 2014 [Resources]: Uploaded Phase2Bencharks.tar.gz and updated Phase3Benchmarks.tar.gz with Combo3v2.

-- March 09, 2014 [TAU Contest 2014][Resources]: Uploaded slides from the TAU 2014 Workshop (winners included in slides!); uploaded evaluation benchmarks.

-- January 29, 2014 [Resources]: Updated contest_rules.pdf to reflect updated evaluation machine specifications and lifted the upper bound of the runtime factor.

-- January 28, 2014 [FAQ]: Updated FAQ about floating point precision mismatches for the checker.

-- January 27, 2014 [Resources]: Released output file checker for comparisons.

-- January 27, 2014 [Resources][FAQ]: Updated ac97_ctrlv2.setup.1000.1 in ac97_ctrlv2.tar.gz to include 1000 tests; updated FAQ to address false positive reporting.

-- January 23, 2014 [Resources]: Updated ac97_ctrlv2.hold.1000.1 in ac97_ctrlv2.tar.gz to remove SDFF_X2_746:SE duplication.

-- January 22, 2014 [Resources][FAQ]: Updated ac97_ctrlv2.tar.gz and aes_corev2.tar.gz to reflect 0 paths for some tests; updated FAQ to address accuracy evaluation for paths and tests.

-- January 20, 2014 [Resources]: Released newer set of larger benchmarks for further testing.

-- January 13, 2014 [Resources][FAQ]: Updated contest_rules.pdf to reflect documentation and evaluation sections; updated FAQ to reflect s526v2.setup timing difference.

-- January 4, 2014 [Resources][FAQ]: Updated golden outputs in sV2.tar.gz to reflect better accuracy; updated FAQ regarding numeric precision.

-- December 20, 2013 [Resources]: Updated golden outputs for sV2.tar.gz benchmarks to reflect correct positive post-CPPR slacks.

-- December 20, 2013 [FAQ]: Updated FAQ.

-- December 19, 2013 [Resources][FAQ]: Updated sV2.tar.gz with the correct timing files. Please re-download. Updated FAQ.

-- December 18, 2013 [Resources]: Released sV2.tar.gz benchmark bundle. These 9 benchmarks are more complex, and should give your CPPR code more exercise.

-- December 11, 2013 [Resources]: Released aes_core.tar.gz benchmark + updated contest_education.pdf to reflect clarification on post-CPPR test slack.

-- December 9, 2013 [Resources]: Released tv80.tar.gz benchmark

-- December 6, 2013 [Resources]: Updated contest_file_formats.pdf to include modified output format and sample golden output

-- December 5, 2013 [Resources]: Updated s* benchmarks for connectivity

-- December 3, 2013 [Resources]: Updated example.tar.gz and s349.tar.gz benchmarks

-- December 2, 2013 [Resources]: Uploaded 7 new benchmarks, updated FAQ

--November 26, 2013 [Resources][FAQ]: Updated s27.timing to remove the extra arrival time input for clk_net, added s349.tar.gz, updated FAQ

--November 25, 2013 [Resources]: Updated Equation 2 in contest_education.pdf and clarified outputs in contest_rules.pdf

Contest Organizers

• Jin Hu, IBM Corp.
• Debjit Sinha, IBM Corp.
• Igor Keller, Cadence Design Systems

Acknowledgments

The contest organizers would like to thank the TAU 2013 Variation-aware Timing Contest winners IITimer for providing their source code.

Contact

If you have any questions or concerns regarding the contest, please send an email to tau <dot> contest <at> gmail <dot> com.