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.


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

isabelle-dev mailing list

Reply via email to