Signature Classifications

Signature classifications are an optional grouping system used to calculate statistics on a segment of Signatures. The default classification in Simscope is fail, indicating a functional failure.

For example, a manager can bucket signatures as fail (normal) vs testbench signatures, and can ignore testbench errors in statistics.

Tuple

Classifications are described by a three-part-tuple:

• Signature
• Component
• Branch

This allows signatures to be classified as, for example:

• Flow
• Test
• Perf

This also means each Component and Branch can classify Signatures differently.

Categorized and Classified Scoring

Signatures which have been Classified, can be removed from Net Score calculation.

For example, a Manager can determine that some signatures are due to timeout, rather than a functional fail, and thus should be omitted from pass/fail score calculation.

• Score — Raw Score
• Net — Net Score, after applying Classification filters

Example

In the above example:

• The top row shows all jobs in Category build
• The second row shows all jobs in Category simulate
  • 200 timeout jobs have been omitted from the Net Score calculation
    • This improves the simulate score from .333 to 1.000
• The bottom row shows the totals for all Categories

Viewing signature classifications

For classifications column to show up in the Signature Search results, the user must be filtering to a single component and branch.

Changing classifications

To classify signatures, a user must have a User Role of Manager or above.

Classifying can be done either from the Regression Details page, or from a Signature Search page, with a specified Component and Branch.

Just choose from the dropdown box, and the Signature has been re-classified.