Home
last modified time | relevance | path

Searched refs:CPUDevice (Results 1 – 25 of 246) sorted by relevance

12345678910

/external/tensorflow/tensorflow/contrib/mpi_collectives/
Dring.cc26 using CPUDevice = Eigen::ThreadPoolDevice; typedef
36 template Status RingAllreduce<CPUDevice, int>(OpKernelContext*, const Tensor*,
38 template Status RingAllreduce<CPUDevice, long long>(OpKernelContext*,
41 template Status RingAllreduce<CPUDevice, float>(OpKernelContext*, const Tensor*,
45 template Status RingAllgather<CPUDevice, int>(OpKernelContext*, const Tensor*,
48 template Status RingAllgather<CPUDevice, long long>(OpKernelContext*,
52 template Status RingAllgather<CPUDevice, float>(OpKernelContext*, const Tensor*,
58 void CopyTensorData<CPUDevice>(void* dst, void* src, size_t size) { in CopyTensorData()
65 void AccumulateTensorData<CPUDevice, type>(type * dst, type * src, \
Dmpi_ops.cc79 using CPUDevice = Eigen::ThreadPoolDevice; typedef
105 extern template Status RingAllreduce<CPUDevice, int>(OpKernelContext*,
108 extern template Status RingAllreduce<CPUDevice, long long>(OpKernelContext*,
111 extern template Status RingAllreduce<CPUDevice, float>(OpKernelContext*,
114 extern template Status RingAllgather<CPUDevice, int>(OpKernelContext*,
118 extern template Status RingAllgather<CPUDevice, long long>(
120 extern template Status RingAllgather<CPUDevice, float>(
133 bool IsGPUDevice<CPUDevice>() { in IsGPUDevice()
465 #define GPU_DEVICE_IF_CUDA CPUDevice in PerformCollectiveOp()
474 : RingAllgather<CPUDevice, float>( in PerformCollectiveOp()
[all …]
/external/tensorflow/tensorflow/contrib/mpi_collectives/kernels/
Dring.cc26 using CPUDevice = Eigen::ThreadPoolDevice; typedef
36 template Status RingAllreduce<CPUDevice, int>(OpKernelContext*, const Tensor*,
38 template Status RingAllreduce<CPUDevice, long long>(OpKernelContext*,
41 template Status RingAllreduce<CPUDevice, float>(OpKernelContext*, const Tensor*,
45 template Status RingAllgather<CPUDevice, int>(OpKernelContext*, const Tensor*,
48 template Status RingAllgather<CPUDevice, long long>(OpKernelContext*,
52 template Status RingAllgather<CPUDevice, float>(OpKernelContext*, const Tensor*,
58 void CopyTensorData<CPUDevice>(void* dst, void* src, size_t size) { in CopyTensorData()
65 void AccumulateTensorData<CPUDevice, type>(type * dst, type * src, \
Dmpi_ops.cc78 using CPUDevice = Eigen::ThreadPoolDevice; typedef
104 extern template Status RingAllreduce<CPUDevice, int>(OpKernelContext*,
107 extern template Status RingAllreduce<CPUDevice, long long>(OpKernelContext*,
110 extern template Status RingAllreduce<CPUDevice, float>(OpKernelContext*,
113 extern template Status RingAllgather<CPUDevice, int>(OpKernelContext*,
117 extern template Status RingAllgather<CPUDevice, long long>(
119 extern template Status RingAllgather<CPUDevice, float>(
132 bool IsGPUDevice<CPUDevice>() { in IsGPUDevice()
464 #define GPU_DEVICE_IF_CUDA CPUDevice in PerformCollectiveOp()
473 : RingAllgather<CPUDevice, float>( in PerformCollectiveOp()
[all …]
/external/tensorflow/tensorflow/core/kernels/
Daggregate_ops_cpu.h24 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
36 struct Add2Functor<CPUDevice, T> {
37 void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
40 Add2EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2);
44 struct Add3Functor<CPUDevice, T> {
45 void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
49 Add3EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3);
53 struct Add4Functor<CPUDevice, T> {
54 void operator()(const CPUDevice& d, typename TTypes<T>::Flat out,
59 Add4EigenImpl<CPUDevice, T>::Compute(d, out, in1, in2, in3, in4);
[all …]
Dcwise_op_clip.cc20 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
91 struct UnaryClipOp<CPUDevice, T> {
92 void operator()(const CPUDevice& d, typename TTypes<T>::ConstFlat& in0_flat, in operator ()()
110 struct BinaryRightClipOp<CPUDevice, T> {
111 void operator()(const CPUDevice& d, typename TTypes<T>::ConstFlat& in0_flat, in operator ()()
130 struct BinaryLeftClipOp<CPUDevice, T> {
131 void operator()(const CPUDevice& d, typename TTypes<T>::ConstFlat& in0_flat, in operator ()()
142 struct TernaryClipOp<CPUDevice, T> {
143 void operator()(const CPUDevice& d, typename TTypes<T>::ConstFlat& in0_flat, in operator ()()
152 template struct UnaryClipOp<CPUDevice, T>; \
[all …]
Dtranspose_functor_cpu.cc29 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
35 void TransposeSimple(const CPUDevice& device, const Tensor& in, in TransposeSimple()
71 struct Transpose<CPUDevice, T, conjugate> {
72 static void run(const CPUDevice& d, const Tensor& in, in run()
76 internal::TransposeUsingEigen<CPUDevice, T, 2>(d, in, perm, conjugate, in run()
80 internal::TransposeUsingEigen<CPUDevice, T, 3>(d, in, perm, conjugate, in run()
84 internal::TransposeUsingEigen<CPUDevice, T, 4>(d, in, perm, conjugate, in run()
88 internal::TransposeUsingEigen<CPUDevice, T, 5>(d, in, perm, conjugate, in run()
92 internal::TransposeUsingEigen<CPUDevice, T, 6>(d, in, perm, conjugate, in run()
96 internal::TransposeUsingEigen<CPUDevice, T, 7>(d, in, perm, conjugate, in run()
[all …]
Dconv_3d.h40 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
43 struct CuboidConvolution<CPUDevice, T> {
44 void operator()(const CPUDevice& d, typename TTypes<T, 5>::Tensor output,
55 struct CuboidConvolutionBackwardInput<CPUDevice, T> {
56 void operator()(const CPUDevice& d,
72 struct CuboidConvolutionBackwardFilter<CPUDevice, T> {
73 void operator()(const CPUDevice& d,
Ddense_update_functor.h27 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
45 struct DenseUpdate<CPUDevice, T, ADD> {
46 void operator()(const CPUDevice& d, typename TTypes<T>::Flat params,
53 struct DenseUpdate<CPUDevice, T, SUB> {
54 void operator()(const CPUDevice& d, typename TTypes<T>::Flat params,
61 struct DenseUpdate<CPUDevice, T, ASSIGN> {
62 void operator()(const CPUDevice& d, typename TTypes<T>::Flat params,
100 Status VariantCopyFn<CPUDevice>(OpKernelContext* context, const Tensor& from,
Dinplace_ops.cc27 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
46 Status DoParallelConcat(const CPUDevice& d, const Tensor& value, int32 loc, in DoParallelConcat()
52 return DoParallelConcatUpdate<CPUDevice, type>(d, value, loc, output); in DoParallelConcat()
158 ParallelConcatUpdate<CPUDevice>);
166 ParallelConcatStart<CPUDevice, type>)
211 ParallelConcatUpdate<CPUDevice>);
250 ParallelConcatUpdate<CPUDevice>);
296 void DoInplaceOp(const CPUDevice& d, InplaceOpType op, const Tensor& i, in DoInplaceOp()
319 void DoInplaceStringUpdateOp(const CPUDevice& d, const Tensor& i, in DoInplaceStringUpdateOp()
332 Status DoInplace(const CPUDevice& device, InplaceOpType op, const Tensor& i, in DoInplace()
[all …]
Dl2loss_op.cc29 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
32 class L2LossOp<CPUDevice, T> : public OpKernel {
44 const CPUDevice& d = context->eigen_device<CPUDevice>(); in Compute()
53 L2LossOp<CPUDevice, T>);
Dmatrix_diag_op.cc40 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
122 MatrixDiagOp<CPUDevice, type>); \
125 MatrixDiagPartOp<CPUDevice, type>);
134 MatrixDiagOp<CPUDevice, type>); \
138 MatrixDiagPartOp<CPUDevice, type>);
145 struct MatrixDiag<CPUDevice, T> {
146 static void Compute(const CPUDevice& d, in Compute()
159 struct MatrixDiagPart<CPUDevice, T> {
160 static void Compute(const CPUDevice& d, in Compute()
Drelu_op.cc30 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
39 ReluOp<CPUDevice, type>); \
42 ReluGradOp<CPUDevice, type>); \
45 Relu6Op<CPUDevice, type>); \
48 Relu6GradOp<CPUDevice, type>) \
51 LeakyReluOp<CPUDevice, type>); \
54 LeakyReluGradOp<CPUDevice, type>);
62 EluOp<CPUDevice, type>); \
65 EluGradOp<CPUDevice, type>); \
68 SeluOp<CPUDevice, type>); \
[all …]
Dreduction_ops_sum.cc26 ReductionOp<CPUDevice, type, int32, Eigen::internal::SumReducer<type>>); \
32 ReductionOp<CPUDevice, type, int64, Eigen::internal::SumReducer<type>>);
70 ReductionOp<CPUDevice, int32, int32, Eigen::internal::SumReducer<int32>>);
79 ReductionOp<CPUDevice, int32, int64, Eigen::internal::SumReducer<int32>>);
110 ReductionOp<CPUDevice, int32, int32, Eigen::internal::SumReducer<int32>>);
119 ReductionOp<CPUDevice, int32, int64, Eigen::internal::SumReducer<int32>>);
Dreduction_ops_max.cc26 ReductionOp<CPUDevice, type, int32, Eigen::internal::MaxReducer<type>>); \
32 ReductionOp<CPUDevice, type, int64, Eigen::internal::MaxReducer<type>>);
70 ReductionOp<CPUDevice, int32, int32, Eigen::internal::MaxReducer<int32>>);
79 ReductionOp<CPUDevice, int32, int64, Eigen::internal::MaxReducer<int32>>);
112 ReductionOp<CPUDevice, int32, int32, Eigen::internal::MaxReducer<int32>>);
121 ReductionOp<CPUDevice, int32, int64, Eigen::internal::MaxReducer<int32>>);
Dreduction_ops_min.cc26 ReductionOp<CPUDevice, type, int32, Eigen::internal::MinReducer<type>>); \
32 ReductionOp<CPUDevice, type, int64, Eigen::internal::MinReducer<type>>);
68 ReductionOp<CPUDevice, int32, int32, Eigen::internal::MinReducer<int32>>);
77 ReductionOp<CPUDevice, int32, int64, Eigen::internal::MinReducer<int32>>);
110 ReductionOp<CPUDevice, int32, int32, Eigen::internal::MinReducer<int32>>);
119 ReductionOp<CPUDevice, int32, int64, Eigen::internal::MinReducer<int32>>);
Ddense_update_functor.cc28 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
34 struct DenseUpdate<CPUDevice, string, ASSIGN> {
35 void operator()(const CPUDevice& d, typename TTypes<string>::Flat params, in operator ()()
76 functor::DenseUpdate<CPUDevice, T, ASSIGN> copy_functor_; \
77 copy_functor_(context->eigen_device<CPUDevice>(), tensor->flat<T>(), \
106 INSTANTIATE_GET_VARIANT_COPY_FN(CPUDevice, TF_CALL_ALL_TYPES, CPU_DENSE_COPY);
Dhistogram_op.cc29 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
35 struct HistogramFixedWidthFunctor<CPUDevice, T, Tout> {
40 const CPUDevice& d = context->eigen_device<CPUDevice>(); in Compute()
122 HistogramFixedWidthOp<CPUDevice, type, int32>) \
127 HistogramFixedWidthOp<CPUDevice, type, int64>)
Ddequantize_op.cc38 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
120 DequantizeOp<CPUDevice, quint8>);
123 DequantizeOp<CPUDevice, qint8>);
126 DequantizeOp<CPUDevice, quint16>);
129 DequantizeOp<CPUDevice, qint16>);
133 DequantizeOp<CPUDevice, qint32>);
Dsoftmax_op.cc30 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
47 struct SoftmaxFunctor<CPUDevice, T> : SoftmaxFunctorBase<CPUDevice, T> {};
84 SoftmaxOp<CPUDevice, T>);
93 SoftmaxOp<CPUDevice, T>);
Dpad_op.cc39 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
271 PadOp<CPUDevice, type, int32>); \
277 PadOp<CPUDevice, type, int64>); \
284 PadOp<CPUDevice, type, int32>); \
291 PadOp<CPUDevice, type, int64>);
371 PadOp<CPUDevice, int32, int32>);
379 PadOp<CPUDevice, int32, int64>);
388 PadOp<CPUDevice, int32, int32>);
397 PadOp<CPUDevice, int32, int64>);
438 PadOp<CPUDevice, int32, int32>);
[all …]
Dcwise_op_equal_to_1.cc23 ApproximateEqualOp<CPUDevice, float>);
26 ApproximateEqualOp<CPUDevice, double>);
46 BinaryOp<CPUDevice, functor::equal_to<int32>>);
58 BinaryOp<CPUDevice, functor::equal_to<int32>>);
/external/tensorflow/tensorflow/contrib/image/kernels/
Dimage_ops.cc33 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
35 template struct FillProjectiveTransform<CPUDevice, uint8>;
36 template struct FillProjectiveTransform<CPUDevice, int32>;
37 template struct FillProjectiveTransform<CPUDevice, int64>;
38 template struct FillProjectiveTransform<CPUDevice, Eigen::half>;
39 template struct FillProjectiveTransform<CPUDevice, float>;
40 template struct FillProjectiveTransform<CPUDevice, double>;
44 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
126 ImageProjectiveTransform<CPUDevice, TYPE>); \
130 ImageProjectiveTransform<CPUDevice, TYPE>)
/external/tensorflow/tensorflow/core/framework/
Dregister_types_traits.h21 typedef Eigen::ThreadPoolDevice CPUDevice; typedef
41 struct proxy_type_pod<CPUDevice, 16> {
45 struct proxy_type_pod<CPUDevice, 8> {
49 struct proxy_type_pod<CPUDevice, 4> {
53 struct proxy_type_pod<CPUDevice, 2> {
57 struct proxy_type_pod<CPUDevice, 1> {
/external/tensorflow/tensorflow/contrib/reduce_slice_ops/kernels/
Dreduce_slice_ops.cc28 using CPUDevice = Eigen::ThreadPoolDevice; typedef
40 struct ReduceSliceFunctor##reduceop<CPUDevice, T, Index> { \
57 virtual void operator()(OpKernelContext* ctx, const CPUDevice& d, \
98 template struct ReduceSliceFunctorSum<CPUDevice, T, Index>; \
99 template struct ReduceSliceFunctorProd<CPUDevice, T, Index>;
102 template struct ReduceSliceFunctorMax<CPUDevice, T, Index>; \
103 template struct ReduceSliceFunctorMin<CPUDevice, T, Index>;
162 ReduceSliceKernel<CPUDevice, type, index_type, \
168 ReduceSliceKernel<CPUDevice, type, index_type, \
176 ReduceSliceKernel<CPUDevice, type, index_type, \
[all …]

12345678910