Random number generation

Overview

Multik provides the mk.rand() family of functions for creating arrays filled with random numbers. All values follow a uniform distribution. You can control the range, seed, and random generator.

Function	Description
`mk.rand<T>(dim0, ...)`	Array with default random range for type `T`.
`mk.rand<T, D>(from, until, dims)`	Array with values in `[from, until)`.
`mk.rand<T, D>(seed, from, until, ...)`	Reproducible random array with a seed.
`mk.rand<T, D>(gen, from, until, ...)`	Random array using a custom `kotlin.random.Random`.

Default ranges by type

Type	Range
`Int`	`[Int.MIN_VALUE, Int.MAX_VALUE)`
`Long`	`[Long.MIN_VALUE, Long.MAX_VALUE)`
`Float`	`[0.0, 1.0)`
`Double`	`[0.0, 1.0)`

Example

val a = mk.rand<Double>(3, 3)                     // 3×3, values in [0.0, 1.0)
val b = mk.rand<Int, D1>(from = 1, until = 100, 10)  // 10 ints in [1, 100)
val c = mk.rand<Double, D2>(42, 0.0, 1.0, 3, 3)  // seeded 3×3

Pitfalls

Complex types (ComplexFloat, ComplexDouble) are not supported. Use d1array/d2array with a custom init lambda for complex random arrays.
The from bound is inclusive, the until bound is exclusive.

19 February 2026