p47068915
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Support int32 as input type for Abs of GPU op, float64 as input type for IsFinite GPU op.

This commit is contained in:
hezhenhao1 2021-11-04 11:10:02 +08:00
parent 3cb39df189
commit d61b089f6b
5 changed files with 56 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
mindspore/ccsrc/backend/kernel_compiler/gpu/cuda_impl/float_status_impl.cu
View File

@ -128,9 +128,13 @@ void CalIsFinite(const size_t size, const T* input, bool* output, cudaStream_t c
template void CalFloatStatus<float>(const size_t size, const float* input, float* output, cudaStream_t cuda_stream);
template void CalFloatStatus<half>(const size_t size, const half* input, float* output, cudaStream_t cuda_stream);
template void CalFloatStatus<double>(const size_t size, const double* input, float* output, cudaStream_t cuda_stream);
template void CalIsInf<float>(const size_t size, const float* input, bool* output, cudaStream_t cuda_stream);
template void CalIsInf<half>(const size_t size, const half* input, bool* output, cudaStream_t cuda_stream);
template void CalIsInf<double>(const size_t size, const double* input, bool* output, cudaStream_t cuda_stream);
template void CalIsNan<float>(const size_t size, const float* input, bool* output, cudaStream_t cuda_stream);
template void CalIsNan<half>(const size_t size, const half* input, bool* output, cudaStream_t cuda_stream);
template void CalIsNan<double>(const size_t size, const double* input, bool* output, cudaStream_t cuda_stream);
template void CalIsFinite<float>(const size_t size, const float* input, bool* output, cudaStream_t cuda_stream);
template void CalIsFinite<half>(const size_t size, const half* input, bool* output, cudaStream_t cuda_stream);
template void CalIsFinite<double>(const size_t size, const double* input, bool* output, cudaStream_t cuda_stream);

8
mindspore/ccsrc/backend/kernel_compiler/gpu/math/float_status_gpu_kernel.cc
View File

@ -22,17 +22,25 @@ MS_REG_GPU_KERNEL_ONE(FloatStatus, KernelAttr().AddInputAttr(kNumberTypeFloat32)
FloatStatusGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(FloatStatus, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeFloat32),
FloatStatusGpuKernel, half)
MS_REG_GPU_KERNEL_ONE(FloatStatus, KernelAttr().AddInputAttr(kNumberTypeFloat64).AddOutputAttr(kNumberTypeFloat32),
FloatStatusGpuKernel, double)
MS_REG_GPU_KERNEL_ONE(IsInf, KernelAttr().AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(IsInf, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, half)
MS_REG_GPU_KERNEL_ONE(IsInf, KernelAttr().AddInputAttr(kNumberTypeFloat64).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, double)
MS_REG_GPU_KERNEL_ONE(IsNan, KernelAttr().AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(IsNan, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, half)
MS_REG_GPU_KERNEL_ONE(IsNan, KernelAttr().AddInputAttr(kNumberTypeFloat64).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, double)
MS_REG_GPU_KERNEL_ONE(IsFinite, KernelAttr().AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(IsFinite, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, half)
MS_REG_GPU_KERNEL_ONE(IsFinite, KernelAttr().AddInputAttr(kNumberTypeFloat64).AddOutputAttr(kNumberTypeBool),
FloatStatusGpuKernel, double)
} // namespace kernel
} // namespace mindspore

2
mindspore/ccsrc/backend/kernel_compiler/gpu/math/unary_op_gpu_kernel.cc
View File

@ -110,6 +110,8 @@ MS_REG_GPU_KERNEL_ONE(Abs, KernelAttr().AddInputAttr(kNumberTypeFloat32).AddOutp
UnaryOpGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(Abs, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeFloat16),
UnaryOpGpuKernel, half)
MS_REG_GPU_KERNEL_ONE(Abs, KernelAttr().AddInputAttr(kNumberTypeInt32).AddOutputAttr(kNumberTypeInt32),
UnaryOpGpuKernel, int)
MS_REG_GPU_KERNEL_ONE(Floor, KernelAttr().AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kNumberTypeFloat32),
UnaryOpGpuKernel, float)
MS_REG_GPU_KERNEL_ONE(Floor, KernelAttr().AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kNumberTypeFloat16),

4
mindspore/ops/operations/math_ops.py
View File

@ -4110,7 +4110,7 @@ class FloatStatus(PrimitiveWithInfer):
Tensor, has the shape of `(1,)`, and the dtype is `mindspore.dtype.float32`.
Raises:
TypeError: If dtype of `x` is neither float16 nor float32.
TypeError: If dtype of `x` is not in [float16, float32, float64].
Supported Platforms:
``GPU``
@ -4132,7 +4132,7 @@ class FloatStatus(PrimitiveWithInfer):
return [1]
def infer_dtype(self, x_dtype):
validator.check_tensor_dtype_valid('x', x_dtype, [mstype.float32, mstype.float16], self.name)
validator.check_tensor_dtype_valid('x', x_dtype, [mstype.float32, mstype.float16, mstype.float64], self.name)
return mstype.float32

56
tests/st/ops/gpu/test_float_status_op.py
View File

@ -67,17 +67,23 @@ x3 = np.array([[1, 2], [3, 4], [5.0, 88.0]]).astype(np.float32)
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_status():
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_status(dtype):
"""
Feature: ALL To ALL
Description: test cases for FloatStatus
Expectation: the result match to expectation
"""
ms_status = Net()
output1 = ms_status(Tensor(x1))
output1 = ms_status(Tensor(x1.astype(dtype)))
expect1 = 1
assert output1.asnumpy()[0] == expect1
output2 = ms_status(Tensor(x2))
output2 = ms_status(Tensor(x2.astype(dtype)))
expect2 = 1
assert output2.asnumpy()[0] == expect2
output3 = ms_status(Tensor(x3))
output3 = ms_status(Tensor(x3.astype(dtype)))
expect3 = 0
assert output3.asnumpy()[0] == expect3
@ -85,17 +91,23 @@ def test_status():
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_nan():
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_nan(dtype):
"""
Feature: ALL To ALL
Description: test cases for IsNan
Expectation: the result match to expectation
"""
ms_isnan = Netnan()
output1 = ms_isnan(Tensor(x1))
output1 = ms_isnan(Tensor(x1.astype(dtype)))
expect1 = [[False, False, True, False]]
assert (output1.asnumpy() == expect1).all()
output2 = ms_isnan(Tensor(x2))
output2 = ms_isnan(Tensor(x2.astype(dtype)))
expect2 = [[False, False, False, False]]
assert (output2.asnumpy() == expect2).all()
output3 = ms_isnan(Tensor(x3))
output3 = ms_isnan(Tensor(x3.astype(dtype)))
expect3 = [[False, False], [False, False], [False, False]]
assert (output3.asnumpy() == expect3).all()
@ -103,17 +115,23 @@ def test_nan():
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_inf():
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_inf(dtype):
"""
Feature: ALL To ALL
Description: test cases for IsInf
Expectation: the result match to expectation
"""
ms_isinf = Netinf()
output1 = ms_isinf(Tensor(x1))
output1 = ms_isinf(Tensor(x1.astype(dtype)))
expect1 = [[False, False, False, False]]
assert (output1.asnumpy() == expect1).all()
output2 = ms_isinf(Tensor(x2))
output2 = ms_isinf(Tensor(x2.astype(dtype)))
expect2 = [[True, False, False, False]]
assert (output2.asnumpy() == expect2).all()
output3 = ms_isinf(Tensor(x3))
output3 = ms_isinf(Tensor(x3.astype(dtype)))
expect3 = [[False, False], [False, False], [False, False]]
assert (output3.asnumpy() == expect3).all()
@ -121,16 +139,22 @@ def test_inf():
@pytest.mark.level0
@pytest.mark.platform_x86_gpu_training
@pytest.mark.env_onecard
def test_finite():
@pytest.mark.parametrize('dtype', [np.float16, np.float32, np.float64])
def test_finite(dtype):
"""
Feature: ALL To ALL
Description: test cases for Netfinite
Expectation: the result match to expectation
"""
ms_isfinite = Netfinite()
output1 = ms_isfinite(Tensor(x1))
output1 = ms_isfinite(Tensor(x1.astype(dtype)))
expect1 = [[True, True, False, True]]
assert (output1.asnumpy() == expect1).all()
output2 = ms_isfinite(Tensor(x2))
output2 = ms_isfinite(Tensor(x2.astype(dtype)))
expect2 = [[False, True, True, True]]
assert (output2.asnumpy() == expect2).all()
output3 = ms_isfinite(Tensor(x3))
output3 = ms_isfinite(Tensor(x3.astype(dtype)))
expect3 = [[True, True], [True, True], [True, True]]
assert (output3.asnumpy() == expect3).all()