KEVOS
ArticlesServicesCase studiesAboutContact
ArticlesServicesCase studiesAboutContact
← ArticlesFactor Base Selection and SmoothnessEngineering · Engineering MathematicsLesson 830/884← PrevNext →
ArticlePublished 7 Aug 20262 min readBy Kevin Joginfactor basesmoothnessbound selectiontrade-off
On this page

Ask about this page

KEVOS AIFactor Base Selection and Smoothness

KEVOS knowledge first · trusted web sources when needed

Sub-exponential Class Group Computation

Factor Base Selection and Smoothness

Choosing the factor base for class group computation, the smoothness trade-off, and how base size interacts with the linear algebra.

Engineering / MathematicsSub-exponential Class Group Computation2 min readKV-MATH-0628

The factor base is the set of prime ideals over which relations are expressed. Its size controls a trade-off that determines the entire running time of the algorithm.

The trade-off

The factor base size trade-off
Larger factor baseSmaller factor base
Relations easier to findRelations harder to find
More relations neededFewer relations needed
Larger matrix; harder linear algebraSmaller matrix
Higher smoothness probabilityLower smoothness probability

Key point

The optimum balances relation collection time against linear algebra time. Because these scale differently, the balance point moves with the discriminant, and a base size tuned for one problem size is wrong for another.

Construction

Building the factor base

  1. Choose a norm boundGuided by the theory and by measurement.
  2. Enumerate rational primesUp to the bound.
  3. Decompose eachFind the primes above it — see simple decomposition.
  4. Filter by normKeep prime ideals whose norm is below the bound; inert primes of high residue degree are usually excluded.
  5. IndexStore in a structure supporting fast smoothness testing.

Note

Prime ideals of residue degree greater than one have large norm and rarely appear in smooth relations, so they are often excluded. Including them enlarges the matrix for little gain.

The generation guarantee

Under GRH, prime ideals of norm below a bound proportional to the square of the logarithm of the discriminant generate the class group. This is what makes a small factor base sufficient — see Minkowski and Bach bounds.

Caution

If the base fails to generate, the computed group is a quotient of the true class group. This failure is invisible in the relation matrix and is caught only by external verification.

Smoothness

An ideal is smooth over the base when its norm factors entirely over the corresponding rational primes and every prime ideal appearing is in the base.

Pitfall

Norm smoothness is necessary but not sufficient. A norm may factor over the base primes while the ideal itself involves a prime above one of them that was excluded from the base. Both conditions must be checked.

Testing smoothness

Cost

Smoothness testing is the inner loop and dominates relation collection. Trial division against the base primes is usually the right method because the base is exactly the set of primes of interest; batch methods amortise across many candidates.

Practical sizing

In practice the base is sized by measurement rather than by the theoretical optimum: run a short relation collection at several base sizes and extrapolate. The theoretical bound then serves as a completeness guarantee rather than as the operative sizing rule.

Source. Henri Cohen, A Course in Computational Algebraic Number Theory, Springer GTM 138 — 6.5.2. 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

  • Minkowski and Bach Bounds
  • Smoothness and Sub-exponential Complexity
  • Quadratic Sieve Factor Base Selection
  • Ideal Reduction in Number Fields
  • Relation Matrix Construction

Continue learning

Ideal Reduction in Number FieldsArticle · Engineering MathematicsNEXT LESSON →Relation Matrix ConstructionArticle · Engineering MathematicsBuchmann's Sub-exponential Algorithm: OverviewArticle · Engineering MathematicsRegulator and Fundamental Unit RecoveryArticle · Engineering Mathematics
KEVOS · Engineering, manufacturing and project improvement
ArticlesServicesCase studiesAboutContact
© 2026 KEVOS®