How do you ensure the quality and coverage of your RTL waveform comparison tests?

1 File-based comparison

One of the most basic methods for waveform comparison is to use file-based comparison tools, such as diff or cmp, to compare the output files generated by your simulation or emulation tools. This method is simple to automate and incorporate into your regression flow, and can deal with large amounts of data. Nonetheless, file-based comparison has some drawbacks, like needing the output files to have the same format, structure, and order of signals, which may not be the case for different tools or versions. Additionally, it does not factor in timing differences, glitches, or noise that may occur in the waveforms, potentially leading to false positives or negatives. Moreover, it does not provide any insight into the root cause of the mismatches or the impact on the functionality of the design. Thus, file-based comparison is best suited for straightforward scenarios where the output files are consistent and the waveforms are expected to be identical.

2 Graphical comparison

Another method for waveform comparison is to use graphical comparison tools, such as GTKWave or SimVision, to visually inspect and compare the waveforms in a graphical user interface (GUI). This method allows you to easily spot and highlight the differences between the waveforms with color codes, markers, or annotations, as well as zoom in or out, filter, or group the signals to focus on the relevant parts. You can also apply timing or value tolerances to ignore minor deviations or glitches that do not affect the functionality of the design. Additionally, you can trace back the source of the mismatches by cross-probing the signals with the source code or schematic. However, this method can be time-consuming and tedious for large designs, prone to human errors, and difficult to automate or reproduce. Therefore, graphical comparison is best suited for interactive and detailed analysis when you need to understand and debug the differences between the waveforms.

3 Assertion-based comparison

A third method for waveform comparison is the use of assertion-based comparison tools, such as SVA or PSL, to specify and check the expected behavior and properties of the design using formal logic. This approach enables you to express the waveform comparison criteria in a concise, declarative way with temporal operators, variables, and functions, allowing for verification in a pass or fail manner. Furthermore, it can generate coverage reports and metrics to measure how well the waveform comparison tests cover the design functionality and corner cases. However, this method requires a good knowledge and skill of assertion languages and tools which may have a steep learning curve and a high entry barrier. Additionally, it may not be able to handle all types of waveform comparison scenarios, such as analog or mixed-signal waveforms, or complex mathematical operations. And it may introduce some overhead or performance issues depending on the complexity and number of assertions. Therefore, assertion-based comparison is best suited for rigorous and comprehensive verification where you need to ensure the correctness and completeness of the waveform comparison tests.