OCEAN script only works on re-evaluation

Dear imulsion,

imulsion said:

When I run my simulations, the script always results in eval err when executing axlGetMainSetupDB, because axlGetWindowSession returns nil.

Does the use of the following verbose form of the command provide a response of "nil" on the first attempt to run your script?

axlGetWindowSession( hiGetCurrentWindow())

Shawn

That's not a valid thing to do in an OCEAN script output. For a start, the script is run at each sweep point/Monte Carlo point/corner (or combinations of the above), and it's not going to work when simulations are running because the scripts (as are other output expressions) are run in a background process (either a background virtuoso if using ICRP or a background expression evaluator if using LSCS) and there's no "window" for axGetWindowSession to retrieve. That's also no axl/mae session in those background sessions because it would be locked by the parent virtuoso - so even if you could find a way of getting a session it wouldn't be able to write to it. This will therefore only work when running a single simulation (because single point optimisation in the job policy runs everything in the virtuoso session itself, at least for ICRP), or when doing re-evaluation (and even then, not when using the new distributed re-evaluation that's available in Virtuoso Studio IC23.1).

Andrew 

Hi Andrew,

Thanks for your reply. The reason I opted for this approach is so I could use axlGetCornerNameForCurrentPointInRun() (otherwise I couldn't think of a way to apply each trim code to each corner). Is there an alternative approach that is less "hacky", or is this kind of thing not really possible?

Cheers

I get why you've done it this way - it's probably simpler than the alternative which is to use the maeReadResDB() function to get a handle and then iterate over the results (I'd have a scalar result value to make it easier to find rather than needing to read waveform data) and then update the corners that way - and run this manually after simulation whenever you want to update the corner settings. What you've done is a bit more self-contained, but it just can't work at run time without needing a re-eval.

Andrew

Hi Andrew,

That makes a lot of sense. I'll stick with my current method, even though it doesn't quite work it's probably the easiest way of doing it (at the very least it executes automatically!) 

Thanks very much for your time

Cheers