0
$\begingroup$

I have a frustration for the term "supersingular" elliptic curve and "non-singular" elliptic curve.

In some paper, performance evaluation is done using supersingular elliptic curve and in some cases, they use non-singular elliptic curve. Are they same or different curves?

$\endgroup$

1 Answer 1

3
$\begingroup$

Every elliptic curve is non-singular. Some elliptic curves are supersingular.

A quote from "The Arithmetic of Elliptic Curves" by Silverman, Remark V.3.2.2:

Do not confuse the notions of singularity and supersingularity. A supersingular elliptic curve is, by definition, an elliptic curve, so it is nonsingular. The origin of the potentially confusing terminology is as follows. Historically, elliptic curves defined over $\mathbb{C}$ whose endomorphism rings are larger than $\mathbb{Z}$ were called singular, where "singular" was used in the sense of "unusual" or "rare". However, in this sense, all elliptic curves defined over $\bar{\mathbb{F}}_p$ are singular! The endomorphism rings of most elliptic curves over $\bar{\mathbb{F}}_p$ are order in imaginary quadratic fields. It is only the rare and unusual curve whose endomorphism ring is an order in a quaternion algebra, whence the term "supersingular".

$\endgroup$
6
  • $\begingroup$ It means all non-singular curve is not supersingular? $\endgroup$
    – myat
    Commented Jan 31, 2017 at 12:09
  • $\begingroup$ You beat me to it. I knew I had read a remark about that terminology somewhere... $\endgroup$
    – yyyyyyy
    Commented Jan 31, 2017 at 12:11
  • $\begingroup$ @CurveEnthusiast. In some performance evaluation, they describe two using supersingular and another with non-singular. I don't know they are same or not. $\endgroup$
    – myat
    Commented Jan 31, 2017 at 12:13
  • 1
    $\begingroup$ @myat Read my statement carefully. An elliptic curve is non-singular by definition. There do not exist non-singular elliptic curves. Perhaps they meant non-supersingular, and just made a typo. $\endgroup$ Commented Jan 31, 2017 at 12:29
  • $\begingroup$ @CurveEnthusiast I have difficulties with the singular definition. E.g. an Edwards curve have singularities if I embed it in the projective plane. Should I just have to look for singularities in the affine plane ? $\endgroup$
    – Ruggero
    Commented Jan 31, 2017 at 12:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.