1 // Copyright (c) 2017-2023, Lawrence Livermore National Security, LLC and other CEED contributors. 2 // All Rights Reserved. See the top-level LICENSE and NOTICE files for details. 3 // 4 // SPDX-License-Identifier: BSD-2-Clause 5 // 6 // This file is part of CEED: http://github.com/ceed 7 8 #include "../../qfunctions/sgs_dd_training.h" 9 10 #include <petscdmplex.h> 11 12 #include "../../include/smartsim.h" 13 #include "../../navierstokes.h" 14 15 typedef struct { 16 CeedElemRestriction elem_restr_grid_aniso; 17 CeedVector grid_aniso_ceed; 18 CeedQFunctionContext sgs_dd_train_qfctx; 19 } *SGS_DD_TrainingSetupData; 20 21 static PetscErrorCode SGS_DD_TrainingSetupDataDestroy(SGS_DD_TrainingSetupData sgs_dd_train_setup_data) { 22 Ceed ceed; 23 24 PetscFunctionBeginUser; 25 PetscCall(CeedElemRestrictionGetCeed(sgs_dd_train_setup_data->elem_restr_grid_aniso, &ceed)); 26 27 PetscCallCeed(ceed, CeedElemRestrictionDestroy(&sgs_dd_train_setup_data->elem_restr_grid_aniso)); 28 PetscCallCeed(ceed, CeedVectorDestroy(&sgs_dd_train_setup_data->grid_aniso_ceed)); 29 PetscCallCeed(ceed, CeedQFunctionContextDestroy(&sgs_dd_train_setup_data->sgs_dd_train_qfctx)); 30 PetscCall(PetscFree(sgs_dd_train_setup_data)); 31 PetscFunctionReturn(PETSC_SUCCESS); 32 } 33 34 // @brief Create DM for storing data-drive SGS model inputs 35 static PetscErrorCode SGS_DD_TrainingCreateDM(DM dm_source, DM *dm_dd_training, PetscInt degree, PetscInt q_extra, PetscInt *num_components) { 36 PetscSection section; 37 38 PetscFunctionBeginUser; 39 *num_components = 12; 40 41 PetscCall(DMClone(dm_source, dm_dd_training)); 42 PetscCall(PetscObjectSetName((PetscObject)*dm_dd_training, "Data-Driven SGS Training Data")); 43 44 PetscCall(DMSetupByOrder_FEM(PETSC_TRUE, PETSC_TRUE, degree, 1, q_extra, 1, num_components, *dm_dd_training)); 45 46 PetscCall(DMGetLocalSection(*dm_dd_training, §ion)); 47 PetscCall(PetscSectionSetFieldName(section, 0, "Data-Driven SGS Training Data")); 48 PetscCall(PetscSectionSetComponentName(section, 0, 0, "SGSInput1")); 49 PetscCall(PetscSectionSetComponentName(section, 0, 1, "SGSInput2")); 50 PetscCall(PetscSectionSetComponentName(section, 0, 2, "SGSInput3")); 51 PetscCall(PetscSectionSetComponentName(section, 0, 3, "SGSInput4")); 52 PetscCall(PetscSectionSetComponentName(section, 0, 4, "SGSInput5")); 53 PetscCall(PetscSectionSetComponentName(section, 0, 5, "SGSInput6")); 54 PetscCall(PetscSectionSetComponentName(section, 0, 6, "FilteredSGSXX")); 55 PetscCall(PetscSectionSetComponentName(section, 0, 7, "FilteredSGSYY")); 56 PetscCall(PetscSectionSetComponentName(section, 0, 8, "FilteredSGSZZ")); 57 PetscCall(PetscSectionSetComponentName(section, 0, 9, "FilteredSGSYZ")); 58 PetscCall(PetscSectionSetComponentName(section, 0, 10, "FilteredSGSXZ")); 59 PetscCall(PetscSectionSetComponentName(section, 0, 11, "FilteredSGSXY")); 60 PetscFunctionReturn(PETSC_SUCCESS); 61 }; 62 63 // @brief Create CeedOperator to calculate training data for data-drive SGS model at nodes 64 static PetscErrorCode SetupTrainingDataCalculation(Ceed ceed, User user, CeedData ceed_data, ProblemData *problem, 65 SGS_DD_TrainingSetupData sgs_dd_train_setup_data) { 66 SGS_DD_TrainingData sgs_dd_train = user->sgs_dd_train; 67 CeedQFunction qf_sgs_dd_train; 68 CeedOperator op_sgs_dd_train; 69 CeedInt num_comp_grad_velo, num_comp_grid_aniso; 70 CeedVector inv_multiplicity, filtered_fields; 71 CeedElemRestriction elem_restr_inv_multiplicity, elem_restr_grad_velo, elem_restr_sgs_train; 72 DMLabel domain_label = NULL; 73 PetscInt label_value = 0, height = 0, dm_field = 0; 74 75 PetscFunctionBeginUser; 76 PetscCallCeed(ceed, CeedElemRestrictionGetNumComponents(sgs_dd_train_setup_data->elem_restr_grid_aniso, &num_comp_grid_aniso)); 77 78 PetscCall(DMPlexCeedElemRestrictionCreate(ceed, sgs_dd_train->dm_dd_training, domain_label, label_value, height, dm_field, &elem_restr_sgs_train)); 79 80 { // -- Create inverse multiplicity for correcting nodal assembly 81 CeedVector multiplicity; 82 CeedQFunction qf_multiplicity; 83 CeedOperator op_multiplicity; 84 CeedInt num_comp_q; 85 86 PetscCallCeed(ceed, CeedElemRestrictionGetNumComponents(ceed_data->elem_restr_q, &num_comp_q)); 87 PetscCallCeed(ceed, CeedElemRestrictionCreateVector(ceed_data->elem_restr_q, &multiplicity, NULL)); 88 PetscCallCeed(ceed, CeedElemRestrictionGetMultiplicity(ceed_data->elem_restr_q, multiplicity)); 89 PetscCall(DMPlexCeedElemRestrictionCollocatedCreate(ceed, sgs_dd_train->dm_dd_training, domain_label, label_value, height, 1, 90 &elem_restr_inv_multiplicity)); 91 PetscCallCeed(ceed, CeedElemRestrictionCreateVector(elem_restr_inv_multiplicity, &inv_multiplicity, NULL)); 92 93 PetscCallCeed(ceed, CeedQFunctionCreateInterior(ceed, 1, InverseMultiplicity, InverseMultiplicity_loc, &qf_multiplicity)); 94 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_multiplicity, "multiplicity", num_comp_q, CEED_EVAL_NONE)); 95 PetscCallCeed(ceed, CeedQFunctionAddOutput(qf_multiplicity, "inverse multiplicity", 1, CEED_EVAL_NONE)); 96 97 PetscCallCeed(ceed, CeedOperatorCreate(ceed, qf_multiplicity, NULL, NULL, &op_multiplicity)); 98 PetscCallCeed(ceed, CeedOperatorSetName(op_multiplicity, "SGS DD Training Inputs - Create Multiplicity Scaling")); 99 PetscCallCeed(ceed, CeedOperatorSetField(op_multiplicity, "multiplicity", ceed_data->elem_restr_q, CEED_BASIS_NONE, CEED_VECTOR_ACTIVE)); 100 PetscCallCeed(ceed, 101 CeedOperatorSetField(op_multiplicity, "inverse multiplicity", elem_restr_inv_multiplicity, CEED_BASIS_NONE, CEED_VECTOR_ACTIVE)); 102 103 PetscCallCeed(ceed, CeedOperatorApply(op_multiplicity, multiplicity, inv_multiplicity, CEED_REQUEST_IMMEDIATE)); 104 105 PetscCallCeed(ceed, CeedVectorDestroy(&multiplicity)); 106 PetscCallCeed(ceed, CeedQFunctionDestroy(&qf_multiplicity)); 107 PetscCallCeed(ceed, CeedOperatorDestroy(&op_multiplicity)); 108 } 109 110 CeedElemRestriction elem_restr_filtered_state; 111 CeedInt num_comp_filtered_state; 112 { // -- Setup filtered velocity gradient projection 113 CeedBasis basis_filtered_state; 114 CeedOperatorField op_field; 115 PetscCallCeed(ceed, CeedOperatorGetFieldByName(user->diff_filter->op_rhs_ctx->op, "v0", &op_field)); 116 PetscCallCeed(ceed, CeedOperatorFieldGetElemRestriction(op_field, &elem_restr_filtered_state)); 117 PetscCallCeed(ceed, CeedElemRestrictionGetNumComponents(elem_restr_filtered_state, &num_comp_filtered_state)); 118 PetscCallCeed(ceed, CeedOperatorFieldGetBasis(op_field, &basis_filtered_state)); 119 PetscCall(VelocityGradientProjectionSetup(ceed, user, ceed_data, problem, STATEVAR_PRIMITIVE, elem_restr_filtered_state, basis_filtered_state, 120 &sgs_dd_train->filtered_grad_velo_proj)); 121 // Get velocity gradient information 122 PetscCallCeed(ceed, CeedOperatorGetFieldByName(sgs_dd_train->filtered_grad_velo_proj->l2_rhs_ctx->op, "velocity gradient", &op_field)); 123 PetscCallCeed(ceed, CeedOperatorFieldGetElemRestriction(op_field, &elem_restr_grad_velo)); 124 PetscCallCeed(ceed, CeedElemRestrictionGetNumComponents(elem_restr_grad_velo, &num_comp_grad_velo)); 125 } 126 127 CeedElemRestriction elem_restr_filtered_velo_prod; 128 CeedInt num_comp_filtered_velo_prod; 129 { // Get filtered velocity product information 130 CeedOperatorField op_field; 131 PetscCallCeed(ceed, CeedOperatorGetFieldByName(user->diff_filter->op_rhs_ctx->op, "v1", &op_field)); 132 PetscCallCeed(ceed, CeedOperatorFieldGetElemRestriction(op_field, &elem_restr_filtered_velo_prod)); 133 PetscCallCeed(ceed, CeedElemRestrictionGetNumComponents(elem_restr_filtered_velo_prod, &num_comp_filtered_velo_prod)); 134 } 135 136 // -- Create operator for generating training data at nodes 137 // Differential Filter only provides filtered primitive variables 138 PetscCallCeed(ceed, CeedQFunctionCreateInterior(ceed, 1, ComputeSGS_DDAnisotropicTrainingDataNodal_Prim, 139 ComputeSGS_DDAnisotropicTrainingDataNodal_Prim_loc, &qf_sgs_dd_train)); 140 141 PetscCallCeed(ceed, CeedQFunctionSetContext(qf_sgs_dd_train, sgs_dd_train_setup_data->sgs_dd_train_qfctx)); 142 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_sgs_dd_train, "q", num_comp_filtered_state, CEED_EVAL_NONE)); 143 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_sgs_dd_train, "velocity product", num_comp_filtered_velo_prod, CEED_EVAL_NONE)); 144 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_sgs_dd_train, "gradient velocity", num_comp_grad_velo, CEED_EVAL_NONE)); 145 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_sgs_dd_train, "anisotropy tensor", num_comp_grid_aniso, CEED_EVAL_NONE)); 146 PetscCallCeed(ceed, CeedQFunctionAddInput(qf_sgs_dd_train, "inverse multiplicity", 1, CEED_EVAL_NONE)); 147 PetscCallCeed(ceed, CeedQFunctionAddOutput(qf_sgs_dd_train, "training data", sgs_dd_train->num_comp_dd_inputs, CEED_EVAL_NONE)); 148 149 PetscCallCeed(ceed, CeedElemRestrictionCreateVector(elem_restr_filtered_state, &filtered_fields, NULL)); 150 PetscCallCeed(ceed, CeedOperatorCreate(ceed, qf_sgs_dd_train, NULL, NULL, &op_sgs_dd_train)); 151 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "q", elem_restr_filtered_state, CEED_BASIS_NONE, filtered_fields)); 152 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "velocity product", elem_restr_filtered_velo_prod, CEED_BASIS_NONE, filtered_fields)); 153 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "gradient velocity", elem_restr_grad_velo, CEED_BASIS_NONE, CEED_VECTOR_ACTIVE)); 154 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "anisotropy tensor", sgs_dd_train_setup_data->elem_restr_grid_aniso, CEED_BASIS_NONE, 155 sgs_dd_train_setup_data->grid_aniso_ceed)); 156 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "inverse multiplicity", elem_restr_inv_multiplicity, CEED_BASIS_NONE, inv_multiplicity)); 157 PetscCallCeed(ceed, CeedOperatorSetField(op_sgs_dd_train, "training data", elem_restr_sgs_train, CEED_BASIS_NONE, CEED_VECTOR_ACTIVE)); 158 159 PetscCall(OperatorApplyContextCreate(sgs_dd_train->filtered_grad_velo_proj->dm, sgs_dd_train->dm_dd_training, ceed, op_sgs_dd_train, NULL, NULL, 160 NULL, NULL, &sgs_dd_train->op_training_data_calc_ctx)); 161 162 PetscCallCeed(ceed, CeedVectorDestroy(&inv_multiplicity)); 163 PetscCallCeed(ceed, CeedVectorDestroy(&filtered_fields)); 164 PetscCallCeed(ceed, CeedElemRestrictionDestroy(&elem_restr_inv_multiplicity)); 165 PetscCallCeed(ceed, CeedQFunctionDestroy(&qf_sgs_dd_train)); 166 PetscCallCeed(ceed, CeedOperatorDestroy(&op_sgs_dd_train)); 167 PetscFunctionReturn(PETSC_SUCCESS); 168 } 169 170 PetscErrorCode SGS_DD_TrainingSetup(Ceed ceed, User user, CeedData ceed_data, ProblemData *problem) { 171 SGS_DDTrainingContext sgsdd_train_qfctx; 172 SGS_DD_TrainingSetupData sgs_dd_train_setup_data; 173 174 PetscFunctionBeginUser; 175 if (!user->diff_filter) PetscCall(DifferentialFilterSetup(ceed, user, ceed_data, problem)); 176 if (!user->smartsim) PetscCall(SmartSimSetup(user)); 177 178 PetscCall(PetscNew(&sgsdd_train_qfctx)); 179 PetscCall(PetscNew(&sgs_dd_train_setup_data)); 180 PetscCall(PetscNew(&user->sgs_dd_train)); 181 SGS_DD_TrainingData sgs_dd_train = user->sgs_dd_train; 182 183 sgs_dd_train->overwrite_training_data = PETSC_TRUE; 184 sgs_dd_train->write_data_interval = 1; 185 PetscOptionsBegin(user->comm, NULL, "SGS Data-Driven Training Options", NULL); 186 PetscCall(PetscOptionsInt("-sgs_train_write_data_interval", "Number of timesteps between writing data into database", NULL, 187 sgs_dd_train->write_data_interval, &sgs_dd_train->write_data_interval, NULL)); 188 PetscCall(PetscOptionsBool("-sgs_train_overwrite_data", "Overwrite old training data in the database", NULL, sgs_dd_train->overwrite_training_data, 189 &sgs_dd_train->overwrite_training_data, NULL)); 190 PetscOptionsEnd(); 191 192 // -- Create DM for storing training data 193 PetscCall(SGS_DD_TrainingCreateDM(user->dm, &sgs_dd_train->dm_dd_training, user->app_ctx->degree, user->app_ctx->q_extra, 194 &sgs_dd_train->num_comp_dd_inputs)); 195 196 { // -- Create QFunction Context 197 NewtonianIdealGasContext gas; 198 PetscCallCeed(ceed, CeedQFunctionContextGetDataRead(problem->apply_vol_ifunction.qfunction_context, CEED_MEM_HOST, &gas)); 199 sgsdd_train_qfctx->gas = *gas; 200 PetscCallCeed(ceed, CeedQFunctionContextRestoreDataRead(problem->apply_vol_ifunction.qfunction_context, &gas)); 201 PetscCallCeed(ceed, CeedQFunctionContextCreate(user->ceed, &sgs_dd_train_setup_data->sgs_dd_train_qfctx)); 202 PetscCallCeed(ceed, CeedQFunctionContextSetData(sgs_dd_train_setup_data->sgs_dd_train_qfctx, CEED_MEM_HOST, CEED_USE_POINTER, 203 sizeof(*sgsdd_train_qfctx), sgsdd_train_qfctx)); 204 PetscCallCeed(ceed, CeedQFunctionContextSetDataDestroy(sgs_dd_train_setup_data->sgs_dd_train_qfctx, CEED_MEM_HOST, FreeContextPetsc)); 205 } 206 207 { // -- Send training data array info to SmartRedis database 208 PetscMPIInt rank, num_ranks; 209 SmartSimData smartsim = user->smartsim; 210 PetscCallMPI(MPI_Comm_rank(user->comm, &rank)); 211 PetscCallMPI(MPI_Comm_size(user->comm, &num_ranks)); 212 213 { 214 PetscSection global_section; 215 PetscInt num_dofs, num_comps; 216 PetscCall(DMGetGlobalSection(sgs_dd_train->dm_dd_training, &global_section)); 217 PetscCall(DMGetGlobalVectorInfo(sgs_dd_train->dm_dd_training, &num_dofs, NULL, NULL)); 218 PetscCall(PetscSectionGetFieldComponents(global_section, 0, &num_comps)); 219 sgs_dd_train->training_data_array_dims[0] = num_dofs / num_comps; 220 sgs_dd_train->training_data_array_dims[1] = num_comps; 221 } 222 223 if (rank % smartsim->collocated_database_num_ranks == 0) { 224 size_t array_info_dim = 6; 225 PetscInt64 array_info[6] = {0}, num_features = 6; 226 227 array_info[0] = sgs_dd_train->training_data_array_dims[0]; 228 array_info[1] = sgs_dd_train->training_data_array_dims[1]; 229 array_info[2] = num_features; 230 array_info[3] = num_ranks; 231 array_info[4] = smartsim->collocated_database_num_ranks; 232 array_info[5] = rank; 233 234 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 235 PetscSmartRedisCall(put_tensor(smartsim->client, "sizeInfo", 8, array_info, &array_info_dim, 1, SRTensorTypeInt64, SRMemLayoutContiguous)); 236 PetscCall(SmartRedisVerifyPutTensor(smartsim->client, "sizeInfo", 8)); 237 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 238 239 // -- Send array that communicates if tensors are overwritten in database 240 PetscInt64 tensor_overwrite[2] = {sgs_dd_train->overwrite_training_data}; 241 size_t dim_2[1] = {2}; 242 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 243 PetscSmartRedisCall(put_tensor(smartsim->client, "tensor-ow", 9, tensor_overwrite, dim_2, 1, SRTensorTypeInt64, SRMemLayoutContiguous)); 244 PetscCall(SmartRedisVerifyPutTensor(smartsim->client, "tensor-ow", 9)); 245 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 246 } 247 } 248 249 // -- Compute and store anisotropy tensor 250 PetscCall(GridAnisotropyTensorProjectionSetupApply(ceed, user, ceed_data, &sgs_dd_train_setup_data->elem_restr_grid_aniso, 251 &sgs_dd_train_setup_data->grid_aniso_ceed)); 252 253 // -- Create Nodal Evaluation Operator 254 PetscCall(SetupTrainingDataCalculation(ceed, user, ceed_data, problem, sgs_dd_train_setup_data)); 255 256 PetscCall(SGS_DD_TrainingSetupDataDestroy(sgs_dd_train_setup_data)); 257 PetscFunctionReturn(PETSC_SUCCESS); 258 } 259 260 PetscErrorCode TSMonitor_SGS_DD_Training(TS ts, PetscInt step_num, PetscReal solution_time, Vec Q, void *ctx) { 261 User user = (User)ctx; 262 Ceed ceed = user->ceed; 263 SGS_DD_TrainingData sgs_dd_train = user->sgs_dd_train; 264 SmartSimData smartsim = user->smartsim; 265 Vec TrainingData; 266 267 PetscFunctionBeginUser; 268 if (step_num % sgs_dd_train->write_data_interval != 0) PetscFunctionReturn(PETSC_SUCCESS); 269 PetscCall(DMGetGlobalVector(sgs_dd_train->dm_dd_training, &TrainingData)); 270 271 PetscCall(PetscLogEventBegin(FLUIDS_TrainDataCompute, 0, 0, 0, 0)); 272 { // -- Compute and assemble training data 273 Vec FilteredVelocityGradient, FilteredFields, FilteredFields_loc; 274 PetscMemType filtered_fields_mem_type; 275 CeedVector filtered_fields; 276 277 PetscCall(DMGetGlobalVector(user->diff_filter->dm_filter, &FilteredFields)); 278 PetscCall(DMGetLocalVector(user->diff_filter->dm_filter, &FilteredFields_loc)); 279 280 PetscCall(DifferentialFilterApply(user, solution_time, Q, FilteredFields)); 281 PetscCall(DMGlobalToLocal(user->diff_filter->dm_filter, FilteredFields, INSERT_VALUES, FilteredFields_loc)); 282 283 PetscCall(DMGetGlobalVector(sgs_dd_train->filtered_grad_velo_proj->dm, &FilteredVelocityGradient)); 284 PetscCall(VelocityGradientProjectionApply(sgs_dd_train->filtered_grad_velo_proj, FilteredFields_loc, FilteredVelocityGradient)); 285 286 { 287 CeedOperatorField op_field; 288 PetscCallCeed(ceed, CeedOperatorGetFieldByName(sgs_dd_train->op_training_data_calc_ctx->op, "q", &op_field)); 289 PetscCallCeed(ceed, CeedOperatorFieldGetVector(op_field, &filtered_fields)); 290 } 291 PetscCall(VecP2C(FilteredFields_loc, &filtered_fields_mem_type, filtered_fields)); // filtered_fields is an implicit input 292 293 PetscCall(ApplyCeedOperatorGlobalToGlobal(FilteredVelocityGradient, TrainingData, sgs_dd_train->op_training_data_calc_ctx)); 294 295 PetscCall(VecC2P(filtered_fields, filtered_fields_mem_type, FilteredFields_loc)); 296 297 PetscCall(DMRestoreGlobalVector(sgs_dd_train->filtered_grad_velo_proj->dm, &FilteredVelocityGradient)); 298 PetscCall(DMRestoreGlobalVector(user->diff_filter->dm_filter, &FilteredFields)); 299 PetscCall(DMRestoreLocalVector(user->diff_filter->dm_filter, &FilteredFields_loc)); 300 } 301 PetscCall(PetscLogEventEnd(FLUIDS_TrainDataCompute, 0, 0, 0, 0)); 302 303 { // -- Send training data to SmartSim 304 char array_key[PETSC_MAX_PATH_LEN]; 305 size_t array_key_len; 306 PetscMPIInt rank; 307 308 PetscCallMPI(MPI_Comm_rank(user->comm, &rank)); 309 310 if (sgs_dd_train->overwrite_training_data) { 311 PetscCall(PetscSNPrintf(array_key, sizeof array_key, "%s", smartsim->rank_id_name)); 312 } else { 313 PetscCall(PetscSNPrintf(array_key, sizeof array_key, "%s.%" PetscInt_FMT, smartsim->rank_id_name, step_num)); 314 } 315 PetscCall(PetscStrlen(array_key, &array_key_len)); 316 317 { 318 const PetscScalar *training_data; 319 PetscCall(VecGetArrayRead(TrainingData, &training_data)); 320 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Train, 0, 0, 0, 0)); 321 PetscSmartRedisCall(put_tensor(smartsim->client, array_key, array_key_len, (void *)training_data, sgs_dd_train->training_data_array_dims, 2, 322 SRTensorTypeDouble, SRMemLayoutContiguous)); 323 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Train, 0, 0, 0, 0)); 324 PetscCall(VecRestoreArrayRead(TrainingData, &training_data)); 325 } 326 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 327 PetscCall(SmartRedisVerifyPutTensor(smartsim->client, array_key, array_key_len)); 328 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 329 330 if (rank % smartsim->collocated_database_num_ranks == 0) { 331 size_t dim_2[1] = {2}; 332 PetscInt64 step_array[2] = {step_num, step_num}; 333 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 334 PetscSmartRedisCall(put_tensor(smartsim->client, "step", 4, step_array, dim_2, 1, SRTensorTypeInt64, SRMemLayoutContiguous)); 335 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 336 } 337 } 338 339 PetscCall(DMRestoreGlobalVector(user->sgs_dd_train->dm_dd_training, &TrainingData)); 340 PetscFunctionReturn(PETSC_SUCCESS); 341 } 342 343 PetscErrorCode TSPostStep_SGS_DD_Training(TS ts) { 344 User user; 345 const char check_run_key[] = "check-run"; 346 PetscReal check_run[2] = {1}; 347 const size_t check_run_dims[1] = {2}; 348 size_t check_run_key_size; 349 350 PetscFunctionBeginUser; 351 PetscCall(PetscStrlen(check_run_key, &check_run_key_size)); 352 PetscCall(TSGetApplicationContext(ts, &user)); 353 SmartSimData smartsim = user->smartsim; 354 355 PetscCall(PetscLogEventBegin(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 356 PetscSmartRedisCall( 357 unpack_tensor(smartsim->client, check_run_key, check_run_key_size, check_run, check_run_dims, 1, SRTensorTypeDouble, SRMemLayoutContiguous)); 358 PetscCall(PetscLogEventEnd(FLUIDS_SmartRedis_Meta, 0, 0, 0, 0)); 359 if (check_run[0] == 0) { 360 PetscCall(PetscPrintf(user->comm, "-- Simulation stopped by 'check-run' tensor in Redis database\n")); 361 PetscCall(TSSetConvergedReason(ts, TS_CONVERGED_USER)); 362 } 363 364 PetscFunctionReturn(PETSC_SUCCESS); 365 } 366 367 PetscErrorCode SGS_DD_TrainingDataDestroy(SGS_DD_TrainingData sgs_dd_train) { 368 PetscFunctionBeginUser; 369 if (!sgs_dd_train) PetscFunctionReturn(PETSC_SUCCESS); 370 371 PetscCall(OperatorApplyContextDestroy(sgs_dd_train->op_training_data_calc_ctx)); 372 PetscCall(NodalProjectionDataDestroy(sgs_dd_train->filtered_grad_velo_proj)); 373 PetscCall(DMDestroy(&sgs_dd_train->dm_dd_training)); 374 PetscCall(PetscFree(sgs_dd_train)); 375 376 PetscFunctionReturn(PETSC_SUCCESS); 377 } 378