It's to allow the config to be read so that name mapping can be performed (particularly if an extracted view is used - then it can do name mapping between schematic and extracted views for out-of-context probing).

What I don't understand is why reading the cds.lib would cause any problems - this is an expected part of the flow, so I'm not sure why you want to disable it?

Regards,

Andrew.

You might think it's not needed, but any operation which requires save statements which were schematic names and have to map to the corresponding names in an extracted view needs the information about the config view.

It may also be needed to map names when accessing results.

However, it's quite possible when you are not using an extracted view that this isn't needed (I can't be certain without seeing your setup). I don't believe there's an option to turn it off - normally it wouldn't cause any performance issues (reading a cds.lib should be quick) and I wouldn't expect warnings in a  properly set up environment.

The best thing to do is to contact Cadence customer support at support.cadence.com - that way we can better understand what the impact is in your specific case and maybe see if an option can be added to prevent this being written (to mitigate a performance issue that can't be solved another way - I'd hope that the performance issue should be solved rather than requiring a user switch to improve performance with the potential loss of functionality). To do that though, it requires better understanding of why this is causing a problem for you.

Can you contact customer support please?

Regards,

Andrew.

For a schematic view, it's not needed because it knows the global switch view list. If a config view is used, the view list could be non-uniform throughout the hierarchy, and to do the appropriate mapping it needs to know which view was used for each cell - and that information is contained in the config - which is why it needs to know what the config was if a config was used and not when a schematic was used at the top level.

I can't comment as to whether it's essential for some features in the case when an extracted view is not used (as in your case) - I think this needs input from Cadence R&D. There may be good reasons to need to know what the view switching is throughout the hierarchy in other situations too - I don't know for certain though. it would be useful to understand what the impact is in terms of performance - ideally so that a reproducible testcase can be created and provided to R&D to improve the tool.

That's why I was asking you to contact customer support. Nothing will happen just because I've been involved in a discussion here (especially as I don't know enough to reproduce the performance problem you are having, but even then it would need me to have sufficient time to work on it; this is precisely what the support mechanism is for). Have you contacted customer support?

Thanks!

Andrew.