-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fpRealToFP and fpToReal fail on trivial problems #6548
Comments
a couple of issues.
|
The model appears to be incorrect. |
…d on relevancy Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
I am going to close this for now as resolved:
CVC5 uses a different approach for enforcing to_fp constraints. It is very possible better as it appears to expand axioms on the fly instead of using a one-shot encoding into arithmetic. On the other hand, it appears also incomplete relative to the semantics of to_fp form what I could see. The corner cases may not manifest in standard use cases (and what is a standard use case for to_fp? It seems it wasn't well used prior to this report). There is a potentially nice project involved in handling other dynamic axioms for powers/exponentiation. It is self-contained and good entry point for someone interested in tinkering with internals while diving into approaches for handling non-linear functions. It is encapsulated here: There is good quality literature already on one approach to creating polynomial approximations. |
See: https://stackoverflow.com/questions/75186194
Even trivial usage of
fpRealToFP
andfpToReal
cause anunknown
result. I haven't yet seen these conversions work under any circumstances.According to an answer by Alias on the linked SO question, this fails in Z3 but works in CVC5.
The text was updated successfully, but these errors were encountered: