Smith Normal Form Calculator

Reduce an integer matrix to Smith normal form, track row and column operations, and test integer-system compatibility.

Results are calculated automatically as you enter data.

Integer matrix

Enter integer matrix values. The reduction supports up to a 6 by 6 matrix.

Formula and interpretation
Result Enter an integer matrix to begin.

▼ See explanations and tips below ▼

Related Calculators

What Is Smith Normal Form?

Smith normal form is a way to simplify an integer matrix without leaving integer arithmetic. It turns a matrix into a diagonal matrix by using reversible integer row and column operations.

For an integer matrix \(A\), the goal is to find integer unimodular matrices \(U\) and \(V\) such that

$$ UAV = D $$

where \(D\) is diagonal and has the form

$$ D = \operatorname{diag}(d_1,d_2,\ldots,d_r,0,\ldots,0) $$

with

$$ d_1 \mid d_2 \mid \cdots \mid d_r. $$

The nonzero diagonal entries \(d_1,d_2,\ldots,d_r\) are called invariant factors. They are not just a byproduct of the calculation; they are the key information preserved by the integer matrix. They reveal the matrix rank, describe quotient structures such as cokernels, and turn many integer linear systems into divisibility tests.

Smith normal form is especially useful when the question is not “what is the real-number solution?” but “does an integer solution exist?” Ordinary row reduction can divide by numbers such as \(2\) or \(5\), which is harmless over the real numbers but not reversible over the integers. Smith normal form keeps track of only operations that preserve the integer structure.

Why Smith Normal Form Matters

Smith normal form matters because many mathematical problems depend on integer structure rather than decimal or real-number approximation.

It is used to study:

  • Linear Diophantine systems, where all coefficients and solutions are integers.
  • Finitely generated abelian groups, where invariant factors describe the group up to isomorphism.
  • Integer lattices and quotient modules, where the diagonal form makes the quotient easy to understand.
  • Algebraic topology, where integer boundary matrices can be reduced to compute homology groups.
  • Exact matrix invariants, where the diagonal entries give information that ordinary row-reduced echelon form does not show.

For students, Smith normal form is often the bridge between linear algebra and abstract algebra. It looks like matrix reduction, but its logic comes from divisibility, greatest common divisors, and modules over principal ideal domains.

Key Terms to Know

  • Integer matrix: A matrix whose entries are integers.
  • Unimodular matrix: A square integer matrix with determinant \(1\) or \(-1\). Its inverse is also an integer matrix.
  • Invariant factor: A nonzero diagonal entry in Smith normal form, usually chosen positive and arranged so that each one divides the next.
  • Rank: The number of nonzero invariant factors.
  • Principal ideal domain: A ring in which every ideal is generated by one element. The integers \(\mathbb{Z}\) are the main example for this calculator.
  • Greatest common divisor: The largest positive integer dividing a set of integers. GCD calculations are central to Smith normal form.
  • Linear Diophantine equation: A linear equation or system whose coefficients and solutions are required to be integers.
  • Free variable: A variable not fixed by the diagonal system. In an underdetermined compatible system, free variables create infinitely many possible integer solutions.

How Smith Normal Form Works

Smith normal form uses three types of reversible integer operations:

  1. Swap two rows or two columns.
  2. Multiply a row or column by \(-1\).
  3. Add an integer multiple of one row or column to another row or column.

These operations correspond to multiplying by unimodular matrices. Row operations are recorded in a left multiplier \(U\), and column operations are recorded in a right multiplier \(V\). Because \(U\) and \(V\) are unimodular, they do not destroy the integer nature of the problem.

The diagonal form is written as

$$ UAV = D. $$

The matrix \(D\) is not just diagonal; its nonzero diagonal entries must satisfy the divisibility chain

$$ d_i \mid d_{i+1}. $$

That condition is what makes the Smith normal form canonical over the integers, up to signs. By convention, the invariant factors are usually taken to be nonnegative.

A useful way to understand the invariant factors is through minors. Let \(\Delta_k(A)\) be the greatest common divisor of all \(k \times k\) minors of \(A\), with \(\Delta_0(A)=1\). When \(A\) has nonzero invariant factors \(d_1,\ldots,d_r\),

$$ \Delta_k(A)=d_1d_2\cdots d_k $$

for \(1 \le k \le r\), so

$$ d_k=\frac{\Delta_k(A)}{\Delta_{k-1}(A)}. $$

This formula explains why invariant factors are true invariants of the matrix: row and column operations of the allowed kind do not change these determinant-divisor relationships.

Smith Normal Form and Integer Systems

Smith normal form is especially powerful for integer systems

$$ Ax=b. $$

If

$$ UAV = D, $$

then set

$$ x = Vy. $$

Substituting gives

$$ A(Vy)=b. $$

Multiplying both sides by \(U\) gives

$$ UAVy=Ub, $$

so the original integer system becomes

$$ Dy = Ub. $$

This is the main advantage. The matrix \(D\) is diagonal, so the system breaks into simple equations

$$ d_i y_i = (Ub)_i. $$

For an integer solution to exist, each nonzero diagonal entry must divide the corresponding transformed right-hand-side entry:

$$ d_i \mid (Ub)_i. $$

Any zero row in \(D\) gives a consistency condition:

$$ 0 = (Ub)_i. $$

If a zero row has a nonzero transformed right-hand side, the system is incompatible over the integers.

The calculator reports this zero-row inconsistency separately from a failed nonzero-diagonal divisibility condition, because the equations \(0=(Ub)_i\) and \(d_i y_i=(Ub)_i\) have different compatibility rules.

When the conditions pass, the diagonal system gives integer values for the constrained \(y_i\) values. Any remaining variables are free. The calculator displays one solution by setting free variables to zero, but an underdetermined compatible system can have more integer solutions.

Examples of Smith Normal Form in Practice

Example 1: A Small Square Matrix

Consider

$$ A= \begin{bmatrix} 2 & 4 \\ 6 & 10 \end{bmatrix}. $$

First compute the greatest common divisor of all entries:

$$ \Delta_1(A)=\gcd(2,4,6,10)=2. $$

For a full-rank \(2 \times 2\) matrix, \(\Delta_2(A)\) is the absolute value of the determinant:

$$ \Delta_2(A)=|2\cdot 10-4\cdot 6|=|20-24|=4. $$

The invariant factors are

$$ d_1=\Delta_1(A)=2 $$

and

$$ d_2=\frac{\Delta_2(A)}{\Delta_1(A)}=\frac{4}{2}=2. $$

So the Smith normal form diagonal is

$$ (2,2), $$

and the rank is \(2\).

This tells you more than the determinant alone. The determinant shows that the matrix is full rank, while the invariant factors show the exact integer divisibility structure.

Example 2: A Linear Diophantine Equation

The equation

$$ 6x+10y=b $$

can be viewed as the matrix equation

$$ \begin{bmatrix} 6 & 10 \end{bmatrix} \begin{bmatrix} x \\ y \end{bmatrix} =b. $$

The Smith normal form of the row matrix \(\begin{bmatrix}6 & 10\end{bmatrix}\) has diagonal entry

$$ \gcd(6,10)=2. $$

So the equation has an integer solution exactly when

$$ 2 \mid b. $$

For example, \(b=14\) is compatible because \(2\) divides \(14\):

$$ 6(-1)+10(2)=14. $$

But \(b=15\) is not compatible because \(2\) does not divide \(15\).

This is the simplest version of the same divisibility test used for larger matrices.

Example 3: A Rank-Deficient Edge Case

A zero matrix has no nonzero invariant factors:

$$ A= \begin{bmatrix} 0 & 0 \\ 0 & 0 \end{bmatrix}. $$

Its Smith normal form is also the zero matrix, and its rank is \(0\). The system

$$ Ax=b $$

then means

$$ 0=b. $$

So if \(b\) is the zero vector, every integer vector \(x\) is a solution. If \(b\) has even one nonzero entry, there is no solution.

This example shows why zero diagonal rows matter: they are not ignored. They impose consistency conditions on the transformed right-hand side.

How to Interpret the Result

The Smith normal form result should be read as a summary of the integer structure of the matrix.

Result part What it means
SNF diagonal The diagonal entries of \(D\) in \(UAV=D\). The nonzero entries are the invariant factors.
Rank The number of nonzero invariant factors.
Invariants The nonzero diagonal values, usually positive and arranged so that each divides the next.
Divisibility status A check that the diagonal values satisfy the Smith normal form condition \(d_i \mid d_{i+1}\).
Shape The original number of rows and columns in the matrix.
Step visual and operation table A record of the row and column operations used during reduction.
Compatibility for \(Ax=b\) A divisibility and zero-row check after transforming the system to \(Dy=Ub\).
Displayed solution One integer solution when compatible, with free variables set to zero.

A small invariant factor means a weaker divisibility restriction. A larger invariant factor means a stronger divisibility restriction. A zero diagonal entry after the nonzero invariant factors means that some equations in the diagonalized system are consistency checks rather than equations determining variables.

The rank should not be confused with the number of columns or rows. A \(2 \times 4\) matrix, for example, can have rank at most \(2\), and its Smith normal form can still be meaningful.

Common Mistakes and Misconceptions

One common mistake is to treat Smith normal form like ordinary row-reduced echelon form. Row reduction over the real numbers allows division by any nonzero number, but Smith normal form over the integers allows only reversible integer operations.

Another mistake is to enter decimal or fractional values. Smith normal form is an integer-matrix concept in this setting. A value such as \(1.5\) or \(\frac{3}{2}\) belongs to a different arithmetic setting and is not supported by this calculator.

Users also sometimes paste matrix rows with different lengths. A matrix must have the same number of entries in every row.

For systems \(Ax=b\), the vector \(b\) must have one entry for each row of \(A\). If \(A\) has \(m\) rows, then \(b\) must have length \(m\).

A displayed solution should not be read as the complete set of all solutions. If the system has free variables, there may be infinitely many integer solutions. The displayed solution is one representative obtained by setting free variables to zero.

Finally, the divisibility chain matters. A diagonal matrix is not in Smith normal form unless the nonzero diagonal entries satisfy

$$ d_1 \mid d_2 \mid \cdots \mid d_r. $$

When to Use Smith Normal Form

Use Smith normal form when you need to understand integer structure, not just real-number solvability.

It is useful for:

  • Testing whether an integer system \(Ax=b\) has an integer solution.
  • Finding invariant factors of an integer matrix.
  • Studying finitely generated abelian groups from generators and relations.
  • Understanding quotient groups or cokernels such as \(\mathbb{Z}^m/\operatorname{im}(A)\).
  • Checking rank and divisibility structure for rectangular as well as square matrices.
  • Comparing integer matrices under reversible row and column operations.

Do not use Smith normal form as a substitute for numerical linear algebra when the entries are measured data, decimals, floating-point values, or approximate real numbers. It is designed for exact arithmetic.

Limitations and Things to Keep in Mind

Smith normal form is exact mathematics, but an interactive calculator still has practical limits.

This calculator is intended for integer matrices up to \(6 \times 6\). It rejects decimal, fractional, symbolic, complex, non-finite, exponent-form, and unsafe-integer inputs. Blank matrix cells are treated as \(0\).

Matrix reduction, unimodular transformations, invariant factors, and optional integer-system solutions use exact integer arithmetic. To keep the browser responsive, the calculation stops with an error rather than an approximation if it exceeds 5,000 recorded operations or if an intermediate integer exceeds 10,000 decimal digits.

The operation table displays only the first recorded operations up to its interface limit. The visual step selector is useful for learning the reduction process, but it should not be treated as a complete proof for every large or complicated input.

For \(Ax=b\), the calculator checks compatibility and displays one solution when possible. It does not give a full parametric description of all integer solutions. If the system is underdetermined, additional solutions may be obtained by varying the free variables.

Every displayed integer solution is automatically substituted back into the original equation

$$ Ax=b $$

and is shown only when that exact verification succeeds. Independent verification is still appropriate for high-stakes work.

How to Use This Calculator

  1. Enter integer values in the matrix grid for \(A\). Blank cells count as \(0\).
  2. Use the add-row, add-column, delete-row, and delete-column controls to set the matrix size, up to the \(6 \times 6\) limit.
  3. Use the import control if you want to paste rows of integers. Separate entries with spaces, commas, or semicolons, and keep every row the same length.
  4. Optionally choose a built-in example preset.
  5. Optionally enter a vector \(b\) with one entry for each row of \(A\) to test the integer system \(Ax=b\).
  6. Review the SNF diagonal, rank, invariant factors, divisibility status, and original shape.
  7. Move the reduction step selector to inspect the row and column operations in the visual and operation table.
  8. Use the PNG download button if you want to save the current matrix visual.

Frequently Asked Questions

Is Smith normal form the same as row-reduced echelon form?

No. Row-reduced echelon form is usually computed over a field, such as the real numbers, where division by any nonzero value is allowed. Smith normal form works over the integers and uses only reversible integer row and column operations.

Can rectangular matrices have Smith normal form?

Yes. Smith normal form applies to rectangular integer matrices as well as square matrices. The diagonal entries continue only as far as the smaller matrix dimension allows, and the remaining rows or columns are handled as zeros or free variables depending on the context.

What does it mean if an invariant factor is \(1\)?

An invariant factor of \(1\) means the corresponding diagonal equation imposes no divisibility restriction, because \(1\) divides every integer. In group or quotient interpretations, a factor of \(1\) often contributes a trivial component.

What do zero diagonal entries mean?

Zero diagonal entries after the nonzero invariant factors mean the matrix has rank less than its maximum possible value. In a system \(Ax=b\), the corresponding diagonal equations become consistency checks of the form \(0=c_i\).

Why does the calculator require integers?

This version of Smith normal form is for matrices over \(\mathbb{Z}\). Decimals, fractions, symbolic expressions, and complex numbers belong to different arithmetic settings and require different algorithms or assumptions.

Does the displayed solution show all solutions of \(Ax=b\)?

No. When the system is compatible, the displayed solution is one integer solution with free variables set to zero. If the matrix has more variables than independent constraints, there may be infinitely many integer solutions.

Sources and References

Books

  1. Morris Newman. Integral Matrices. Academic Press, 1972. Relevant topics: equivalence of integer matrices, unimodular transformations, Smith normal form, and applications to integer linear equations. Google Books bibliographic record.
  2. David S. Dummit and Richard M. Foote. Abstract Algebra. 3rd ed., John Wiley & Sons, 2004. Relevant topics: principal ideal domains, modules over PIDs, invariant factor decompositions, and finitely generated abelian groups. Google Books bibliographic record.
  3. Nathan Jacobson. Basic Algebra I. 2nd ed., Dover Publications reprint, 2009. Relevant topics: modules, principal ideal domains, and Smith normal form background used in linear Diophantine systems. Google Books bibliographic record.

Online and Research Sources

  1. Richard P. Stanley. “Smith Normal Form in Combinatorics.” Journal of Combinatorial Theory, Series A, 2016. Used for general algebraic properties of Smith normal form, diagonal forms, determinant-divisor interpretation, and applications. MIT PDF.
  2. Raymond N. Greenwell and Stanley Kertzner. “Solving Linear Diophantine Matrix Equations Using the Smith Normal Form (More or Less).” International Journal of Pure and Applied Mathematics, 55(1), 2009, 49–60. Used for the integer-system compatibility conditions and solution form for \(AX=B\). PDF.
  3. Omar Antolín Camarena. “Using the Smith Normal Form to Compute Homology.” Universidad Nacional Autónoma de México course notes. Used for an accessible explanation of Smith normal form, invertible integer matrices, uniqueness up to signs, and cokernel interpretation. Course note.
  4. Morris Newman. “The Smith Normal Form.” Linear Algebra and its Applications, 254, 1997, 367–381. Used as a review source for the Smith normal form and invariant factors. ScienceDirect record.