repmat — Tile arrays across one or more dimensions using MATLAB-compatible repmat rules.
repmat(A, reps) tiles A according to replication factors in reps. It follows MATLAB-compatible shape rules and supports numeric, logical, char, string, cell, and GPU-backed array inputs.
Syntax
B = repmat(A, r)
B = repmat(A, m, n)
B = repmat(A, d1, d2, ...)Inputs
A- Type
Any- Required
- Yes
- Description
- Input array to replicate.
r- Type
Any- Required
- Yes
- Description
- Replication vector or scalar factor.
m- Type
NumericScalar- Required
- Yes
- Description
- Row replication factor.
n- Type
NumericScalar- Required
- Yes
- Description
- Column replication factor.
dims- Type
NumericScalar- Required
- Variadic
- Description
- Per-dimension replication factors.
Returns
B- Type
Any- Description
- Replicated array tiled by requested replication factors.
Errors
RunMat:repmat:MissingReplicationFactors- When
- No replication factors are provided.
- Message
repmat: replication factors must be specified
RunMat:repmat:InvalidReplicationFactors- When
- Replication factors are non-scalar, non-integer, negative, or otherwise invalid.
- Message
repmat: invalid replication factors
RunMat:repmat:UnsupportedInput- When
- Input type is unsupported for repmat.
- Message
repmat: unsupported input type
RunMat:repmat:Internal- When
- Internal allocation/indexing/provider path fails.
- Message
repmat: internal operation failed
Behavior
repmat(A, m, n) or repmat(A, [m n]) repeats A m times along rows and n times along columns.
A size vector [r1 r2 … rN] can be given as a row or column; trailing dimensions default to 1, letting you replicate only the axes you care about.
Supplying a scalar replication factor (e.g., repmat(A, k)) duplicates non-structure inputs across every stored dimension, ensuring at least the first two dimensions are replicated. Structure arrays retain their existing per-axis rule: the supplied factor applies to the first axis and omitted factors default to one.
Zero replication factors produce empty dimensions; repmat(A, 0, 3) has shape [0, 3*size(A,2)].
Character and cell arrays use row/column tiling with trailing factors at most one; numeric, logical, complex, string, and structure arrays use their existing N-D tiling paths.
Replication factors must be finite, non-negative integers; RunMat raises descriptive errors for negative, fractional, or excessively large sizes.
All eight integer data classes are tiled from authoritative native storage and returned in the same class; typed integer replication factors are decoded structurally without a floating round trip.
Admitted provider tiling keeps the result on the device. A host fallback tiles exact-class values and attempts restoration through the input's owning provider. Explicit gpuArray input cannot silently produce a host result; automatically placed floating-point values may remain on the host when restoration upload is unsupported.
Execution and residency
The runtime uses admitted provider tiling or exact-class host tiling restored to the owning provider. Operational or unphased provider errors are propagated without replay.
You usually do not need to call gpuArray directly. RunMat's planner keeps values on the GPU when it detects that further operations benefit from staying there. Explicit gpuArray calls remain available for compatibility with legacy MATLAB code and when you want to steer residency manually.
Structural data retains its admitted single, double, logical, or native signed/unsigned integer storage, including supported complex arrays. This does not upgrade preceding numerical accuracy or authorize an unsupported carrier.
The first argument is the data operand; remaining arguments, where present, are dimension or order controls. No scalar broadcasting, elementwise/reduction fusion template, or standalone fusion constant ABI is declared.
Examples
Tiling a matrix across rows and columns
A = [1 2; 3 4];
B = repmat(A, 2, 3)Expected output:
B =
1 2 1 2 1 2
3 4 3 4 3 4
1 2 1 2 1 2
3 4 3 4 3 4Using a scalar replication factor
row = 1:4;
Tiled = repmat(row, 3);
size(Tiled)Expected output:
ans =
3 12Replicating into three dimensions
A = reshape(1:6, [1 3 2]);
T = repmat(A, [2 1 4]);
size(T)Expected output:
ans =
2 3 8Repeating logical masks with zero dimensions
mask = logical([1 0 1]);
emptyMask = repmat(mask, 0, 3);
size(emptyMask)Expected output:
ans =
0 9Replicating string scalars into string arrays
name = "runmat";
names = repmat(name, 2, 2)Expected output:
names =
2x2 string array
"runmat" "runmat"
"runmat" "runmat"Replicating data that lives on the GPU
G = gpuArray(magic(3));
T = repmat(G, [2 1]);
result = gather(T)Using repmat with coding agents
Open a RunMat example with live inputs, then ask the agent to explain how repmat changes the result.
Run a small repmat example, explain the result, then change one input and compare the output.
FAQ
Are replication factors required to be integers?⌄
Yes. RunMat follows MATLAB and requires every replication factor to be a non-negative integer. Non-integers raise the descriptive error repmat: replication factor <n> must be an integer.
What happens when I pass a scalar replication factor?⌄
For non-structure inputs, RunMat duplicates every stored dimension, ensuring at least the first two dimensions receive the factor. For structure arrays, the factor applies to the first axis and omitted factors default to one.
Can I request zero replication along a dimension?⌄
Yes. Zero factors produce empty dimensions while preserving the remaining sizes, which is useful when constructing placeholder tensors or short-circuiting loops.
Does repmat work with char, string, or cell arrays?⌄
Yes. Character and cell arrays tile rows and columns with trailing factors at most one. String, numeric, logical, complex, and structure arrays use their existing N-D replication paths.
How does repmat handle GPU tensors today?⌄
Device tiling preserves residency. An admitted host fallback attempts exact-owner restoration; explicit gpuArray intent forbids silent host publication, while automatically placed floating-point values may remain on the host when upload is unsupported.
Does the result reuse backing storage from the input?⌄
No. repmat always creates a new array so modifying the result never mutates the original input.
Can replication overflow memory?⌄
RunMat checks for overflow when multiplying shape dimensions. If the requested size cannot fit in native address space, the builtin raises a descriptive error before attempting allocation.
Does repmat preserve data ordering?⌄
Yes. Column-major ordering is maintained for numeric, logical, string, and complex arrays, while char and cell arrays respect their row-major storage conventions within RunMat.
Related Array functions
Shape
blkdiag · cat · circshift · diag · flip · fliplr · flipud · horzcat · ipermute · kron · permute · repelem · reshape · rot90 · shiftdim · squeeze · toeplitz · tril · triu · vertcat
Accumulation
Sorting Sets
argsort · intersect · ismember · ismembertol · issorted · issortedrows · setdiff · setxor · sort · sortrows · union · unique
Creation
colon · createArray · empty · eye · false · full · inf · int16.empty · int32.empty · int64.empty · int8.empty · linspace · logspace · magic · meshgrid · nan · ndgrid · nonzeros · ones · peaks · rand · randi · randn · randperm · range · sparse · spdiags · speye · spones · sprand · true · uint16.empty · uint32.empty · uint64.empty · uint8.empty · zeros
Combinatorics
Binning
Grouping
Source you can read
Every RunMat function's source is public. Read exactly how repmat is executed, line by line, in Rust.
- View the source for repmat in Rust on GitHub
- Learn how the RunMat runtime works
- Found a bug? Open an issue with a minimal reproduction.
About RunMat
RunMat is a runtime that executes MATLAB-syntax code fast on any GPU. The runtime is free for individuals, including freelancers; students and faculty at an academic institution, for teaching or research there; contributors to an open-source project, for work on that project; organizations with 250 or fewer people and US$10M or less in annual revenue; and anyone, for non-production use. See runmat.com/license.
- RunMat automatically optimizes your math for GPU execution on Apple, Nvidia, and AMD hardware. No code changes needed. Simulations that took hours now take minutes.
- Start running code in seconds. RunMat runs in the browser, on the desktop, or from the CLI. No license server, no IT ticket.