class ARROW(T)
inherits Num::Backend::Storage

Constructors

.new(shape : Array(Int), order : Num::OrderType, value : T)

Initialize an ARROW storage from an initial capacity and an initial value, which will fill the buffer

Arguments

• shape : Array(Int) - Shape of the parent Tensor
• order : Array(Int) - Memory layout of the parent Tensor
• value : T - Initial value to populate the buffer

Examples

ARROW.new([10, 10], 3.4)

View source

.new(shape : Array(Int), strides : Array(Int), value : T)

Initialize an ARROW storage from an initial capacity and an initial value, which will fill the buffer

Arguments

• shape : Array(Int) - Shape of the parent Tensor
• strides : Array(Int) - Strides of the parent Tensor
• value : T - Initial value to populate the buffer

Examples

ARROW.new([10, 10], [10, 1], 3.4)

View source

.new(data : Pointer(T), shape : Array(Int), strides : Array(Int))

Initialize an ARROW storage from a hostptr and initial shape. The shape is not required for this storage type, but is needed by other implementations to ensure copy requirements have the right pointer size.

Arguments

• data : Pointer(T) - Existing databuffer for a Tensor
• shape : Array(Int) - Shape of the parent Tensor
• strides : Array(Int) - Strides of the parent Tensor

Examples

a = Pointer(Int32).malloc(10)
s = ARROW.new(a, [5, 2])

View source

.new(shape : Array(Int), order : Num::OrderType)

Initialize an ARROW backed storage from an initial capacity. The data will be filled with zeros

Arguments

• shape : Array(Int) - Shape of the parent Tensor
• order : Array(Int) - Memory layout of the parent Tensor

Examples

CPU.new([2, 3, 4])

View source

.new(shape : Array(Int), strides : Array(Int))

Initialize an ARROW storage from an initial capacity. The data will be filled with zeros

Arguments

• shape : Array(Int) - Shape of the parent Tensor
• strides : Array(Int) - Strides of the parent Tensor

Examples

ARROW(Int32).new([2, 3, 4])

View source

Class methods

.base(dtype : U.class) : ARROW(U).class forall U

Return a generic class of a specific generic type, to allow for explicit return types in functions that return a different storage type than the parent Tensor

Examples

a = ARROW(Float32).new([10])

# Cannot do
# a.class.new ...

a.class.base(Float64).new([10])

View source

Methods

#data : Arrow::NumericArray

Raw Crystal pointer that holds an ARROW(T)s data

View source

#to_hostptr : Pointer(T)

Converts ARROW storage to a crystal pointer

Examples

a = ARROW(Int32).new([3, 3, 2])
a.to_hostptr

View source

#to_unsafe

Returns the raw Arrow::Array associated with an ARROW(T)

View source