KEVOS
ArticlesServicesCase studiesAboutContact
ArticlesServicesCase studiesAboutContact
← ArticlesComputing with Elliptic Curves over CEngineering · Engineering MathematicsLesson 845/884← PrevNext →
ArticlePublished 7 Aug 20262 min readBy Kevin Jogincomplex numbersperiodsarithmetic geometric meanelliptic logarithm
On this page

Ask about this page

KEVOS AIComputing with Elliptic Curves over C

KEVOS knowledge first · trusted web sources when needed

Elliptic Curves

Computing with Elliptic Curves over C

Practical computation with elliptic curves over the complex numbers: periods, the torus map, and heights.

Engineering / MathematicsElliptic Curves2 min readKV-MATH-0643

Analytic methods over the complex numbers supply practical tools for curves over the rationals: periods, elliptic logarithms, and the real-place contribution to heights.

Computing periods

The period lattice is computed by the arithmetic-geometric mean, which converges quadratically. The real period and the complex one follow from the roots of the cubic.

Computing the period lattice

  1. Find the rootsOf the cubic defining the curve, numerically — see root finding.
  2. Apply the AGMTo appropriate combinations of the roots.
  3. Assemble the latticeThe periods follow from the limits.

Cost

The arithmetic-geometric mean doubles the correct digits each iteration, so periods to hundreds of digits cost only tens of iterations. This makes high-precision analytic work on elliptic curves entirely practical.

The elliptic logarithm

Mapping a point on the curve back to the torus is the elliptic logarithm. It is the inverse of the Weierstrass parametrisation and is computed by numerical integration or by the AGM.

Key point

The elliptic logarithm converts the group law into addition of complex numbers. Independence of points becomes linear independence of their logarithms, which is testable by lattice methods.

Heights

Naive height
Measures the size of the coordinates. Easy to compute, not quite additive under the group law.
Canonical height
A quadratic form on the Mordell-Weil group, obtained by a limiting process from the naive height. Genuinely quadratic.
Local decomposition
The canonical height decomposes into contributions from each place, with the real place computed analytically.

Key point

The canonical height being a quadratic form is what makes rank computation possible. Independence of points is detected by the height pairing matrix being non-singular, and the regulator of the curve is its determinant.

Finding points

Searching for rational points is guided by height bounds. Points of bounded canonical height lie in a bounded region of the torus, which converts the search into a lattice point enumeration — see short vector search.

Caution

Height bounds from descent can be very large, making the search region enormous. This is the practical obstacle to finding generators even when the rank is known.

Precision

Pitfall

Height computations involve cancellation between local contributions, so precision must comfortably exceed the size of the result. Rank determination from a numerically singular height matrix is unreliable unless the precision is known to be adequate.

Source. Henri Cohen, A Course in Computational Algebraic Number Theory, Springer GTM 138 — 7.4.1. Structural reference unverified: the source file was not available during authoring; chapter and section numbers are taken from the published edition and have not been checked against a physical copy.

Related pages

  • Lattices, Complex Tori and the Weierstrass p-Function
  • L-Functions and the Birch-Swinnerton-Dyer Conjecture
  • Curve Reduction and Tate's Algorithm

Continue learning

L-Functions and the Birch-Swinnerton-Dyer ConjectureArticle · Engineering MathematicsNEXT LESSON →Curve Reduction and Tate's AlgorithmArticle · Engineering MathematicsZeta Functions of Elliptic CurvesArticle · Engineering MathematicsSchoof's Point Counting AlgorithmArticle · Engineering Mathematics
KEVOS · Engineering, manufacturing and project improvement
ArticlesServicesCase studiesAboutContact
© 2026 KEVOS®