I was trying to calculate the number of points of the curve $E:y^2 = x^3 + 4x^2 + 2x$ over $\mathbb{F}_p$ for $p\equiv 1\bmod 8$ (In order to have $\sqrt{-2}\in\mathbb{F}_p$) but I did not succeed. This curve is mentioned in Silverman's Advanced Topics in the Arithmetic of Elliptic curves (Proposition 2.3.1) to have multiplication by $\sqrt{-2}$.

Over these primes $p\equiv 1\bmod 8$ the curve $E$ has full $2$-torsion so $E(\mathbb{F}_p)\cong \mathbb{Z}/(2)\times \mathbb{Z}/(k)$.

In this case my conjecture is that the size will be related to the factorization of $p=(a+b\sqrt{-2})(a-b\sqrt{-2})$ over $\mathbb{Z}[i\sqrt{2}]$, that is $p=a^2 + 2b^2$. Hence, $\#E(\mathbb{F}_p)=p+1\pm 2a$ where $a$ is odd (and the sign I do not know how to choose it yet). Calculating this reminds me to the proof of the Last Entry of Gauss Tagebuch.

I would like to have an elliptic curve with CM by $\sqrt{-2}$ such that I can know the number of points in terms of $p$. Does anybody has a suggestion? or maybe another curve?



2 Answers 2


Sorry for self-advertisement: the question of the "sign" that one must choose (equivalently which $\pi$ or $\overline{\pi}$) is entirely answered in my book GTM239, Section 8.5.2, some of the results being due to Mark Watkins (unpublished I believe). The main idea is to show that any CM curve is equivalent in an evident sense to a "basic" CM elliptic curve with a specific equation, giving a relation between the number of points of the initial curve and the "basic" curve, and then Theorem 8.5.8 determines explicitly the number of points of that basic curve. I did not do the exercise for your specific example, but it is completely algorithmic.

  • 2
    $\begingroup$ I'm fortunate to have purchased an electronic copy of GTM239. The basic CM elliptic curve for $\mathbb{Z}[\sqrt{-2}]$ is isomorphic to $E : y^{2} = x^{3} + 4x^{2} + 2x$, and for this curve Theorem 8.5.8 gives that $a_{p}(E) \equiv 2 \pmod{8}$ when $p \equiv 1 \text{ or } 11 \pmod{16}$ and $a_{p}(E) \equiv -2 \pmod{8}$ when $p \equiv 3 \text{ or } 9 \pmod{16}$. $\endgroup$ Jul 7, 2017 at 1:30
  • 1
    $\begingroup$ I should add that I just learnt that Rubin--Silverberg published a paper on this subject in Math Comp 2009, including proofs this time. $\endgroup$ Jul 7, 2017 at 20:55

Your conjecture is true and follows from a theorem in Cox's book ''Primes of the form $x^{2} + ny^{2}$.'' (Theorem 14.16 on page 317, although this is from the first edition.) In particular, if $\mathcal{O}$ is an order in an imaginary quadratic field, $L$ is the ring class field associated to $\mathcal{O}$ and $E/L$ is an elliptic curve, with good reduction at a degree one prime $\mathfrak{P}$ of $L$, then $\mathcal{O}/\mathfrak{P} \cong \mathbb{F}_{p}$ and $$ |E(\mathbb{F}_{p})| = p+1 - (\pi + \overline{\pi}) $$ where $\pi \in \mathcal{O}$ and $p = \pi \overline{\pi}$. [ This result was originally proven by Deuring, and in fact Deuring's result is more general. ]

Your conjecture follows by taking $\mathcal{O} = \mathbb{Z}[\sqrt{-2}]$, $L = \mathbb{Q}(\sqrt{-2})$ and $p$ to be a prime $\equiv 1 \text{ or 3 } \pmod{8}$ (which implies that $\sqrt{-2} \in \mathbb{F}_{p}$). In this case, your $\pi = a + b \sqrt{-2}$. This theorem doesn't answer the question of what the sign is in the equation $|E(\mathbb{F}_{p})| = p+1 \pm 2a$.

  • $\begingroup$ Thank you for your proof of the conjecture. I gave the answer to @Henri Cohen because now I solved my problem. Thanks Jeremy! $\endgroup$ Jul 12, 2017 at 15:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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