numba-cuda 0.8.1__py3-none-any.whl → 0.10.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- _numba_cuda_redirector.py +17 -13
- numba_cuda/VERSION +1 -1
- numba_cuda/_version.py +4 -1
- numba_cuda/numba/cuda/__init__.py +6 -2
- numba_cuda/numba/cuda/api.py +129 -86
- numba_cuda/numba/cuda/api_util.py +3 -3
- numba_cuda/numba/cuda/args.py +12 -16
- numba_cuda/numba/cuda/cg.py +6 -6
- numba_cuda/numba/cuda/codegen.py +74 -43
- numba_cuda/numba/cuda/compiler.py +246 -114
- numba_cuda/numba/cuda/cpp_function_wrappers.cu +1 -2
- numba_cuda/numba/cuda/cuda_bf16.py +5155 -0
- numba_cuda/numba/cuda/cuda_paths.py +293 -99
- numba_cuda/numba/cuda/cudadecl.py +93 -79
- numba_cuda/numba/cuda/cudadrv/__init__.py +3 -1
- numba_cuda/numba/cuda/cudadrv/devicearray.py +185 -135
- numba_cuda/numba/cuda/cudadrv/devices.py +16 -11
- numba_cuda/numba/cuda/cudadrv/driver.py +460 -297
- numba_cuda/numba/cuda/cudadrv/drvapi.py +241 -207
- numba_cuda/numba/cuda/cudadrv/dummyarray.py +66 -54
- numba_cuda/numba/cuda/cudadrv/enums.py +1 -1
- numba_cuda/numba/cuda/cudadrv/error.py +6 -2
- numba_cuda/numba/cuda/cudadrv/libs.py +67 -63
- numba_cuda/numba/cuda/cudadrv/linkable_code.py +27 -3
- numba_cuda/numba/cuda/cudadrv/mappings.py +16 -14
- numba_cuda/numba/cuda/cudadrv/nvrtc.py +146 -30
- numba_cuda/numba/cuda/cudadrv/nvvm.py +296 -161
- numba_cuda/numba/cuda/cudadrv/rtapi.py +1 -1
- numba_cuda/numba/cuda/cudadrv/runtime.py +20 -8
- numba_cuda/numba/cuda/cudaimpl.py +296 -275
- numba_cuda/numba/cuda/cudamath.py +1 -1
- numba_cuda/numba/cuda/debuginfo.py +99 -7
- numba_cuda/numba/cuda/decorators.py +87 -45
- numba_cuda/numba/cuda/descriptor.py +1 -1
- numba_cuda/numba/cuda/device_init.py +68 -18
- numba_cuda/numba/cuda/deviceufunc.py +143 -98
- numba_cuda/numba/cuda/dispatcher.py +300 -213
- numba_cuda/numba/cuda/errors.py +13 -10
- numba_cuda/numba/cuda/extending.py +55 -1
- numba_cuda/numba/cuda/include/11/cuda_bf16.h +3749 -0
- numba_cuda/numba/cuda/include/11/cuda_bf16.hpp +2683 -0
- numba_cuda/numba/cuda/{cuda_fp16.h → include/11/cuda_fp16.h} +1090 -927
- numba_cuda/numba/cuda/{cuda_fp16.hpp → include/11/cuda_fp16.hpp} +468 -319
- numba_cuda/numba/cuda/include/12/cuda_bf16.h +5118 -0
- numba_cuda/numba/cuda/include/12/cuda_bf16.hpp +3865 -0
- numba_cuda/numba/cuda/include/12/cuda_fp16.h +5363 -0
- numba_cuda/numba/cuda/include/12/cuda_fp16.hpp +3483 -0
- numba_cuda/numba/cuda/initialize.py +5 -3
- numba_cuda/numba/cuda/intrinsic_wrapper.py +0 -39
- numba_cuda/numba/cuda/intrinsics.py +203 -28
- numba_cuda/numba/cuda/kernels/reduction.py +13 -13
- numba_cuda/numba/cuda/kernels/transpose.py +3 -6
- numba_cuda/numba/cuda/libdevice.py +317 -317
- numba_cuda/numba/cuda/libdeviceimpl.py +3 -2
- numba_cuda/numba/cuda/locks.py +16 -0
- numba_cuda/numba/cuda/lowering.py +43 -0
- numba_cuda/numba/cuda/mathimpl.py +62 -57
- numba_cuda/numba/cuda/models.py +1 -5
- numba_cuda/numba/cuda/nvvmutils.py +103 -88
- numba_cuda/numba/cuda/printimpl.py +9 -5
- numba_cuda/numba/cuda/random.py +46 -36
- numba_cuda/numba/cuda/reshape_funcs.cu +1 -1
- numba_cuda/numba/cuda/runtime/__init__.py +1 -1
- numba_cuda/numba/cuda/runtime/memsys.cu +1 -1
- numba_cuda/numba/cuda/runtime/memsys.cuh +1 -1
- numba_cuda/numba/cuda/runtime/nrt.cu +3 -3
- numba_cuda/numba/cuda/runtime/nrt.py +48 -43
- numba_cuda/numba/cuda/simulator/__init__.py +22 -12
- numba_cuda/numba/cuda/simulator/api.py +38 -22
- numba_cuda/numba/cuda/simulator/compiler.py +2 -2
- numba_cuda/numba/cuda/simulator/cudadrv/__init__.py +8 -2
- numba_cuda/numba/cuda/simulator/cudadrv/devicearray.py +63 -55
- numba_cuda/numba/cuda/simulator/cudadrv/devices.py +13 -11
- numba_cuda/numba/cuda/simulator/cudadrv/driver.py +5 -5
- numba_cuda/numba/cuda/simulator/cudadrv/drvapi.py +2 -2
- numba_cuda/numba/cuda/simulator/cudadrv/libs.py +1 -1
- numba_cuda/numba/cuda/simulator/cudadrv/nvvm.py +3 -3
- numba_cuda/numba/cuda/simulator/cudadrv/runtime.py +3 -3
- numba_cuda/numba/cuda/simulator/kernel.py +43 -34
- numba_cuda/numba/cuda/simulator/kernelapi.py +31 -26
- numba_cuda/numba/cuda/simulator/reduction.py +1 -0
- numba_cuda/numba/cuda/simulator/vector_types.py +13 -9
- numba_cuda/numba/cuda/simulator_init.py +2 -4
- numba_cuda/numba/cuda/stubs.py +134 -108
- numba_cuda/numba/cuda/target.py +92 -47
- numba_cuda/numba/cuda/testing.py +24 -19
- numba_cuda/numba/cuda/tests/__init__.py +14 -12
- numba_cuda/numba/cuda/tests/cudadrv/test_array_attr.py +16 -17
- numba_cuda/numba/cuda/tests/cudadrv/test_context_stack.py +7 -7
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_array_slicing.py +73 -54
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_auto_context.py +1 -1
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_devicerecord.py +48 -50
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_driver.py +47 -29
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_libraries.py +3 -3
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_memory.py +19 -19
- numba_cuda/numba/cuda/tests/cudadrv/test_cuda_ndarray.py +108 -103
- numba_cuda/numba/cuda/tests/cudadrv/test_deallocations.py +20 -11
- numba_cuda/numba/cuda/tests/cudadrv/test_detect.py +20 -17
- numba_cuda/numba/cuda/tests/cudadrv/test_emm_plugins.py +8 -6
- numba_cuda/numba/cuda/tests/cudadrv/test_events.py +1 -1
- numba_cuda/numba/cuda/tests/cudadrv/test_host_alloc.py +8 -7
- numba_cuda/numba/cuda/tests/cudadrv/test_init.py +13 -13
- numba_cuda/numba/cuda/tests/cudadrv/test_inline_ptx.py +12 -9
- numba_cuda/numba/cuda/tests/cudadrv/test_linker.py +36 -31
- numba_cuda/numba/cuda/tests/cudadrv/test_managed_alloc.py +8 -7
- numba_cuda/numba/cuda/tests/cudadrv/test_module_callbacks.py +294 -0
- numba_cuda/numba/cuda/tests/cudadrv/test_mvc.py +10 -7
- numba_cuda/numba/cuda/tests/cudadrv/test_nvjitlink.py +24 -15
- numba_cuda/numba/cuda/tests/cudadrv/test_nvvm_driver.py +43 -41
- numba_cuda/numba/cuda/tests/cudadrv/test_pinned.py +4 -5
- numba_cuda/numba/cuda/tests/cudadrv/test_profiler.py +2 -2
- numba_cuda/numba/cuda/tests/cudadrv/test_ptds.py +28 -17
- numba_cuda/numba/cuda/tests/cudadrv/test_reset_device.py +1 -2
- numba_cuda/numba/cuda/tests/cudadrv/test_runtime.py +22 -14
- numba_cuda/numba/cuda/tests/cudadrv/test_select_device.py +1 -1
- numba_cuda/numba/cuda/tests/cudadrv/test_streams.py +4 -3
- numba_cuda/numba/cuda/tests/cudapy/cache_usecases.py +10 -4
- numba_cuda/numba/cuda/tests/cudapy/cache_with_cpu_usecases.py +1 -0
- numba_cuda/numba/cuda/tests/cudapy/extensions_usecases.py +10 -7
- numba_cuda/numba/cuda/tests/cudapy/jitlink.ptx +0 -2
- numba_cuda/numba/cuda/tests/cudapy/recursion_usecases.py +1 -0
- numba_cuda/numba/cuda/tests/cudapy/test_alignment.py +6 -5
- numba_cuda/numba/cuda/tests/cudapy/test_array.py +52 -42
- numba_cuda/numba/cuda/tests/cudapy/test_array_args.py +5 -6
- numba_cuda/numba/cuda/tests/cudapy/test_array_methods.py +1 -1
- numba_cuda/numba/cuda/tests/cudapy/test_atomics.py +501 -304
- numba_cuda/numba/cuda/tests/cudapy/test_bfloat16_bindings.py +257 -0
- numba_cuda/numba/cuda/tests/cudapy/test_blackscholes.py +59 -23
- numba_cuda/numba/cuda/tests/cudapy/test_boolean.py +3 -3
- numba_cuda/numba/cuda/tests/cudapy/test_caching.py +50 -37
- numba_cuda/numba/cuda/tests/cudapy/test_casting.py +29 -24
- numba_cuda/numba/cuda/tests/cudapy/test_cffi.py +11 -6
- numba_cuda/numba/cuda/tests/cudapy/test_compiler.py +84 -50
- numba_cuda/numba/cuda/tests/cudapy/test_complex.py +144 -73
- numba_cuda/numba/cuda/tests/cudapy/test_complex_kernel.py +2 -2
- numba_cuda/numba/cuda/tests/cudapy/test_const_string.py +37 -27
- numba_cuda/numba/cuda/tests/cudapy/test_constmem.py +43 -45
- numba_cuda/numba/cuda/tests/cudapy/test_cooperative_groups.py +21 -14
- numba_cuda/numba/cuda/tests/cudapy/test_cuda_array_interface.py +60 -55
- numba_cuda/numba/cuda/tests/cudapy/test_cuda_jit_no_types.py +3 -2
- numba_cuda/numba/cuda/tests/cudapy/test_datetime.py +26 -22
- numba_cuda/numba/cuda/tests/cudapy/test_debug.py +29 -27
- numba_cuda/numba/cuda/tests/cudapy/test_debuginfo.py +77 -28
- numba_cuda/numba/cuda/tests/cudapy/test_device_func.py +52 -45
- numba_cuda/numba/cuda/tests/cudapy/test_dispatcher.py +55 -43
- numba_cuda/numba/cuda/tests/cudapy/test_enums.py +24 -7
- numba_cuda/numba/cuda/tests/cudapy/test_errors.py +30 -15
- numba_cuda/numba/cuda/tests/cudapy/test_exception.py +11 -12
- numba_cuda/numba/cuda/tests/cudapy/test_extending.py +21 -12
- numba_cuda/numba/cuda/tests/cudapy/test_fastmath.py +77 -66
- numba_cuda/numba/cuda/tests/cudapy/test_forall.py +5 -3
- numba_cuda/numba/cuda/tests/cudapy/test_freevar.py +5 -3
- numba_cuda/numba/cuda/tests/cudapy/test_frexp_ldexp.py +1 -1
- numba_cuda/numba/cuda/tests/cudapy/test_globals.py +3 -5
- numba_cuda/numba/cuda/tests/cudapy/test_gufunc.py +144 -126
- numba_cuda/numba/cuda/tests/cudapy/test_gufunc_scalar.py +23 -18
- numba_cuda/numba/cuda/tests/cudapy/test_gufunc_scheduling.py +16 -22
- numba_cuda/numba/cuda/tests/cudapy/test_idiv.py +1 -3
- numba_cuda/numba/cuda/tests/cudapy/test_inline.py +59 -0
- numba_cuda/numba/cuda/tests/cudapy/test_inspect.py +29 -20
- numba_cuda/numba/cuda/tests/cudapy/test_intrinsics.py +147 -99
- numba_cuda/numba/cuda/tests/cudapy/test_ipc.py +50 -36
- numba_cuda/numba/cuda/tests/cudapy/test_iterators.py +1 -2
- numba_cuda/numba/cuda/tests/cudapy/test_lang.py +4 -4
- numba_cuda/numba/cuda/tests/cudapy/test_laplace.py +7 -7
- numba_cuda/numba/cuda/tests/cudapy/test_libdevice.py +24 -20
- numba_cuda/numba/cuda/tests/cudapy/test_lineinfo.py +36 -31
- numba_cuda/numba/cuda/tests/cudapy/test_localmem.py +13 -13
- numba_cuda/numba/cuda/tests/cudapy/test_mandel.py +13 -6
- numba_cuda/numba/cuda/tests/cudapy/test_math.py +83 -66
- numba_cuda/numba/cuda/tests/cudapy/test_matmul.py +1 -3
- numba_cuda/numba/cuda/tests/cudapy/test_minmax.py +19 -58
- numba_cuda/numba/cuda/tests/cudapy/test_montecarlo.py +4 -4
- numba_cuda/numba/cuda/tests/cudapy/test_multigpu.py +9 -7
- numba_cuda/numba/cuda/tests/cudapy/test_multiprocessing.py +9 -8
- numba_cuda/numba/cuda/tests/cudapy/test_multithreads.py +12 -10
- numba_cuda/numba/cuda/tests/cudapy/test_nondet.py +1 -1
- numba_cuda/numba/cuda/tests/cudapy/test_operator.py +180 -96
- numba_cuda/numba/cuda/tests/cudapy/test_optimization.py +5 -5
- numba_cuda/numba/cuda/tests/cudapy/test_overload.py +37 -18
- numba_cuda/numba/cuda/tests/cudapy/test_powi.py +7 -7
- numba_cuda/numba/cuda/tests/cudapy/test_print.py +9 -7
- numba_cuda/numba/cuda/tests/cudapy/test_py2_div_issue.py +1 -1
- numba_cuda/numba/cuda/tests/cudapy/test_random.py +15 -10
- numba_cuda/numba/cuda/tests/cudapy/test_record_dtype.py +88 -87
- numba_cuda/numba/cuda/tests/cudapy/test_recursion.py +12 -10
- numba_cuda/numba/cuda/tests/cudapy/test_reduction.py +26 -11
- numba_cuda/numba/cuda/tests/cudapy/test_retrieve_autoconverted_arrays.py +7 -10
- numba_cuda/numba/cuda/tests/cudapy/test_serialize.py +4 -6
- numba_cuda/numba/cuda/tests/cudapy/test_slicing.py +1 -1
- numba_cuda/numba/cuda/tests/cudapy/test_sm.py +10 -9
- numba_cuda/numba/cuda/tests/cudapy/test_sm_creation.py +62 -43
- numba_cuda/numba/cuda/tests/cudapy/test_stream_api.py +7 -3
- numba_cuda/numba/cuda/tests/cudapy/test_sync.py +7 -5
- numba_cuda/numba/cuda/tests/cudapy/test_transpose.py +18 -11
- numba_cuda/numba/cuda/tests/cudapy/test_ufuncs.py +111 -88
- numba_cuda/numba/cuda/tests/cudapy/test_userexc.py +2 -3
- numba_cuda/numba/cuda/tests/cudapy/test_vector_type.py +305 -130
- numba_cuda/numba/cuda/tests/cudapy/test_vectorize.py +33 -36
- numba_cuda/numba/cuda/tests/cudapy/test_vectorize_complex.py +5 -5
- numba_cuda/numba/cuda/tests/cudapy/test_vectorize_decor.py +16 -12
- numba_cuda/numba/cuda/tests/cudapy/test_vectorize_device.py +7 -7
- numba_cuda/numba/cuda/tests/cudapy/test_vectorize_scalar_arg.py +6 -7
- numba_cuda/numba/cuda/tests/cudapy/test_warning.py +31 -29
- numba_cuda/numba/cuda/tests/cudapy/test_warp_ops.py +81 -30
- numba_cuda/numba/cuda/tests/cudasim/test_cudasim_issues.py +19 -13
- numba_cuda/numba/cuda/tests/data/jitlink.cu +1 -1
- numba_cuda/numba/cuda/tests/data/jitlink.ptx +0 -2
- numba_cuda/numba/cuda/tests/doc_examples/test_cg.py +15 -8
- numba_cuda/numba/cuda/tests/doc_examples/test_cpu_gpu_compat.py +4 -7
- numba_cuda/numba/cuda/tests/doc_examples/test_ffi.py +14 -9
- numba_cuda/numba/cuda/tests/doc_examples/test_laplace.py +22 -18
- numba_cuda/numba/cuda/tests/doc_examples/test_matmul.py +7 -4
- numba_cuda/numba/cuda/tests/doc_examples/test_montecarlo.py +2 -0
- numba_cuda/numba/cuda/tests/doc_examples/test_random.py +8 -4
- numba_cuda/numba/cuda/tests/doc_examples/test_reduction.py +2 -1
- numba_cuda/numba/cuda/tests/doc_examples/test_sessionize.py +94 -19
- numba_cuda/numba/cuda/tests/doc_examples/test_vecadd.py +2 -2
- numba_cuda/numba/cuda/tests/nocuda/test_dummyarray.py +91 -62
- numba_cuda/numba/cuda/tests/nocuda/test_function_resolution.py +14 -5
- numba_cuda/numba/cuda/tests/nocuda/test_import.py +25 -25
- numba_cuda/numba/cuda/tests/nocuda/test_library_lookup.py +40 -40
- numba_cuda/numba/cuda/tests/nocuda/test_nvvm.py +12 -10
- numba_cuda/numba/cuda/tests/nrt/test_nrt.py +16 -20
- numba_cuda/numba/cuda/tests/nrt/test_nrt_refct.py +12 -10
- numba_cuda/numba/cuda/tests/test_binary_generation/generate_raw_ltoir.py +2 -2
- numba_cuda/numba/cuda/types.py +5 -2
- numba_cuda/numba/cuda/ufuncs.py +382 -362
- numba_cuda/numba/cuda/utils.py +2 -2
- numba_cuda/numba/cuda/vector_types.py +5 -3
- numba_cuda/numba/cuda/vectorizers.py +38 -33
- {numba_cuda-0.8.1.dist-info → numba_cuda-0.10.0.dist-info}/METADATA +1 -1
- numba_cuda-0.10.0.dist-info/RECORD +263 -0
- {numba_cuda-0.8.1.dist-info → numba_cuda-0.10.0.dist-info}/WHEEL +1 -1
- numba_cuda-0.8.1.dist-info/RECORD +0 -251
- {numba_cuda-0.8.1.dist-info → numba_cuda-0.10.0.dist-info}/licenses/LICENSE +0 -0
- {numba_cuda-0.8.1.dist-info → numba_cuda-0.10.0.dist-info}/top_level.txt +0 -0
@@ -21,6 +21,7 @@ class CudaRuntimeAPIError(CudaRuntimeError):
|
|
21
21
|
"""
|
22
22
|
Raised when there is an error accessing a C API from the CUDA Runtime.
|
23
23
|
"""
|
24
|
+
|
24
25
|
def __init__(self, code, msg):
|
25
26
|
self.code = code
|
26
27
|
self.msg = msg
|
@@ -44,11 +45,13 @@ class Runtime:
|
|
44
45
|
_logger = make_logger()
|
45
46
|
|
46
47
|
if config.DISABLE_CUDA:
|
47
|
-
msg = (
|
48
|
-
|
49
|
-
|
48
|
+
msg = (
|
49
|
+
"CUDA is disabled due to setting NUMBA_DISABLE_CUDA=1 "
|
50
|
+
"in the environment, or because CUDA is unsupported on "
|
51
|
+
"32-bit systems."
|
52
|
+
)
|
50
53
|
raise CudaSupportError(msg)
|
51
|
-
self.lib = open_cudalib(
|
54
|
+
self.lib = open_cudalib("cudart")
|
52
55
|
|
53
56
|
self.is_initialized = True
|
54
57
|
|
@@ -76,9 +79,10 @@ class Runtime:
|
|
76
79
|
def _wrap_api_call(self, fname, libfn):
|
77
80
|
@functools.wraps(libfn)
|
78
81
|
def safe_cuda_api_call(*args):
|
79
|
-
_logger.debug(
|
82
|
+
_logger.debug("call runtime api: %s", libfn.__name__)
|
80
83
|
retcode = libfn(*args)
|
81
84
|
self._check_error(fname, retcode)
|
85
|
+
|
82
86
|
return safe_cuda_api_call
|
83
87
|
|
84
88
|
def _check_error(self, fname, retcode):
|
@@ -125,11 +129,19 @@ class Runtime:
|
|
125
129
|
def supported_versions(self):
|
126
130
|
"""A tuple of all supported CUDA toolkit versions. Versions are given in
|
127
131
|
the form ``(major_version, minor_version)``."""
|
128
|
-
if sys.platform not in (
|
132
|
+
if sys.platform not in ("linux", "win32") or config.MACHINE_BITS != 64:
|
129
133
|
# Only 64-bit Linux and Windows are supported
|
130
134
|
return ()
|
131
|
-
return (
|
132
|
-
|
135
|
+
return (
|
136
|
+
(11, 0),
|
137
|
+
(11, 1),
|
138
|
+
(11, 2),
|
139
|
+
(11, 3),
|
140
|
+
(11, 4),
|
141
|
+
(11, 5),
|
142
|
+
(11, 6),
|
143
|
+
(11, 7),
|
144
|
+
)
|
133
145
|
|
134
146
|
|
135
147
|
runtime = Runtime()
|