Data types¶

dpctl.tensor supports the following data types:

Data Type	Description
`bool`	Boolean (`True` or `False`)
`int8`	An 8-bit signed integer type capable of representing \(v\) subject to \(-2^7 \le v < 2^7\)
`int16`	A 16-bit signed integer type capable of representing \(v\) subject to \(-2^{15} \le v < 2^{15}\)
`int32`	A 32-bit signed integer type capable of representing \(v\) subject to \(-2^{31} \le v < 2^{31}\)
`int64`	A 64-bit signed integer type capable of representing \(v\) subject to \(-2^{63} \le v < 2^{63}\)
`uint8`	An 8-bit unsigned integer type capable of representing \(v\) subject to \(0 \le v < 2^8\)
`uint16`	A 16-bit unsigned integer type capable of representing \(v\) subject to \(0 \le v < 2^{16}\)
`uint32`	A 32-bit unsigned integer type capable of representing \(v\) subject to \(0 \le v < 2^{32}\)
`uint64`	A 64-bit unsigned integer type capable of representing \(v\) subject to \(0 \le v < 2^{64}\)
`float16`	An IEEE-754 half-precision (16-bit) binary floating-point number (see IEEE 754-2019)
`float32`	An IEEE-754 single-precision (32-bit) binary floating-point number (see IEEE 754-2019)
`float64`	An IEEE-754 double-precision (64-bit) binary floating-point number (see IEEE 754-2019)
`complex64`	Single-precision (64-bit) complex floating-point number whose real and imaginary components are IEEE 754 single-precision (32-bit) binary floating-point numbers (see IEEE 754-2019)
`complex128`	Double-precision (128-bit) complex floating-point number whose real and imaginary components are IEEE 754 double-precision (64-bit) binary floating-point numbers (see IEEE 754-2019)

Data type support by array object usm_ndarray depends on capabilities of dpctl.SyclDevice where array is allocated.

Half-precision floating-point type float16 is supported only for devices whose attribute dpctl.SyclDevice.has_aspect_fp16 evaluates to True.

Double-precision floating-point type float64 and double-precision complex floating-point type complex128 are supported only for devices whose attribute dpctl.SyclDevice.has_aspect_fp64 evaluates to True.

If prerequisites are not met, requests to create an instance of an array object for these types will raise an exception.

Data type objects are instances of dtype object, and support equality comparison by implementing special method __eq__().

class dpctl.tensor.dtype¶

Same as numpy.dtype

__eq__()¶: Check if data-type instances are equal.

Default integral data type¶

The default integral data type is int64 for all supported devices.

Default indexing data type¶

The default indexing data type is int64 for all supported devices.

Default real floating-point data type¶

The default real floating-point type depends on the capabilities of device where array is allocated. If the device support double precision floating-point types, the default real floating-point type is float64, otherwise float32.

Make sure to select an appropriately capable device for an application that requires use of double precision floating-point type.

Default complex floating-point data type¶

Like for the default real floating-point type, the default complex floating-point type depends on capabilities of device. If the device support double precision real floating-point types, the default complex floating-point type is complex128, otherwise complex64.

Querying default data types programmatically¶

The data type can be discovered programmatically using Array API inspection functions:

from dpctl
from dpctl import tensor

device = dpctl.select_default_device()
# get default data types for default-selected device
default_types = tensor.__array_namespace_info__().default_dtypes(device)
int_dt = default_types["integral"]
ind_dt = default_types["indexing"]
rfp_dt = default_types["real floating"]
cfp_dt = default_types["complex floating"]

Type promotion rules¶

Type promotion rules govern the behavior of an array library when a function does not have a dedicated implementation for the data type(s) of the input array(s).

In such a case, input arrays may be cast to data types for which a dedicated implementation exists. For example, when sin is applied to array of integral values.

Type promotion rules used in dpctl.tensor are consistent with the Python Array API specification’s type promotion rules for devices that support double precision floating-point type.

For devices that do not support double precision floating-point type, the type promotion rule is truncated by removing nodes corresponding to unsupported data types and edges that lead to them.