/third_party/mindspore/mindspore/ccsrc/minddata/dataset/core/ |
D | tensor_row.cc | 27 : id_(kDefaultRowId), path_({}), row_(n, t), tensor_row_flag_(kFlagNone) {} in TensorRow() 30 : id_(kDefaultRowId), path_({}), row_(v), tensor_row_flag_(kFlagNone) {} in TensorRow() 33 : id_(id), path_({}), row_(lst), tensor_row_flag_(kFlagNone) {} in TensorRow() 36 : id_(tr.id_), path_(tr.path_), row_(tr.row_), tensor_row_flag_(tr.tensor_row_flag_) {} in TensorRow() 44 row_ = tr.row_; in operator =() 52 row_ = lst; in operator =() 58 : id_(kDefaultRowId), path_({}), row_(std::move(v)), tensor_row_flag_(kFlagNone) {} in TensorRow() 61 : id_(id), path_({}), row_(std::move(lst)), tensor_row_flag_(kFlagNone) {} in TensorRow() 66 row_ = std::move(tr.row_); in TensorRow() 74 row_ = std::move(tr.row_); in operator =() [all …]
|
D | tensor_row.h | 168 const vector_type &getRow() const { return row_; } in getRow() 172 for (auto &it : row_) { in SizeInBytes() 179 void emplace_back(value_type t) { row_.emplace_back(t); } in emplace_back() 181 void push_back(value_type t) { row_.push_back(t); } in push_back() 183 void clear() noexcept { row_.clear(); } in clear() 185 size_type size() const noexcept { return row_.size(); } in size() 187 void reserve(size_type size) { row_.reserve(size); } in reserve() 189 void resize(size_type size) { row_.resize(size); } in resize() 191 bool empty() { return row_.empty(); } in empty() 193 …void insert(iterator position, iterator first, iterator last) { row_.insert(position, first, last)… in insert() [all …]
|
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/fp32/ |
D | attention_fp32.c | 31 matrix->row_ = row; in InitMatrix() 43 int real_row = matrix->is_transpose_ ? matrix->col_ : matrix->row_; in LeftMatrixPackElementSize() 44 int deep = matrix->is_transpose_ ? matrix->row_ : matrix->col_; in LeftMatrixPackElementSize() 57 int deep = matrix->is_transpose_ ? matrix->col_ : matrix->row_; in RightMatrixPackElementSize() 58 int real_col = matrix->is_transpose_ ? matrix->row_ : matrix->col_; in RightMatrixPackElementSize() 71 int real_row = matrix->is_transpose_ ? matrix->col_ : matrix->row_; in PackLeftMatrix() 72 int deep = matrix->is_transpose_ ? matrix->row_ : matrix->col_; in PackLeftMatrix() 75 int src_area = matrix->row_ * matrix->col_; in PackLeftMatrix() 133 int deep = matrix->is_transpose_ ? matrix->col_ : matrix->row_; in PackRightMatrix() 134 int real_col = matrix->is_transpose_ ? matrix->row_ : matrix->col_; in PackRightMatrix() [all …]
|
/third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp32/ |
D | convolution_1x1_fp32.cc | 57 matmul_param_->row_ = conv_param_->output_h_ * conv_param_->output_w_; in InitConv1x1MatmulParam() 60 matmul_param_->row_align_ = UP_ROUND(matmul_param_->row_, row_tile_); in InitConv1x1MatmulParam() 66 …if ((matmul_param_->row_ > (row_tile_ * op_parameter_->thread_num_)) && (matmul_param_->row_ > mat… in InitConv1x1Param() 68 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->row_, row_tile_)); in InitConv1x1Param() 73 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->row_, row_tile_), thread_count_) * row_tile_; in InitConv1x1Param() 87 …input_ptr_ = reinterpret_cast<float *>(malloc(matmul_param_->row_ * matmul_param_->deep_ * sizeof(… in InitConv1x1Param() 92 memset(input_ptr_, 0, matmul_param_->row_ * matmul_param_->deep_ * sizeof(float)); in InitConv1x1Param() 156 matmul_param_->row_, cur_oc, matmul_param_->col_, OutType_Nhwc); in DoConv1x1() 159 … output_ptr_ + task_id * thread_stride_ * matmul_param_->row_, bias, matmul_param_->act_type_, in DoConv1x1() 160 … matmul_param_->deep_, matmul_param_->row_, cur_oc, matmul_param_->row_, OutType_NC4HW4); in DoConv1x1() [all …]
|
D | matmul_fp32_base.cc | 56 if (params_->row_ == 1) { in ResizeParameter() 66 params_->row_align_ = UP_ROUND(params_->row_, row_tile_); in ResizeParameter() 178 const float *src = src_ptr + i * params_->deep_ * params_->row_; in InitMatrixA() 181 matrix_a_pack_fun_(src, dst, params_->deep_, params_->row_); in InitMatrixA() 183 matrix_a_pack_fun_(src, dst, params_->row_, params_->deep_); in InitMatrixA() 298 …MatMulOpt(batch_a_ptr_, b, c, bias, params_->act_type_, params_->deep_, params_->row_, cur_oc, par… in FloatRun() 326 params_->row_align_ = UP_ROUND(params_->row_, row_tile_); in init_global_variable() 424 params_->batch * params_->row_ * oc_block_num * col_tile_ * static_cast<int>(sizeof(float)))); in InitTmpOutBuffer() 477 batch_c_ptr_ = output_data_ + i * params_->row_ * params_->col_align_; in Run() 480 batch_c_ptr_ = output_data_ + i * params_->row_ * params_->col_; in Run() [all …]
|
/third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp16/ |
D | convolution_1x1_fp16.cc | 31 matmul_param_->row_ = conv_param_->output_h_ * conv_param_->output_w_; in InitMatmulParam() 34 matmul_param_->row_align_ = UP_ROUND(matmul_param_->row_, row_tile_); in InitMatmulParam() 53 …if ((matmul_param_->row_ > (row_tile_ * op_parameter_->thread_num_)) && (matmul_param_->row_ > mat… in InitConv1x1Param() 55 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->row_, row_tile_)); in InitConv1x1Param() 60 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->row_, row_tile_), thread_count_) * row_tile_; in InitConv1x1Param() 73 …input_ptr_ = reinterpret_cast<float16_t *>(malloc(matmul_param_->row_ * matmul_param_->deep_ * siz… in InitConv1x1Param() 78 memset(input_ptr_, 0, matmul_param_->row_ * matmul_param_->deep_ * sizeof(float16_t)); in InitConv1x1Param() 217 matmul_param_->row_, cur_oc, matmul_param_->col_, OutType_Nhwc); in RunOc() 222 matmul_param_->row_, cur_oc, matmul_param_->col_, OutType_Nhwc); in RunOc() 234 int res_stride = matmul_param_->row_ - task_id * thread_stride_; in RunHw() [all …]
|
D | matmul_base_fp16.cc | 119 if (params_->row_ == 1) { in ResizeParameter() 131 params_->row_align_ = UP_ROUND(params_->row_, row_tile_); in ResizeParameter() 177 …const int8_t *src = int8_src + i * params_->deep_ * params_->row_ * lite::DataTypeSize(src_data_ty… in InitMatrixA() 181 RowMajor2RowNMajorFp16((const float16_t *)src, dst, params_->deep_, params_->row_); in InitMatrixA() 183 …RowMajor2Row12MajorFp16(src, dst, params_->deep_, params_->row_, src_data_type == kNumberTypeFloat… in InitMatrixA() 187 RowMajor2ColNMajorFp16((const float16_t *)src, dst, params_->row_, params_->deep_); in InitMatrixA() 189 …RowMajor2Col12MajorFp16(src, dst, params_->row_, params_->deep_, src_data_type == kNumberTypeFloat… in InitMatrixA() 304 …MatmulBaseFp16Neon(batch_a_ptr_, b, c, bias, params_->act_type_, params_->deep_, params_->row_, cu… in RunImpl() 307 …MatMulFp16(batch_a_ptr_, b, c, bias, params_->act_type_, params_->deep_, params_->row_, cur_oc, pa… in RunImpl() 341 batch_c_ptr_ = c_ptr + i * params_->row_ * params_->col_; in Run() [all …]
|
/third_party/boost/boost/qvm/ |
D | map_mat_vec.hpp | 150 row_ class 152 row_( row_ const & ); 153 row_ & operator=( row_ const & ); 154 ~row_(); 160 row_ & 180 vec_traits< qvm_detail::row_<Row,OriginalMatrix> > 182 typedef qvm_detail::row_<Row,OriginalMatrix> this_vector; 233 deduce_vec<qvm_detail::row_<Row,OriginalMatrix>,D> 240 deduce_vec2<qvm_detail::row_<Row,OriginalMatrix>,qvm_detail::row_<Row,OriginalMatrix>,D> 248 qvm_detail::row_<Row,A> const &>::type [all …]
|
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/mkldnn/ |
D | matmul_cpu_kernel.cc | 66 const float *src = src_ptr + i * param_.row_ * param_.deep_; in InitMatrixA() 70 RowMajor2Row6Major(src, dst, param_.deep_, param_.row_); in InitMatrixA() 72 RowMajor2Col6Major(src, dst, param_.row_, param_.deep_); in InitMatrixA() 76 RowMajor2Row4Major(src, dst, param_.deep_, param_.row_); in InitMatrixA() 78 RowMajor2Col4Major(src, dst, param_.row_, param_.deep_); in InitMatrixA() 82 RowMajor2Row12Major(src, dst, param_.deep_, param_.row_); in InitMatrixA() 84 RowMajor2Col12Major(src, dst, param_.row_, param_.deep_); in InitMatrixA() 145 param_.row_ = SizeToInt(o_shape[rank_ - kIndexOffset]); in InitArmKernel() 148 vec_matmul_ = (param_.row_ == 1); in InitArmKernel() 149 param_.row_align_ = vec_matmul_ ? 1 : UP_ROUND(param_.row_, row_tile_); in InitArmKernel() [all …]
|
/third_party/mindspore/mindspore/ccsrc/minddata/dataset/util/ |
D | queue_map.h | 104 CHECK_FAIL_RETURN_UNEXPECTED(!row_.empty(), "Programming error"); in Wait() 105 *out = std::move(row_.front()); in Wait() 106 row_.pop_front(); in Wait() 112 row_.push_back(std::move(row)); in WakeUpAny() 120 out << "sz:" << rq.row_.size() << ",uc:" << rq.use_count_.Peek(); 127 std::deque<T> row_; variable
|
/third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/int8/ |
D | convolution_1x1_int8.cc | 116 …size_t size = support_optimize_ ? UP_ROUND(matmul_param_->row_, C8NUM) * UP_ROUND(matmul_param_->d… in InitRunBuf() 117 … : UP_ROUND(matmul_param_->row_, C4NUM) * UP_ROUND(matmul_param_->deep_, C16NUM); in InitRunBuf() 339 matmul_param_->row_ = conv_param_->output_h_ * conv_param_->output_w_; in InitParam() 342 matmul_param_->row_4_ = UP_ROUND(matmul_param_->row_, C4NUM); in InitParam() 363 input_sum_size_ = UP_ROUND(matmul_param_->row_, row_pack_count); in InitParam() 366 …input_ptr_ = reinterpret_cast<int8_t *>(malloc(matmul_param_->row_ * matmul_param_->deep_ * sizeof… in InitParam() 371 memset(input_ptr_, 0, matmul_param_->row_ * matmul_param_->deep_ * sizeof(int8_t)); in InitParam() 374 int hw_thread_count = UP_DIV(matmul_param_->row_, row_pack_count); in InitParam() 423 int res_stride = matmul_param_->row_ - task_id * thread_stride_hw_ * C4NUM; in RunArmHw() 451 int res_stride = matmul_param_->row_ - task_id * thread_stride_hw_ * C4NUM; in RunArm64OptHw() [all …]
|
D | matmul_base_int8.cc | 50 …pack_a_ptr_, batch_b_ptr_ + cur_stride * param_->deep_16_, batch_c_ptr_ + cur_stride, param_->row_, in RunImpl() 178 param_->row_align_ = UP_ROUND(param_->row_, row_tile_); in ResizeParameter() 328 auto current_src_a = a_ptr + i * param_->row_ * param_->deep_; in Run() 330 RowMajor2Col16x4MajorInt8(current_src_a, param_->deep_, param_->row_, pack_a_ptr_); in Run() 331 … CalcInputSums(current_src_a, param_->row_, param_->deep_, tmp_weight_zp, input_sums_, ColMajor); in Run() 333 RowMajor2Row16x4MajorInt8(current_src_a, pack_a_ptr_, param_->row_, param_->deep_); in Run() 334 … CalcInputSums(current_src_a, param_->row_, param_->deep_, tmp_weight_zp, input_sums_, RowMajor); in Run() 339 batch_c_ptr_ = c_ptr + i * param_->row_ * param_->col_; in Run()
|
D | deconvolution_int8.cc | 132 matmul_param_->row_ = conv_param_->input_h_ * conv_param_->input_w_; in InitParam() 211 size = UP_ROUND(matmul_param_->row_, C4NUM); in InitRunBuf() 264 UP_ROUND(matmul_param_->row_, C4NUM), cur_oc * C4NUM * kernel_plane, in DoDeconv() 290 …RowMajor2Row16x4MajorInt8(src_in + batch_index * matmul_param_->row_ * conv_param_->input_channel_… in Run() 291 matmul_param_->row_, matmul_param_->deep_); in Run() 295 … UP_ROUND(matmul_param_->row_, C4NUM), UP_ROUND(matmul_param_->deep_, C16NUM), support_optimize_); in Run()
|
/third_party/flutter/skia/third_party/externals/libwebp/src/utils/ |
D | quant_levels_dec_utils.c | 48 int row_; // current input row being processed member 103 if (p->row_ >= 0 && p->row_ < p->height_ - 1) { in VFilter() 139 const uint8_t* const dither = kOrderedDither[p->row_ % DSIZE]; in ApplyFilter() 246 p->row_ = -radius; in InitParams() 278 for (; p.row_ < p.height_; ++p.row_) { in WebPDequantizeLevels() 282 if (p.row_ >= p.radius_) { in WebPDequantizeLevels()
|
/third_party/skia/third_party/externals/libwebp/src/utils/ |
D | quant_levels_dec_utils.c | 48 int row_; // current input row being processed member 103 if (p->row_ >= 0 && p->row_ < p->height_ - 1) { in VFilter() 139 const uint8_t* const dither = kOrderedDither[p->row_ % DSIZE]; in ApplyFilter() 246 p->row_ = -radius; in InitParams() 278 for (; p.row_ < p.height_; ++p.row_) { in WebPDequantizeLevels() 282 if (p.row_ >= p.radius_) { in WebPDequantizeLevels()
|
/third_party/mindspore/mindspore/ccsrc/minddata/dataset/include/dataset/ |
D | iterator.h | 70 MSTensorMapChar row_; in GetNextRow() local 71 row_.clear(); in GetNextRow() 73 Status s = GetNextRowCharIF(&row_); in GetNextRow() 74 TensorMapCharToString(&row_, row); in GetNextRow()
|
/third_party/mindspore/mindspore/lite/micro/coder/opcoders/nnacl/int8/ |
D | conv2d_1x1_int8_coder.cc | 85 std::to_string(matmul_param_->row_ * matmul_param_->col_); in DoCode() 238 matmul_param_->row_ = conv_param_->output_h_ * conv_param_->output_w_; in InitParam() 241 matmul_param_->row_4_ = UP_ROUND(matmul_param_->row_, C4NUM); in InitParam() 247 input_sum_size_ = UP_ROUND(matmul_param_->row_, row_pack_count); in InitParam() 251 …allocator_->Malloc(kNumberTypeInt8, matmul_param_->row_ * matmul_param_->deep_ * sizeof(int8_t), k… in InitParam() 263 size_t size = MSMAX(UP_ROUND(matmul_param_->row_, C8NUM) * UP_ROUND(matmul_param_->deep_, C4NUM), in InitRunBuf() 264 … UP_ROUND(matmul_param_->row_, C4NUM) * UP_ROUND(matmul_param_->deep_, C16NUM)); in InitRunBuf()
|
D | matmul_base_int8_coder.cc | 61 param_->row_align_ = UP_ROUND(param_->row_, row_tile_); in ResizeParameter() 222 std::string current_src_a = a_ptr_str + "+" + std::to_string(i * param_->row_ * param_->deep_); in DoCode() 224 …code.CodeFunction("RowMajor2Col16x4MajorInt8", current_src_a, param_->deep_, param_->row_, pack_a_… in DoCode() 225 …code.CodeFunction("CalcInputSums", current_src_a, param_->row_, param_->deep_, "tmp_weight_zp", in… in DoCode() 228 …code.CodeFunction("RowMajor2Row16x4MajorInt8", current_src_a, pack_a_ptr_, param_->row_, param_->d… in DoCode() 229 …code.CodeFunction("CalcInputSums", current_src_a, param_->row_, param_->deep_, "tmp_weight_zp", in… in DoCode() 233 std::string batch_c_ptr_str = c_ptr_str + "+" + std::to_string(i * param_->row_ * param_->col_); in DoCode() 261 …ction("MatmulInt8Opt", pack_a_ptr_, batch_b_ptr_str_final, batch_c_ptr_final, param_->row_, cur_oc, in DoCode()
|
D | deconvolution_int8_coder.cc | 53 matmul_param_->row_ = conv_param_->input_h_ * conv_param_->input_w_; in InitParam() 118 input_sum_size_ = UP_ROUND(matmul_param_->row_, C4NUM) * sizeof(int32_t); in InitRunBuf() 145 …code.CodeFunction("RowMajor2Row16x4MajorInt8", input_tensor_, input_ptr_, matmul_param_->row_, mat… in DoCode() 147 … conv_param_->conv_quant_arg_.filter_quant_args_[0].zp_, UP_ROUND(matmul_param_->row_, C4NUM), in DoCode() 157 UP_ROUND(matmul_param_->row_, C4NUM), cur_oc * C4NUM * kernel_plane, in DoCode()
|
/third_party/mindspore/mindspore/lite/micro/coder/opcoders/nnacl/fp32/ |
D | deconv2d_fp32_coder.cc | 58 matmul_param_.row_ = input_plane_; in InitParam() 61 matmul_param_.row_12_ = UP_ROUND(matmul_param_.row_, C12NUM); in InitParam() 62 matmul_param_.row_4_ = UP_ROUND(matmul_param_.row_, C4NUM); in InitParam() 180 …code.CodeFunction("RowMajor2Col4Major", input_ptr_, packed_input_, matmul_param_.row_, matmul_para… in DoCode() 182 …code.CodeFunction("RowMajor2Col12Major", input_ptr_, packed_input_, matmul_param_.row_, matmul_par… in DoCode()
|
D | full_connection_fp32_coder.cc | 29 params_->row_ = row; in ReSize() 51 params_->row_ = a_shape.at(0); in Init() 64 if (params_->row_ == 1 && !params_->b_const_) { in Init()
|
D | matmul_fp32_base_coder.cc | 71 if (params_->row_ == 1) { in ResizeParameter() 74 params_->row_align_ = vec_matmul_ ? 1 : UP_ROUND(params_->row_, row_tile_); in ResizeParameter() 241 …code << "\t\tfloat *batch_c_ptr = " << c_str << " + i * " << params_->row_ * params_->col_ << ";\n… in DoCode() 245 …code << "\t\tfloat *batch_c_ptr = " << c_str << " + i * " << params_->row_ * params_->col_ << ";\n… in DoCode() 253 params_->deep_, params_->row_, cur_oc, params_->col_, "OutType_Nhwc"); in DoCode()
|
/third_party/mindspore/mindspore/lite/micro/coder/wrapper/fp32/ |
D | matmul_fp32_wrapper.c | 24 const float *src = src_ptr + i * params_->deep_ * params_->row_; in InitMatrixA() 27 RowMajor2Row12Major(src, dst, params_->deep_, params_->row_); in InitMatrixA() 29 RowMajor2Col12Major(src, dst, params_->row_, params_->deep_); in InitMatrixA()
|
/third_party/mindspore/mindspore/lite/micro/coder/wrapper/int8/ |
D | conv1x1_run_int8_wrapper.c | 36 int res_stride = args->matmul_param_->row_ - task_id * args->thread_stride_hw_ * C4NUM; in OcOptPre() 75 … args->matmul_param_->row_, cur_oc, args->matmul_param_->deep_4_, cur_left_shift, cur_right_shift, in RunArm64OptOc() 106 … args->matmul_param_->row_, cur_oc, args->matmul_param_->deep_16_, cur_left_shift, cur_right_shift, in RunArmOc() 114 int res_stride = args->matmul_param_->row_ - task_id * args->thread_stride_hw_ * C4NUM; in RunArm64OptHw() 140 int res_stride = args->matmul_param_->row_ - task_id * args->thread_stride_hw_ * C4NUM; in RunArmHw() 180 int hw_thread_count = UP_DIV(args->matmul_param_->row_, row_pack_count); in Conv1x1PreRun()
|
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/fp16/ |
D | conv_fp16.c | 122 int input_block = UP_DIV(param->row_, tile_n); in Conv1x1OutNc8hw8MultiThreadByInputFp16() 134 int cur_row_cnt = MSMIN(block_per_thread * tile_n, param->row_ - input_start_block * tile_n); in Conv1x1OutNc8hw8MultiThreadByInputFp16() 141 int real_in_row = (i != input_block - 1) ? tile_n : param->row_ - i * tile_n; in Conv1x1OutNc8hw8MultiThreadByInputFp16() 146 int out_offset = j * param->row_ * C8NUM + i * tile_n * real_weight_row; in Conv1x1OutNc8hw8MultiThreadByInputFp16() 164 int input_block = UP_DIV(param->row_, tile_n); in Conv1x1OutNc8hw8MultiThreadByWeightFp16() 174 int real_in_row = (i != input_block - 1) ? tile_n : param->row_ - i * tile_n; in Conv1x1OutNc8hw8MultiThreadByWeightFp16() 179 int out_offset = j * param->row_ * C8NUM + i * tile_n * real_weight_row; in Conv1x1OutNc8hw8MultiThreadByWeightFp16()
|