6701964608

Committed 31 Oct 2023 04:14AM UTC coverage: 78.586% (+0.3%) from 78.255%

Build # 6701964608

Build Type

Pull #1178

github

Committed by

web-flow

Commit Message

Merge d18b8a4d1 into a030c8515

Pull Request Pull Request #1178: An experimental kernel dispatcher for numba_dpex.kernel decorator

Run Details

1132 of 1705 branches covered (0.0%)

Branch coverage included in aggregate %.

363 of 416 new or added lines in 13 files covered. (87.26%)

10 existing lines in 1 file now uncovered.

5015 of 6117 relevant lines covered (81.98%)

0.82 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

26.47

/numba_dpex/experimental/decorators.py

# SPDX-FileCopyrightText: 2023 Intel Corporation
#
# SPDX-License-Identifier: Apache-2.0

import inspect

from numba.core import sigutils

from .kernel_dispatcher import KernelDispatcher


def kernel(func_or_sig=None, debug=False, cache=False, **options):
    """A decorator to define a kernel function.

    A kernel function is conceptually equivalent to a SYCL kernel function, and
    gets compiled into either an OpenCL or a LevelZero SPIR-V binary kernel.
    A kernel decorated Python function has the following restrictions:

        * The function can not return any value.
        * All array arguments passed to a kernel should adhere to compute
          follows data programming model.
    """
    # FIXME: The options need to be evaluated and checked here like it is
    # done in numba.core.decorators.jit

    def _kernel_dispatcher(pyfunc, sigs=None):
        return KernelDispatcher(
            pyfunc=pyfunc,
            debug_flags=debug,
            enable_cache=cache,
            specialization_sigs=sigs,
            targetoptions=options,
        )

    if func_or_sig is None:
        return _kernel_dispatcher

    if isinstance(func_or_sig, str):
        raise NotImplementedError(
            "Specifying signatures as string is not yet supported by numba-dpex"
        )

    if isinstance(func_or_sig, list) or sigutils.is_signature(func_or_sig):
        # String signatures are not supported as passing usm_ndarray type as
        # a string is not possible. Numba's sigutils relies on the type being
        # available in Numba's `types.__dict__` and dpex types are not
        # registered there yet.
        if isinstance(func_or_sig, list):
            for sig in func_or_sig:
                if isinstance(sig, str):
                    raise NotImplementedError(
                        "Specifying signatures as string is not yet supported "
                        "by numba-dpex"
                    )
        # Specialized signatures can either be a single signature or a list.
        # In case only one signature is provided convert it to a list
        if not isinstance(func_or_sig, list):
            func_or_sig = [func_or_sig]

        def _specialized_kernel_dispatcher(pyfunc):
            return KernelDispatcher(
                pyfunc=pyfunc,
                debug_flags=debug,
                enable_cache=cache,
                specialization_sigs=func_or_sig,
            )

        return _specialized_kernel_dispatcher
    func = func_or_sig
    if not inspect.isfunction(func):
        raise ValueError(
            "Argument passed to the kernel decorator is neither a "
            "function object, nor a signature. If you are trying to "
            "specialize the kernel that takes a single argument, specify "
            "the return type as void explicitly."
        )
    return _kernel_dispatcher(func)

1	# SPDX-FileCopyrightText: 2023 Intel Corporation
2	#
3	# SPDX-License-Identifier: Apache-2.0
4
5	import inspect	1✔
6
7	from numba.core import sigutils	1✔
8
9	from .kernel_dispatcher import KernelDispatcher	1✔
10
11
12	def kernel(func_or_sig=None, debug=False, cache=False, **options):	1✔
13	"""A decorator to define a kernel function.
14
15	A kernel function is conceptually equivalent to a SYCL kernel function, and
16	gets compiled into either an OpenCL or a LevelZero SPIR-V binary kernel.
17	A kernel decorated Python function has the following restrictions:
18
19	* The function can not return any value.
20	* All array arguments passed to a kernel should adhere to compute
21	follows data programming model.
22	"""
23	# FIXME: The options need to be evaluated and checked here like it is
24	# done in numba.core.decorators.jit
25
26	def _kernel_dispatcher(pyfunc, sigs=None):	1✔
27	return KernelDispatcher(	1✔
28	pyfunc=pyfunc,
29	debug_flags=debug,
30	enable_cache=cache,
31	specialization_sigs=sigs,
32	targetoptions=options,
33	)
34
35	if func_or_sig is None:	1!
36	return _kernel_dispatcher	1✔
37
NEW 38	if isinstance(func_or_sig, str):	×
39	raise NotImplementedError(
40	"Specifying signatures as string is not yet supported by numba-dpex"
41	)
42
NEW 43	if isinstance(func_or_sig, list) or sigutils.is_signature(func_or_sig):	×
44	# String signatures are not supported as passing usm_ndarray type as
45	# a string is not possible. Numba's sigutils relies on the type being
46	# available in Numba's `types.__dict__` and dpex types are not
47	# registered there yet.
NEW 48	if isinstance(func_or_sig, list):	×
NEW 49	for sig in func_or_sig:	×
NEW 50	if isinstance(sig, str):	×
51	raise NotImplementedError(
52	"Specifying signatures as string is not yet supported "
53	"by numba-dpex"
54	)
55	# Specialized signatures can either be a single signature or a list.
56	# In case only one signature is provided convert it to a list
NEW 57	if not isinstance(func_or_sig, list):	×
NEW 58	func_or_sig = [func_or_sig]	×
59
NEW 60	def _specialized_kernel_dispatcher(pyfunc):	×
NEW 61	return KernelDispatcher(	×
62	pyfunc=pyfunc,
63	debug_flags=debug,
64	enable_cache=cache,
65	specialization_sigs=func_or_sig,
66	)
67
NEW 68	return _specialized_kernel_dispatcher	×
NEW 69	func = func_or_sig	×
NEW 70	if not inspect.isfunction(func):	×
NEW 71	raise ValueError(	×
72	"Argument passed to the kernel decorator is neither a "
73	"function object, nor a signature. If you are trying to "
74	"specialize the kernel that takes a single argument, specify "
75	"the return type as void explicitly."
76	)
NEW 77	return _kernel_dispatcher(func)	×

IntelPython / numba-dpex / 6701964608

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous