numba_dpex.core.targets.dpjit_target#

Defines the target and typing contexts for numba_dpex’s dpjit decorator.

Overview#

Classes#
`Dpex`	Mark the target as CPU.
`DpexTypingContext`	Custom typing context to support dpjit compilation.
`DpexTargetContext`	Changes BaseContext calling convention

Attributes#
`DPEX_TARGET_NAME`	-
`dpex_function_registry`	-

Classes#

class Dpex#

Bases: numba.core.target_extension.CPU

Mark the target as CPU.

class DpexTypingContext#

Bases: numba.core.typing.Context

Custom typing context to support dpjit compilation.

Overview

Methods#
`load_additional_registries`()	Register dpjit specific functions like dpnp ufuncs.

Members

load_additional_registries()#: Register dpjit specific functions like dpnp ufuncs.

class DpexTargetContext(typingctx, target=DPEX_TARGET_NAME)#

Bases: numba.core.cpu.CPUContext

Changes BaseContext calling convention

Overview

Methods#
`init`()	For subclasses to add initializer
`dpexrt`()	-
`load_additional_registries`()	Load dpjit-specific registries.
`get_ufunc_info`(ufunc_key)	Get the ufunc implementation for a given ufunc object.

Members

init()#: For subclasses to add initializer

dpexrt()#

load_additional_registries()#: Load dpjit-specific registries.

get_ufunc_info(ufunc_key)#

Get the ufunc implementation for a given ufunc object.

The default implementation in BaseContext always raises a NotImplementedError exception. Subclasses may raise KeyError to signal that the given ufunc_key is not available.

Parameters:: ufunc_key (NumPy ufunc) –
Returns:: res – A mapping of a NumPy ufunc type signature to a lower-level implementation.
Return type:: dict[str, callable]

Attributes#

DPEX_TARGET_NAME = 'dpex'#

dpex_function_registry#