Created attachment 753 [details]
Adaptor for GMP rational arithmetic type
The attached file enables GMP rationals (mpq_class) support in Eigen. Minor issues arise occationally due to expression templates used by GMP, but in general it works fine, could be added to unsupported modules, next to MPRealSupport.
Thanks but some cleaning are required:
- cdiv is not used anymore
- mini/maxi are commented out
- on our side we should really bypass the heavy product kernels for anything but trivial scalar types. (that's for us!)
Then to add such a support module we need the usual unit tests, cmake detections, etc.
Thanx for the feedback.
I will update/complete the file+tests+cmake and make a pull request when time permits.
I did not quite catch the third point (heavy product kernels).
Do you mean that rationals are heavy (sure they are, yet some users might need it) ?
Or is it that I could bypass gebp_kernel ?
-- GitLab Migration Automatic Message --
This bug has been migrated to gitlab.com's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.com/libeigen/eigen/issues/1354.