KEVOS
ArticlesServicesCase studiesAboutContact
ArticlesServicesCase studiesAboutContact
← ArticlesThe Elliptic Curve Method: Stage OneEngineering · Engineering MathematicsLesson 872/884← PrevNext →
ArticlePublished 7 Aug 20262 min readBy Kevin JoginECMstage onesmooth scalarB1 bound
On this page

Ask about this page

KEVOS AIThe Elliptic Curve Method: Stage One

KEVOS knowledge first · trusted web sources when needed

Modern Factoring Methods

The Elliptic Curve Method: Stage One

ECM stage one: multiplying a point by a highly smooth scalar to reach the identity in one component.

Engineering / MathematicsModern Factoring Methods2 min readKV-MATH-0670

Stage one of ECM multiplies a point by a scalar containing every prime power below a bound. If the curve order in one component divides that scalar, the component reaches the identity and the factor appears.

The scalar

k = product over primes p <= B1 of p^floor(log_p B1)Every prime power up to the bound; a highly smooth integer.

Key point

If the curve group order modulo a prime factor divides this scalar — that is, if the order is smooth to the bound — the point becomes the identity in that component. A GCD then reveals the factor.

The procedure

ECM stage one

  1. Choose a curve and pointBy Suyama's parametrisation, so a point is known without a square root.
  2. Build the scalarOr multiply by each prime power in turn, which avoids forming an enormous integer.
  3. MultiplyUsing the Montgomery ladder in projective coordinates.
  4. Take a GCDOf the final coordinate with the modulus.
  5. Interpret or continueA factor, a restart, or proceed to stage two.

Cost

Multiplying by each prime power in turn rather than forming the full scalar keeps memory bounded and allows intermediate GCD checks. Checking every step is wasteful; checking periodically is the usual compromise.

Choosing the bound

ECM effort scales sharply with target factor size
Target factor sizeTypical boundCurves needed
15 digitsSmallTens
25 digitsModerateHundreds
35 digitsLargeThousands
45 digits and beyondVery largeTens of thousands or more

Key point

The bound is chosen for the size of factor being sought, not for the size of the number. This is what distinguishes ECM from the sieves and makes it the right tool for finding medium factors of very large numbers.

Curve selection

Parametrisations that force small factors into the curve order improve the odds appreciably. Suyama's construction guarantees divisibility by twelve, and other families guarantee more.

Note

Guaranteed small factors help because they remove those primes from the smoothness requirement, effectively shifting the order into a more favourable range.

Parallelism

Key point

Different curves are entirely independent, so ECM parallelises perfectly across any number of machines with no communication. This is a decisive practical advantage over the sieves, whose linear algebra stage does not parallelise nearly as well.

When stage one fails

If the order is smooth except for one larger prime, stage two finds it far more cheaply than raising the stage one bound — see stage two.

Source. Henri Cohen, A Course in Computational Algebraic Number Theory, Springer GTM 138 — 10.3.1-10.3.3. 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

  • Binary Powering and Exponentiation Chains
  • The Pollard p-1 Method
  • Modern Factoring Methods Compared
  • Elliptic Curve Arithmetic Modulo N
  • ECM Stage Two and Practical Tuning

Continue learning

Elliptic Curve Arithmetic Modulo NArticle · Engineering MathematicsNEXT LESSON →ECM Stage Two and Practical TuningArticle · Engineering MathematicsElliptic Curves Modulo NArticle · Engineering MathematicsQuadratic Sieve Factor Base SelectionArticle · Engineering Mathematics
KEVOS · Engineering, manufacturing and project improvement
ArticlesServicesCase studiesAboutContact
© 2026 KEVOS®