Multiply matrices up to 5×5 with dimension compatibility check, partial products breakdown, Strassen complexity comparison, and result magnitude visualization.
Matrix multiplication is one of the most important operations in all of mathematics. It combines two matrices to produce a new matrix where each element is computed as the dot product of a row from the first matrix and a column from the second. Unlike addition, matrix multiplication requires dimension compatibility: if A is m×n, then B must be n×p, and the result AB will be m×p.
This calculator lets you multiply matrices up to 5×5, with full control over dimensions. Enter your matrices, and the tool shows the result matrix with color-coded values and row magnitude bars, a complete partial products breakdown showing exactly how each element was computed, and Frobenius norms for all three matrices.
The partial products table is particularly valuable for learning: for each result element C(i,j), it shows every term A(i,k)×B(k,j) that contributes to the sum. This makes the row-by-column dot product pattern immediately visible.
The calculator also includes a complexity comparison table covering naïve O(n³) multiplication, Strassen's O(n^2.807) algorithm, and more advanced theoretical bounds. For small matrices like those handled here, the naïve algorithm is optimal — Strassen only becomes advantageous around n = 64 or larger. Nevertheless, understanding these bounds gives insight into one of the most important problems in computational mathematics.
Remember: matrix multiplication is not commutative (AB ≠ BA in general), but it is associative ((AB)C = A(BC)) and distributes over addition (A(B+C) = AB + AC).
Multiplying two 3×3 matrices by hand requires computing 9 dot products, each with 3 multiply-add pairs — 27 multiplications and 18 additions total. A single index error produces a wrong row or column in the result. This calculator computes the product instantly, shows every partial product that contributes to each result element, and visualizes row magnitudes. It is indispensable for students learning the row-times-column rule, for verifying hand calculations, and for anyone exploring how dimension compatibility works.
(AB)ᵢⱼ = Σₖ₌₁ⁿ aᵢₖ · bₖⱼ — each result element is the dot product of row i of A and column j of B
Result: [[58,64],[139,154]]
C(1,1) = 1×7 + 2×9 + 3×11 = 7+18+33 = 58; C(1,2) = 1×8 + 2×10 + 3×12 = 8+20+36 = 64; etc.
Matrix multiplication is defined as (AB)ᵢⱼ = Σₖ aᵢₖbₖⱼ — each element of the result is the **dot product** of row i of A with column j of B. This requires the number of columns of A to equal the number of rows of B (the “inner dimensions” must match). An m×n matrix times an n×p matrix produces an m×p matrix. The operation is **associative** (A(BC) = (AB)C) and **distributive** (A(B+C) = AB + AC), but crucially **not commutative** — AB ≠ BA in general, and the two products may not even have the same dimensions.
Naïve matrix multiplication for two n×n matrices requires n³ scalar multiplications and n²(n−1) additions, giving **O(n³)** complexity. Strassen’s algorithm reduces this to O(n^2.807) by cleverly rewriting 2×2 block multiplication with 7 sub-multiplications instead of 8. The current theoretical best is O(n^2.371), though practical implementations rarely go beyond Strassen due to overhead and numerical stability concerns. For sparse matrices, specialized algorithms exploit the zero structure for significant speedups.
Matrix multiplication is the computational backbone of **computer graphics** (applying rotation, scaling, and projection to 3D vertices via 4×4 matrices), **machine learning** (forward and backward passes in neural networks are chains of matrix multiplications), **physics** (composing quantum operators, transforming coordinate systems), and **economics** (Leontief input-output models). Understanding how and why it works — and why order matters — is fundamental to every field that uses linear algebra.
Each element of the result is a dot product of a row from A and a column from B. Rows of A have as many elements as A has columns, and columns of B have as many as B has rows. These must be equal for the dot product to be defined.
No. AB and BA may not even have the same dimensions. Even when both are defined and have the same size, AB ≠ BA in general. Commutativity only holds for special cases like scalar multiples of the identity.
Strassen's algorithm multiplies 2×2 block matrices using 7 multiplications instead of 8, applied recursively. This reduces the complexity from O(n³) to O(n^2.807), but only becomes practical for large matrices (n > ~64) due to overhead.
For an m×n matrix times an n×p matrix: m×n×p scalar multiplications and m×(n−1)×p additions, giving O(mnp) total. Use this as a practical reminder before finalizing the result.
For result element C(i,j), the partial products are the individual terms a(i,k)×b(k,j) for k = 1 to n. The sum of all partial products gives the final value C(i,j).
Yes! A can be m×n and B can be n×p with any m, n, p ≥ 1. The only requirement is that the number of columns of A equals the number of rows of B.