Code Coverage Assessment & Improvement Service


The provision of the code coverage assessment and improvement service is divided into three phases as depicted below.






Phase 1 - Preparation for Code Coverage Assessment and Improvement
The first phase is one of preparation and may be begun at the start of the project, as typically some RTL modules are being re-used and have testbenches in existence. These may be used to set up the environment and pipeclean the flow.

As testbenches are created by your engineers they maybe added to the automation script. The amount of setup time required will vary depending on the complexity of the testbench and its similarity to previous testbenches. This may continue until testbenches for all modules, intermediate integrations and the top level are complete.




Phase 2 - Improvement of Code Coverage - Provided at Levels 1, 2 and 3 for each testbench
At the first level, HTML annotated RTL code is generated which includes details of the selected coverage shortfalls. Your engineers may then use this to improve the coverage of their testbenches.

Level 2 is provided if you require assistance to improve the coverage. For each module where improvement is required, time will be required to become familiar with that module's specification, it's testbench and the RTL code. This time will vary depending on the complexity of the module.

Feedback on testbench coverage, and advice on improvement of coverage, may be provided as soon as the testbenches are available. The testbenches do not have to be complete, and in fact knowledge of the code coverage achieved so far will help the development of the testbench.

Level 3 is provided if you wish Jaskay Technology to both assess and improve the quality of the code coverage by testbench modifications, or perhaps the creation of a testbench. This will require time to become familiar with that module's specification, it's testbench and the RTL code. This time will vary depending on the complexity of the module and the nature of the code coverage shortfall.




Phase 3 - Monitoring of Code Coverage
Once the automation of one or more testbenches for Code Coverage is complete, regression runs may be performed at agreed periodic intervals. Any changes in coverage will be reported.

It is important to note that bugs can be introduced into testbenches during any modification, even late in the design, which can result in whole sections of RTL module code being left untested. Regression runs of code coverage are necessary until RTL freeze and testbench freeze.




The Three Phases working in Parallel
The three phases of Code Coverage assessment and Improvement overlap. For example, testbenches for intermediate integrations and the top level may be being prepared while module level testbenches are being monitored by regression runs at agreed periodic intervals.



Obtain a quotation here