/external/tensorflow/tensorflow/stream_executor/rocm/ |
D | rocm_dnn.h | 67 const DeviceMemory<Eigen::half>& input_data, 69 const DeviceMemory<Eigen::half>& input_h_data, 71 const DeviceMemory<Eigen::half>& input_c_data, 72 const DeviceMemory<Eigen::half>& params, 74 DeviceMemory<Eigen::half>* output_data, 76 DeviceMemory<Eigen::half>* output_h_data, 78 DeviceMemory<Eigen::half>* output_c_data, bool is_training, 85 const DeviceMemory<float>& input_data, 87 const DeviceMemory<float>& input_h_data, 89 const DeviceMemory<float>& input_c_data, [all …]
|
D | rocm_blas.cc | 420 const DeviceMemory<float> &x, int incx, in DoBlasAsum() 421 DeviceMemory<float> *result) { in DoBlasAsum() 428 const DeviceMemory<double> &x, int incx, in DoBlasAsum() 429 DeviceMemory<double> *result) { in DoBlasAsum() 436 const DeviceMemory<std::complex<float>> &x, int incx, in DoBlasAsum() 437 DeviceMemory<float> *result) { in DoBlasAsum() 444 const DeviceMemory<std::complex<double>> &x, int incx, in DoBlasAsum() 445 DeviceMemory<double> *result) { in DoBlasAsum() 452 const DeviceMemory<float> &x, int incx, in DoBlasAxpy() 453 DeviceMemory<float> *y, int incy) { in DoBlasAxpy() [all …]
|
D | rocm_dnn.cc | 1938 const DeviceMemory<T>& input_data, in ExtractAndCheckRnnForward() 1940 const DeviceMemory<T>& input_h_data, in ExtractAndCheckRnnForward() 1942 const DeviceMemory<T>& input_c_data, const DeviceMemory<T>& params, in ExtractAndCheckRnnForward() 1944 const DeviceMemory<T>& output_data, in ExtractAndCheckRnnForward() 1946 const DeviceMemory<T>& output_h_data, in ExtractAndCheckRnnForward() 1948 const DeviceMemory<T>& output_c_data, RnnModelDims* model_dims) { in ExtractAndCheckRnnForward() 2015 DeviceMemory<uint8>* workspace) { in CreateRnnWorkspace() 2037 *workspace = DeviceMemory<uint8>(); in CreateRnnWorkspace() 2048 const DeviceMemory<T>& input_data, in DoRnnForwardImpl() 2050 const DeviceMemory<T>& input_h_data, in DoRnnForwardImpl() [all …]
|
D | rocm_blas.h | 123 DeviceMemory<typename RocBlasTypeConversionHelper<T>::mapped_type> 146 const port::ArraySlice<DeviceMemory<T> *> &a_ptrs_to_wrappers, int lda, 147 const port::ArraySlice<DeviceMemory<T> *> &b_ptrs_to_wrappers, int ldb, 148 T beta, const port::ArraySlice<DeviceMemory<T> *> &c_ptrs_to_wrappers, 160 uint64 n, uint64 k, const CompT &alpha, const DeviceMemory<InT> &a, 161 int lda, const DeviceMemory<InT> &b, int ldb, const CompT &beta, 162 DeviceMemory<OutT> *c, int ldc, blas::ComputationType computation_type, 170 uint64 n, uint64 k, const ParamType &alpha, const DeviceMemory<T> &a, 171 int lda, const DeviceMemory<T> &b, int ldb, const ParamType &beta, 172 DeviceMemory<T> *c, int ldc, blas::ProfileResult *output_profile_result); [all …]
|
/external/tensorflow/tensorflow/stream_executor/cuda/ |
D | cuda_dnn.h | 76 const DeviceMemory<Eigen::half>& input_data, 78 const DeviceMemory<Eigen::half>& input_h_data, 80 const DeviceMemory<Eigen::half>& input_c_data, 81 const DeviceMemory<Eigen::half>& params, 83 DeviceMemory<Eigen::half>* output_data, 85 DeviceMemory<Eigen::half>* output_h_data, 87 DeviceMemory<Eigen::half>* output_c_data, bool is_training, 94 const DeviceMemory<float>& input_data, 96 const DeviceMemory<float>& input_h_data, 98 const DeviceMemory<float>& input_c_data, [all …]
|
D | cuda_blas.cc | 434 const DeviceMemory<float> &x, int incx, in DoBlasAsum() 435 DeviceMemory<float> *result) { in DoBlasAsum() 442 const DeviceMemory<double> &x, int incx, in DoBlasAsum() 443 DeviceMemory<double> *result) { in DoBlasAsum() 450 const DeviceMemory<std::complex<float>> &x, int incx, in DoBlasAsum() 451 DeviceMemory<float> *result) { in DoBlasAsum() 458 const DeviceMemory<std::complex<double>> &x, int incx, in DoBlasAsum() 459 DeviceMemory<double> *result) { in DoBlasAsum() 466 const DeviceMemory<float> &x, int incx, in DoBlasAxpy() 467 DeviceMemory<float> *y, int incy) { in DoBlasAxpy() [all …]
|
D | cuda_dnn.cc | 985 DeviceMemory<uint8> state_memory; in Create() 1575 const DeviceMemory<T>& input_data, in ExtractAndCheckRnnForward() 1577 const DeviceMemory<T>& input_h_data, in ExtractAndCheckRnnForward() 1579 const DeviceMemory<T>& input_c_data, const DeviceMemory<T>& params, in ExtractAndCheckRnnForward() 1581 const DeviceMemory<T>& output_data, in ExtractAndCheckRnnForward() 1583 const DeviceMemory<T>& output_h_data, in ExtractAndCheckRnnForward() 1585 const DeviceMemory<T>& output_c_data) { in ExtractAndCheckRnnForward() 1649 port::StatusOr<DeviceMemory<uint8>> CreateRnnWorkspace( in CreateRnnWorkspace() 1662 return DeviceMemory<uint8>(); in CreateRnnWorkspace() 1668 port::StatusOr<DeviceMemory<uint8>> CreateBatchNormForwardWorkspace( in CreateBatchNormForwardWorkspace() [all …]
|
D | cuda_blas.h | 113 const port::ArraySlice<DeviceMemory<T> *> &a_array, int lda, 114 const port::ArraySlice<DeviceMemory<T> *> &b_array, int ldb, Scalar beta, 115 const port::ArraySlice<DeviceMemory<T> *> &c_array, int ldc, 123 const DeviceMemory<InT> &a, int lda, const DeviceMemory<InT> &b, int ldb, 124 const HostOrDeviceScalar<CompT> &beta, DeviceMemory<OutT> *c, int ldc, 132 uint64 n, uint64 k, const ParamType &alpha, const DeviceMemory<T> &a, 133 int lda, const DeviceMemory<T> &b, int ldb, const ParamType &beta, 134 DeviceMemory<T> *c, int ldc, blas::ProfileResult *output_profile_result); 140 const DeviceMemory<T> &a, int lda, 141 const DeviceMemory<T> &x, int incx, [all …]
|
/external/tensorflow/tensorflow/stream_executor/ |
D | stream.h | 62 class DeviceMemory; variable 233 const DeviceMemory<float> &x, const DeviceMemory<float> &scale, 234 const DeviceMemory<float> &offset, 235 const DeviceMemory<float> &estimated_mean, 236 const DeviceMemory<float> &estimated_variance, 237 const DeviceMemory<float> &side_input, const dnn::BatchDescriptor &x_desc, 240 dnn::ActivationMode activation_mode, DeviceMemory<float> *y, 241 DeviceMemory<float> *batch_mean, DeviceMemory<float> *batch_var, 242 DeviceMemory<float> *saved_mean, DeviceMemory<float> *saved_inv_var, 244 std::function<const DeviceMemory<float> &()> var_to_inv_var, [all …]
|
D | blas.h | 61 class DeviceMemory; variable 183 const DeviceMemory<float> &x, int incx, 184 DeviceMemory<float> *result) = 0; 186 const DeviceMemory<double> &x, int incx, 187 DeviceMemory<double> *result) = 0; 189 const DeviceMemory<std::complex<float>> &x, int incx, 190 DeviceMemory<float> *result) = 0; 192 const DeviceMemory<std::complex<double>> &x, int incx, 193 DeviceMemory<double> *result) = 0; 197 const DeviceMemory<float> &x, int incx, [all …]
|
D | dnn.h | 1040 Stream* stream, const DeviceMemory<float>& x, 1041 const DeviceMemory<float>& scale, const DeviceMemory<float>& offset, 1042 const DeviceMemory<float>& estimated_mean, 1043 const DeviceMemory<float>& estimated_variance, 1044 const DeviceMemory<float>& side_input, const dnn::BatchDescriptor& x_desc, 1047 dnn::ActivationMode activation_mode, DeviceMemory<float>* y, 1048 DeviceMemory<float>* batch_mean, DeviceMemory<float>* batch_var, 1049 DeviceMemory<float>* reserve_space_1, 1050 DeviceMemory<float>* reserve_space_2, bool is_training, 1053 std::function<const DeviceMemory<float>&()> var_to_inv_var, [all …]
|
D | stream.cc | 341 const DeviceMemory<float> &x, const DeviceMemory<float> &scale, in ThenBatchNormalizationForward() 342 const DeviceMemory<float> &offset, in ThenBatchNormalizationForward() 343 const DeviceMemory<float> &estimated_mean, in ThenBatchNormalizationForward() 344 const DeviceMemory<float> &estimated_variance, in ThenBatchNormalizationForward() 345 const DeviceMemory<float> &side_input, const dnn::BatchDescriptor &x_desc, in ThenBatchNormalizationForward() 348 dnn::ActivationMode activation_mode, DeviceMemory<float> *y, in ThenBatchNormalizationForward() 349 DeviceMemory<float> *batch_mean, DeviceMemory<float> *batch_var, in ThenBatchNormalizationForward() 350 DeviceMemory<float> *saved_mean, DeviceMemory<float> *saved_inv_var, in ThenBatchNormalizationForward() 352 std::function<const DeviceMemory<float> &()> var_to_inv_var, in ThenBatchNormalizationForward() 375 const DeviceMemory<float> &y_backprop, const DeviceMemory<float> &x, in ThenBatchNormalizationBackward() [all …]
|
D | fft.h | 55 class DeviceMemory; variable 181 const DeviceMemory<std::complex<float>> &input, 182 DeviceMemory<std::complex<float>> *output) = 0; 184 const DeviceMemory<std::complex<double>> &input, 185 DeviceMemory<std::complex<double>> *output) = 0; 189 const DeviceMemory<float> &input, 190 DeviceMemory<std::complex<float>> *output) = 0; 192 const DeviceMemory<double> &input, 193 DeviceMemory<std::complex<double>> *output) = 0; 197 const DeviceMemory<std::complex<float>> &input, [all …]
|
D | rng.h | 29 class DeviceMemory; variable 54 DeviceMemory<float> *v) = 0; 56 DeviceMemory<double> *v) = 0; 58 DeviceMemory<std::complex<float>> *v) = 0; 60 DeviceMemory<std::complex<double>> *v) = 0; 65 DeviceMemory<float> *v) { in DoPopulateRandGaussian() 71 double stddev, DeviceMemory<double> *v) { in DoPopulateRandGaussian()
|
D | device_memory_allocator.h | 103 const DeviceMemory<ElemT> &cref() const { return wrapped_; } in cref() 108 DeviceMemory<ElemT> *ptr() { return &wrapped_; } in ptr() 109 const DeviceMemory<ElemT> *ptr() const { return &wrapped_; } in ptr() 114 const DeviceMemory<ElemT> &operator*() const { return cref(); } 115 DeviceMemory<ElemT> *operator->() { return ptr(); } 116 const DeviceMemory<ElemT> *operator->() const { return ptr(); } 126 DeviceMemory<ElemT> Release() { in Release() 127 DeviceMemory<ElemT> tmp = wrapped_; in Release() 128 wrapped_ = DeviceMemory<ElemT>{}; in Release() 141 DeviceMemory<ElemT> wrapped_; // Value we wrap with scoped-release. [all …]
|
D | device_memory.h | 122 class DeviceMemory final : public DeviceMemoryBase { 125 DeviceMemory() : DeviceMemoryBase(nullptr, 0) {} in DeviceMemory() function 126 explicit DeviceMemory(std::nullptr_t) : DeviceMemory() {} in DeviceMemory() function 130 explicit DeviceMemory(const DeviceMemoryBase &other) in DeviceMemory() function 146 static DeviceMemory<ElemT> MakeFromByteSize(void *opaque, uint64 bytes) { in MakeFromByteSize() 147 return DeviceMemory<ElemT>(opaque, bytes); in MakeFromByteSize() 168 DeviceMemory(void *opaque, uint64 size) : DeviceMemoryBase(opaque, size) {} in DeviceMemory() function
|
/external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
D | cudnn_batchnorm_runner.cc | 34 se::DeviceMemory<float> scale; 41 se::DeviceMemory<float> offset; 42 se::DeviceMemory<float> mean; 43 se::DeviceMemory<float> variance; 49 se::DeviceMemory<float> offset; 50 se::DeviceMemory<float> output_mean; 51 se::DeviceMemory<float> output_inv_stddev; 58 se::DeviceMemory<float> output_grad_scale; 59 se::DeviceMemory<float> output_grad_offset; 60 se::DeviceMemory<float> mean; [all …]
|
D | cudnn_batchnorm_runner.h | 33 se::DeviceMemoryBase output, se::DeviceMemory<float> scale, 34 se::DeviceMemory<float> offset, se::DeviceMemory<float> mean, 35 se::DeviceMemory<float> variance, float epsilon, int64 feature_index, 40 se::DeviceMemoryBase output_data, se::DeviceMemory<float> output_mean, 41 se::DeviceMemory<float> output_inv_stddev, se::DeviceMemory<float> scale, 42 se::DeviceMemory<float> offset, float epsilon, int64 feature_index, 48 se::DeviceMemory<float> output_grad_scale, 49 se::DeviceMemory<float> output_grad_offset, se::DeviceMemory<float> scale, 50 se::DeviceMemory<float> mean, se::DeviceMemory<float> inv_stddev,
|
D | cudnn_batchnorm_thunk.cc | 125 se::DeviceMemory<float> scale(buffer_allocations.GetDeviceAddress(scale_)); in ExecuteOnStream() 126 se::DeviceMemory<float> offset(buffer_allocations.GetDeviceAddress(offset_)); in ExecuteOnStream() 127 se::DeviceMemory<float> mean(buffer_allocations.GetDeviceAddress(mean_)); in ExecuteOnStream() 128 se::DeviceMemory<float> variance( in ExecuteOnStream() 174 se::DeviceMemory<float> output_mean( in ExecuteOnStream() 176 se::DeviceMemory<float> output_inv_stddev( in ExecuteOnStream() 179 se::DeviceMemory<float> null_device_ptr(nullptr); in ExecuteOnStream() 185 se::DeviceMemory<float>(buffer_allocations.GetDeviceAddress(scale_)), in ExecuteOnStream() 186 se::DeviceMemory<float>(buffer_allocations.GetDeviceAddress(offset_)), in ExecuteOnStream() 195 se::DeviceMemory<void*> tuple_addr( in ExecuteOnStream() [all …]
|
D | cholesky_thunk.cc | 81 se::DeviceMemory<int> info_data( in ExecuteOnStream() 86 context->Potrf(uplo_, n_, se::DeviceMemory<float>(a_data), n_, in ExecuteOnStream() 87 info_data, se::DeviceMemory<float>(workspace_data))); in ExecuteOnStream() 92 uplo_, n_, se::DeviceMemory<double>(a_data), n_, info_data, in ExecuteOnStream() 93 se::DeviceMemory<double>(workspace_data))); in ExecuteOnStream() 98 uplo_, n_, se::DeviceMemory<std::complex<float>>(a_data), n_, in ExecuteOnStream() 99 info_data, se::DeviceMemory<std::complex<float>>(workspace_data))); in ExecuteOnStream() 104 uplo_, n_, se::DeviceMemory<std::complex<double>>(a_data), n_, in ExecuteOnStream() 105 info_data, se::DeviceMemory<std::complex<double>>(workspace_data))); in ExecuteOnStream()
|
D | cusolver_context.h | 58 Status Potrf(se::blas::UpperLower uplo, int n, se::DeviceMemory<T> dev_A, 59 int lda, se::DeviceMemory<int> dev_lapack_info, 60 se::DeviceMemory<T> workspace) = delete; 81 se::blas::UpperLower uplo, int n, se::DeviceMemory<T> A, int lda, \ 82 se::DeviceMemory<int> lapack_info, se::DeviceMemory<T> workspace); 103 Status Potrf(se::blas::UpperLower uplo, int n, se::DeviceMemory<T> dev_A, 104 int lda, se::DeviceMemory<int> dev_lapack_info, 105 se::DeviceMemory<T> workspace) {
|
/external/swiftshader/src/Vulkan/ |
D | VkDeviceMemory.cpp | 24 class DeviceMemory::ExternalBase 94 class DeviceMemoryHostExternalBase : public DeviceMemory::ExternalBase 134 class ExternalMemoryHost : public vk::DeviceMemory::ExternalBase 246 DeviceMemory::DeviceMemory(const VkMemoryAllocateInfo *pAllocateInfo, void *mem) in DeviceMemory() function in vk::DeviceMemory 258 void DeviceMemory::destroy(const VkAllocationCallbacks *pAllocator) in destroy() 269 size_t DeviceMemory::ComputeRequiredAllocationSize(const VkMemoryAllocateInfo *pAllocateInfo) in ComputeRequiredAllocationSize() 276 VkResult DeviceMemory::allocate() in allocate() 291 VkResult DeviceMemory::map(VkDeviceSize pOffset, VkDeviceSize pSize, void **ppData) in map() 298 VkDeviceSize DeviceMemory::getCommittedMemoryInBytes() const in getCommittedMemoryInBytes() 303 void *DeviceMemory::getOffsetPointer(VkDeviceSize pOffset) const in getOffsetPointer() [all …]
|
D | VkDeviceMemory.hpp | 23 class DeviceMemory : public Object<DeviceMemory, VkDeviceMemory> class 26 DeviceMemory(const VkMemoryAllocateInfo *pCreateInfo, void *mem); 61 static inline DeviceMemory *Cast(VkDeviceMemory object) in Cast() 63 return DeviceMemory::Cast(object); in Cast()
|
/external/tensorflow/tensorflow/stream_executor/gpu/ |
D | gpu_rng.h | 30 class DeviceMemory; variable 56 bool DoPopulateRandUniform(Stream* stream, DeviceMemory<float>* v) override; 57 bool DoPopulateRandUniform(Stream* stream, DeviceMemory<double>* v) override; 59 DeviceMemory<std::complex<float>>* v) override; 61 DeviceMemory<std::complex<double>>* v) override; 63 DeviceMemory<float>* v) override; 65 DeviceMemory<double>* v) override; 73 bool DoPopulateRandUniformInternal(Stream* stream, DeviceMemory<T>* v); 76 DeviceMemory<ElemT>* v, FuncT func);
|
D | redzone_allocator.cc | 62 port::StatusOr<DeviceMemory<uint8>> RedzoneAllocator::AllocateBytes( in AllocateBytes() 82 DeviceMemory<uint8> allocated_buffer_memory(*allocated_buffer); in AllocateBytes() 84 DeviceMemory<uint8> lhs_redzone = stream_->parent()->GetSubBuffer( in AllocateBytes() 87 DeviceMemory<uint8> data_chunk = stream_->parent()->GetSubBuffer( in AllocateBytes() 95 DeviceMemory<uint8> rhs_redzone_slop = stream_->parent()->GetSubBuffer( in AllocateBytes() 98 DeviceMemory<uint8> rhs_redzone_nonslop = stream_->parent()->GetSubBuffer( in AllocateBytes() 175 TypedKernel<DeviceMemory<uint8>, uint8, uint64, DeviceMemory<uint64>>; 215 const DeviceMemory<uint8>& redzone, in RunRedzoneChecker() 217 const DeviceMemory<uint64>& out_param, in RunRedzoneChecker() 251 const DeviceMemory<uint64>& out_param, in CheckRedzonesForBuffer() [all …]
|