You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Dec 5, 2023. It is now read-only.
We were originally using Montgomery form for representing base field and scalar field elements, which has then be removed in favour of a canonical encoding for the base prime field, to match the implementation on winterfell and miden.
However, if we stay away from the latter, we may want to go back to Montgomery form and use the same wrapping approach in winterfell for our own version of f64. While Fp6 multiplication and squaring may be slightly slower than it currently is, addition/subtraction/doubling should be faster. In addition, we wouldn't have to deal with canonicalization of elements before feeding them into a Hasher for instance, as all element representations are unique in Montgomery form.
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
We were originally using Montgomery form for representing base field and scalar field elements, which has then be removed in favour of a canonical encoding for the base prime field, to match the implementation on winterfell and miden.
However, if we stay away from the latter, we may want to go back to Montgomery form and use the same wrapping approach in winterfell for our own version of f64. While Fp6 multiplication and squaring may be slightly slower than it currently is, addition/subtraction/doubling should be faster. In addition, we wouldn't have to deal with canonicalization of elements before feeding them into a Hasher for instance, as all element representations are unique in Montgomery form.
The text was updated successfully, but these errors were encountered: