On 07/11/2017 23:13, Makarius wrote:
For Flyspeck-Tame a small performance loss remains. It might be worth
trying to configure the Isabelle/HOL codegen to use FixedInt instead of
regular (unbounded) Int. Thus it could become faster with Poly/ML 5.7.1
than with 5.6.


Just as for Isabelle itself, I don't want generated code to abort with overflow or even worse to overflow silently.

Tobias

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

_______________________________________________
isabelle-dev mailing list
isabelle-...@in.tum.de
https://mailmanbroy.informatik.tu-muenchen.de/mailman/listinfo/isabelle-dev

Reply via email to