cos — Compute cosine element-wise in MATLAB and RunMat.

y = cos(x) evaluates cosine of each element in x using radian input. Real and complex values are handled element-wise with behavior aligned to MATLAB and RunMat.

Syntax

Y = cos(X)
Y = cos(X, "like", P)

Inputs

Returns

Errors

How cos works

• Operates on scalars, vectors, matrices, and N-D tensors with MATLAB-compatible implicit expansion.
• Logical and integer inputs are promoted to double precision before evaluation so downstream arithmetic matches MATLAB’s numeric tower.
• Complex values use the analytic extension cos(a + bi) = cos(a)cosh(b) - i·sin(a)sinh(b) while propagating NaN/Inf components independently.
• Character arrays are interpreted through their Unicode code points and return dense double arrays that mirror MATLAB’s behaviour.
• Appending 'like', prototype mirrors the prototype’s class and residency (host or GPU), re-uploading the result when a device prototype is supplied.
• Empty inputs and singleton dimensions are preserved without introducing extraneous allocations.

Does RunMat run cos on the GPU?

With RunMat Accelerate active, tensors remain on the device and execute through the provider’s unary_cos hook (or fused elementwise kernels) without leaving GPU memory.

If the provider declines the operation—for example, when only CPU precision is available or the operand type is unsupported—RunMat transparently gathers to the host, computes the result, and reapplies the requested residency rules (including 'like' prototypes).

Fusion planning keeps neighbouring elementwise operators grouped, reducing host↔device transfers even when an intermediate fallback occurs.

GPU memory and residency

A = reshape(0:5, [3 2]);
result = cos(A);        % planner keeps the tensor on device when beneficial

Expected output (after gather(result)):

result =
    1.0000   -0.9899
    0.5403   -0.6536
   -0.4161    0.2837

Examples

Cosine of zero

y = cos(0)

Expected output:

y = 1

Cosine of evenly spaced angles

theta = linspace(0, 2*pi, 5);
values = cos(theta)

Expected output:

values = [1 0 -1 0 1]

Cosine of complex data

z = cos(1 + 2i)

Expected output:

z = 2.0327 - 3.0519i

Cosine on a GPU tensor without manual residency

A = reshape(0:5, [3 2]);
result = cos(A)        % planner keeps the tensor on device when beneficial

Expected output:

result =
    1.0000   -0.9900
    0.5403   -0.6536
   -0.4161    0.2837

Keeping results on the GPU with a 'like' prototype

proto = gpuArray.zeros(1, 1, 'single');
angles = [0 pi/2 pi];
deviceResult = cos(angles, 'like', proto);
result = gather(deviceResult)

Expected output:

result = [1 0 -1]

Matching a host prototype while inputs live on the GPU

G = gpuArray([0 1 2]);
hostLike = cos(G, 'like', zeros(1, 'double'))

Expected output:

hostLike = [1 0.5403 -0.4161]

Using cos with coding agents

Open a RunMat example with live inputs, then ask the agent to explain how cos changes the result.

FAQ

Related Math functions

Open-source implementation

Unlike proprietary runtimes, every RunMat function is open-source. Read exactly how cos is executed, line by line, in Rust.

• View the source for cos in Rust on GitHub
• Learn how the RunMat runtime works
• Found a bug? Open an issue with a minimal reproduction.

About RunMat

RunMat is an open-source runtime that executes MATLAB-syntax code blazing on any GPU. It is licensed under the Apache 2.0 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.

Getting started · Benchmarks · Pricing