( DVcon 07 Item 5 ) --------------------------------------------- [ 04/24/07 ]

Subject: Certess Certitude

THE OTHER COVERAGE TOOL -- This is one of those "surprize" tools that popped
out of the user survey comments.  It appears to be some sort of jacked up
coverage tool, but I'm not exactly sure what it is.  (Read for yourself.)
Interestingly this is 1 of 2 new "coverage" tools -- check out the next part
of this report with user comments on the unnamed Nusym "DeNibulator" tool.  


  It's called Certess and their tool Certitude injects errors into the
  DUT and runs you regression to see if you would catch it.  They use
  a form of coverage to find which tests cover which errors and can
  optimize significantly so they do not have to run every test against
  every error.

  We have run Certitude against 5 block test benches and found problems
  with 4 of them including 2 of the test benches where a data bus was
  not being checked.  One had a commented out "Error" message and the
  other was overwriting bad data with good data just before checking
  (a pass by reference not value issue).

  There are some compatibility issues with Certitude and 0-in plus some
  philosophical questions on how they interact with assertions, but we
  are very pleased with the tool.

  The company has been responsive to our open issues (about 15 in all,
  so it's not perfect), and have probably one of the nicest user
  interfaces I have ever seen in a tool.  This would have to get my vote
  for the coolest technology I have seen in the last couple of years.

      - [ An Anon Engineer ]


  We use Spyglass and code coverage with ICC.

  We also use a tool named Certitude from the Certess company, which gives
  much better results than code coverage in order to estimate where we
  are in verification.  We're in the process of replacing traditional code
  coverage.  You should have a look to http://www.certess.com -- one of
  the worst EDA web sites, but you can still get information.

      - Thomas Goust of STmicroelectronics


  Linter: Cadence HAL
  Code coverage: Cadence Conformal
  Testbench checking: Certess Certitude

  Certitude is a new class of tool, btw, not sure where to put it on your
  question list.  It fills what was previously a fairly large hole in most
  DV methodologies.

  You also didn't ask about CDC checking, which is very important on
  today's multi-clock designs.

      - Carey Kloss of Aquantia


  Certess Certitude tool is used for RTL code coverage and verification
  environment qualification.  Certitude is the only tool to measure the
  quality of a verification environment (tests & checkers effectiveness).
  We found several RTL bugs after improving efficiency of verif platform
  following the Certitude report.

      - [ An Anon Engineer ]


  For coverage we use line coverage from Cadence ICC and functional
  coverage from specman e.  We are just starting to ramp up on the
  tool from Certess called Certitude.  It has so far shown itself to
  be useful in closing the gap between what you planned on testing
  and what you should have tested.

      - [ An Anon Engineer ]


  We use Cadence HAL to check  RTL coding and TransEDA Vnavigator for
  code coverage.  Since last year, I use another tool call Certitude
  from Certess.  This tool is now mendatory in our verification flow.
  It allow us to get metrics and answer the question: how confident
  could we have in our verification to release the chip and tape-out?
  We can identify exactly the risk we take releasing the chip with
  the verification level we have achieve.

      - Xavier Jacquart of STmicroelectronics


  Linter: Atrenta Spyglass
  Coverage: builtin coverage tool from VCS -- CoverMeter

  Though it is not coverage tool per se, there is a new tool in this
  space, Certitude, which goes beyond traditional coverage tools, and
  helps address verification quality/closure/issues with propagation
  and detection.  Part of the tool is the 'activation' phase where it
  identifies if a particular line/cond is activated by your existing
  testbench.  Indirectly that gives your coverage numbers.

  Since there is no specific question about this topic, let me cover
  the tool here.  It also injects errors into the design and checks if
  the error is 'propagated' to the checkers and also if verification
  checkers are 'detecting' the errors.  So it helps one assess the
  quality of their verification code. Very unique tool in that respect.

  We have been using it for few months now.  Though, we have not
  finished a project with the tool, so far it proved to be useful.

      - [ An Anon Engineer ]


  Side comment - we recently evaluated Certess Certitude and had good
  results.  It looks for holes and give us feedback on how well we were
  testing our design.  What we really ended up finding out in the end
  is that the tool provided a comprehensive 'checker review'.  We spend
  gobs of time in code reviews, but never really put the same effort
  into reviewing how our checkers are written, leaving open the serious
  possibility of bugs.  Certess is an automated way to determine how
  good a job our test environment was doing at activating and detecting
  bugs in the design.  Highly recommended.

      - [ An Anon Engineer ]


  I've had the opportunity to use a new tool from Certess called
  Certitude, that doesn't fit into any of your questions.  It helps to
  show whether the testbench is capable of propagating and detecting
  bugs in the design.  Think of it as a verification tool for your
  verification environment.

      - Steve Pearlmutter of TwoTents Systems


  We are looking at a tool called Certitude from Certess.  During our
  preliminary it uncovered missing checkers and missing test scenarios
  in our verification environment.

      - [ An Anon Engineer ]


  Recently we looked at a new company Certess and their tool Certitude.

  Its metrics exceeds that of coverage by identifying false positives
  (executed code measured by coverage that either never propagates values
  to your testbench or is not verified correctly) as well as revealing
  weaknesses in your verification strategy (e.g., non-propagation of
  injected faults can suggest the need for additional or more robust test
  cases, which would not be obvious from coverage alone).

      - [ An Anon Engineer ]

Index    Next->Item

 Sign up for the DeepChip newsletter.
Email
 Read what EDA tool users really think.


Feedback About Wiretaps ESNUGs SIGN UP! Downloads Trip Reports Advertise

"Relax. This is a discussion. Anything said here is just one engineer's opinion. Email in your dissenting letter and it'll be published, too."
This Web Site Is Modified Every 2-3 Days
Copyright 1991-2024 John Cooley.  All Rights Reserved.
| Contact John Cooley | Webmaster | Legal | Feedback Form |

   !!!     "It's not a BUG,
  /o o\  /  it's a FEATURE!"
 (  >  )
  \ - / 
  _] [_     (jcooley 1991)