use torch_npu._C instead of option-pytorch-AtomGit

Mtest/npu/test_torch_fill_uninitialized_memory.py+1-1

	@@ -122,7 +122,7 @@ class TorchFillUninitializedMemoryTestCase(TestCase):
122	self.assertTrue(res.eq(max_val).all())	122	self.assertTrue(res.eq(max_val).all())
123		123
124		124
125	-instantiate_device_type_tests(TorchFillUninitializedMemoryTestCase, globals(), only_for='privateuse1')	125	+instantiate_device_type_tests(TorchFillUninitializedMemoryTestCase, globals(), only_for=('privateuse1',))
126		126
127		127
128	if __name__ == '__main__':	128	if __name__ == '__main__':

Mtorch_npu/csrc/aten/common/ResizeNpu.cpp+2-1

	@@ -3,6 +3,7 @@
3	#include <ATen/native/ResizeCommon.h>	3	#include <ATen/native/ResizeCommon.h>
4		4
5	#include "torch_npu/csrc/framework/FormatHelper.h"	5	#include "torch_npu/csrc/framework/FormatHelper.h"
		6	+#include "torch_npu/csrc/framework/interface/EnvVariables.h"
6	#include "torch_npu/csrc/aten/NPUNativeFunctions.h"	7	#include "torch_npu/csrc/aten/NPUNativeFunctions.h"
7	#include "torch_npu/csrc/aten/common/ResizeNpu.h"	8	#include "torch_npu/csrc/aten/common/ResizeNpu.h"
8		9
	@@ -62,7 +63,7 @@ const at::Tensor& NPUNativeFunctions::resize_(
62	// fill newly added storage section with NaN/MAX_INT when requested.	63	// fill newly added storage section with NaN/MAX_INT when requested.
63	if (C10_UNLIKELY(at::globalContext().deterministicAlgorithms() &&	64	if (C10_UNLIKELY(at::globalContext().deterministicAlgorithms() &&
64	at::globalContext().deterministicFillUninitializedMemory() &&	65	at::globalContext().deterministicFillUninitializedMemory() &&
65	- at_npu::native::env::CheckFillUninitializedMemory())) {	66	+ at_npu::native::env::globalNpuContext().npuFillUninitializedMemory())) {
66	at::native::fill_resize_deterministic_(self, static_cast<int64_t>(old_storage_nbytes));	67	at::native::fill_resize_deterministic_(self, static_cast<int64_t>(old_storage_nbytes));
67	}	68	}
68	return self;	69	return self;

Mtorch_npu/csrc/aten/common/TensorFactories.cpp+1-1

	@@ -120,7 +120,7 @@ inline bool should_fill_empty_deterministic()
120	{	120	{
121	return at::globalContext().deterministicAlgorithms() &&	121	return at::globalContext().deterministicAlgorithms() &&
122	at::globalContext().deterministicFillUninitializedMemory() &&	122	at::globalContext().deterministicFillUninitializedMemory() &&
123	- at_npu::native::env::CheckFillUninitializedMemory();	123	+ at_npu::native::env::globalNpuContext().npuFillUninitializedMemory();
124	}	124	}
125		125
126	void window_function_checks(	126	void window_function_checks(

Mtorch_npu/csrc/framework/interface/EnvVariables.cpp+13-3

	@@ -321,9 +321,6 @@ bool CheckCompatibleImplFor(CompatibleKey compatible_key)
321	return CheckCompatibleImpl();	321	return CheckCompatibleImpl();
322	}	322	}
323		323
324	-TORCH_NPU_REGISTER_OPTION(TORCH_NPU_FILL_UNINITIALIZED_MEMORY)
325	-REGISTER_OPTION_BOOL_FUNCTION(CheckFillUninitializedMemory, TORCH_NPU_FILL_UNINITIALIZED_MEMORY, "0", "1")
326	-
327	REGISTER_OPTION_HOOK(ALLOW_CONV_HF32, [](const std::string &val) {	324	REGISTER_OPTION_HOOK(ALLOW_CONV_HF32, [](const std::string &val) {
328	static const std::string mm_hf32_option_name = "ALLOW_MATMUL_HF32";	325	static const std::string mm_hf32_option_name = "ALLOW_MATMUL_HF32";
329	auto mm_hf32_val = c10_npu::option::GetOption(mm_hf32_option_name);	326	auto mm_hf32_val = c10_npu::option::GetOption(mm_hf32_option_name);
	@@ -364,6 +361,19 @@ REGISTER_OPTION_HOOK(ACL_OP_DEBUG_OPTION, [](const std::string &val) {
364		361
365	TORCH_NPU_REGISTER_OPTION(CUBE_MATH_TYPE)	362	TORCH_NPU_REGISTER_OPTION(CUBE_MATH_TYPE)
366		363
		364	+NpuContext& globalNpuContext() {
		365	+ static NpuContext globalNpuContext_;
		366	+ return globalNpuContext_;
		367	+}
		368	+
		369	+bool NpuContext::npuFillUninitializedMemory() const {
		370	+ return _npu_fill_uninitialized_memory;
		371	+}
		372	+
		373	+void NpuContext::setNpuFillUninitializedMemory(bool b) {
		374	+ _npu_fill_uninitialized_memory = b;
		375	+}
		376	+
367	} // namespace env	377	} // namespace env
368	} // namespace native	378	} // namespace native
369	} // namespace at_npu	379	} // namespace at_npu

Mtorch_npu/csrc/framework/interface/EnvVariables.h+11-1

	@@ -32,12 +32,22 @@ enum class CompatibleKey {
32	bool CheckCompatibleImpl();	32	bool CheckCompatibleImpl();
33	bool CheckCompatibleImplFor(CompatibleKey compatible_key);	33	bool CheckCompatibleImplFor(CompatibleKey compatible_key);
34	bool CheckCompatibleImplBlackListFor(CompatibleKey compatible_key);	34	bool CheckCompatibleImplBlackListFor(CompatibleKey compatible_key);
35	-bool CheckFillUninitializedMemory();
36		35
37	bool IsAllowFP32ToFP16();	36	bool IsAllowFP32ToFP16();
38	bool IsAllowConvHF32();	37	bool IsAllowConvHF32();
39	bool IsAllowMatmulHF32();	38	bool IsAllowMatmulHF32();
40		39
		40	+class NpuContext {
		41	+ public:
		42	+ bool npuFillUninitializedMemory() const;
		43	+ void setNpuFillUninitializedMemory(bool /b/);
		44	+
		45	+ private:
		46	+ bool _npu_fill_uninitialized_memory = false;
		47	+};
		48	+
		49	+NpuContext& globalNpuContext();
		50	+
41	} // namespace env	51	} // namespace env
42	} // namespace native	52	} // namespace native
43	} // namespace at_npu	53	} // namespace at_npu

Mtorch_npu/csrc/npu/Module.cpp+29-0

	@@ -2530,6 +2530,27 @@ PyObject* THNPModule_get_deterministic_level(PyObject* self, PyObject* noargs) {
2530	END_HANDLE_TH_ERRORS	2530	END_HANDLE_TH_ERRORS
2531	}	2531	}
2532		2532
		2533	+static PyObject* THNPModule_setNpuFillUninitializedMemory(
		2534	+ PyObject* _unused,
		2535	+ PyObject* arg) {
		2536	+ HANDLE_TH_ERRORS
		2537	+ TORCH_CHECK(
		2538	+ PyBool_Check(arg), "expected a bool, but got ", THPUtils_typename(arg));
		2539	+ at_npu::native::env::globalNpuContext().setNpuFillUninitializedMemory(
		2540	+ arg == Py_True);
		2541	+ Py_RETURN_NONE;
		2542	+ END_HANDLE_TH_ERRORS
		2543	+}
		2544	+
		2545	+static PyObject* THNPModule_npuFillUninitializedMemory(
		2546	+ PyObject* _unused,
		2547	+ PyObject* noargs) {
		2548	+ if (at_npu::native::env::globalNpuContext().npuFillUninitializedMemory())
		2549	+ Py_RETURN_TRUE;
		2550	+ else
		2551	+ Py_RETURN_FALSE;
		2552	+}
		2553	+
2533	PyObject* THNPModule_hasPrimaryContext_wrap(PyObject* self, PyObject* arg) {	2554	PyObject* THNPModule_hasPrimaryContext_wrap(PyObject* self, PyObject* arg) {
2534	HANDLE_TH_ERRORS	2555	HANDLE_TH_ERRORS
2535	TORCH_CHECK(	2556	TORCH_CHECK(
	@@ -2860,6 +2881,14 @@ static struct PyMethodDef THNPModule_methods[] = {
2860	(PyCFunction)THNPModule_get_deterministic_level,	2881	(PyCFunction)THNPModule_get_deterministic_level,
2861	METH_NOARGS,	2882	METH_NOARGS,
2862	nullptr},	2883	nullptr},
		2884	+ {"_npu_set_fill_uninitialized_memory",
		2885	+ (PyCFunction)THNPModule_setNpuFillUninitializedMemory,
		2886	+ METH_O,
		2887	+ nullptr},
		2888	+ {"_npu_get_fill_uninitialized_memory",
		2889	+ (PyCFunction)THNPModule_npuFillUninitializedMemory,
		2890	+ METH_NOARGS,
		2891	+ nullptr},
2863	{"_npu_sleep",	2892	{"_npu_sleep",
2864	(PyCFunction)THNPModule_npuSleep,	2893	(PyCFunction)THNPModule_npuSleep,
2865	METH_O,	2894	METH_O,

Mtorch_npu/utils/_deterministic.py+2-3

	@@ -2,7 +2,7 @@ def _add_deterministic_patch():
2	"""	2	"""
3	Patch torch.utils.deterministic.fill_uninitialized_memory setter so that	3	Patch torch.utils.deterministic.fill_uninitialized_memory setter so that
4	setting it also synchronizes the NPU-side option	4	setting it also synchronizes the NPU-side option
5	- TORCH_NPU_FILL_UNINITIALIZED_MEMORY via torch_npu._C._npu_setOption.	5	+ _npu_fill_uninitialized_memory via torch_npu._C._npu_set_fill_uninitialized_memory.
6	"""	6	"""
7	import torch_npu	7	import torch_npu
8	import torch.utils.deterministic as det_mod	8	import torch.utils.deterministic as det_mod
	@@ -19,7 +19,6 @@ def _add_deterministic_patch():
19	# 1. call original setter: torch._C._set_deterministic_fill_uninitialized_memory(mode)	19	# 1. call original setter: torch._C._set_deterministic_fill_uninitialized_memory(mode)
20	_orig_fset(self, mode)	20	_orig_fset(self, mode)
21	# 2. sync NPU-side option	21	# 2. sync NPU-side option
22	- option = {"TORCH_NPU_FILL_UNINITIALIZED_MEMORY": "1" if mode else "0"}	22	+ torch_npu._C._npu_set_fill_uninitialized_memory(mode)
23	- torch_npu._C._npu_setOption(option)
24		23
25	_Deterministic.fill_uninitialized_memory = property(_orig_fget, _new_fset)	24	_Deterministic.fill_uninitialized_memory = property(_orig_fget, _new_fset)