1 // Copyright (c) 2017-2025, 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 <ceed.h> 9 #include <ceed/backend.h> 10 #include <stdbool.h> 11 #include <stdlib.h> 12 #include <string.h> 13 14 #include "ceed-memcheck.h" 15 16 //------------------------------------------------------------------------------ 17 // Set backend strides 18 //------------------------------------------------------------------------------ 19 static inline int CeedElemRestrictionGetBackendStrides_Memcheck(CeedElemRestriction rstr, CeedInt strides[3]) { 20 CeedInt elem_size, num_comp, num_elem; 21 22 CeedCallBackend(CeedElemRestrictionGetElementSize(rstr, &elem_size)); 23 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 24 CeedCallBackend(CeedElemRestrictionGetNumElements(rstr, &num_elem)); 25 // Memcheck default, contiguous by component, then node 26 strides[0] = num_comp; 27 strides[1] = 1; 28 strides[2] = num_comp * elem_size; 29 /** 30 // CPU default, contiguous by node, then component 31 strides[0] = 1; 32 strides[1] = elem_size; 33 strides[2] = elem_size * num_comp; 34 35 // GPU default, contiguous by node, then element 36 strides[0] = 1; 37 strides[1] = num_elem * elem_size; 38 strides[2] = elem_size; 39 **/ 40 return CEED_ERROR_SUCCESS; 41 } 42 43 //------------------------------------------------------------------------------ 44 // Core ElemRestriction Apply Code 45 //------------------------------------------------------------------------------ 46 static inline int CeedElemRestrictionApplyStridedNoTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 47 CeedInt start, CeedInt stop, CeedInt num_elem, CeedInt elem_size, 48 CeedSize v_offset, const CeedScalar *__restrict__ uu, 49 CeedScalar *__restrict__ vv) { 50 // Get strides 51 bool has_backend_strides; 52 CeedInt strides[3] = {0}; 53 54 CeedCallBackend(CeedElemRestrictionHasBackendStrides(rstr, &has_backend_strides)); 55 if (has_backend_strides) CeedCallBackend(CeedElemRestrictionGetBackendStrides_Memcheck(rstr, strides)); 56 else CeedCallBackend(CeedElemRestrictionGetStrides(rstr, strides)); 57 58 // Apply restriction 59 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 60 for (CeedSize k = 0; k < num_comp; k++) { 61 for (CeedSize n = 0; n < elem_size; n++) { 62 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 63 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 64 uu[n * strides[0] + k * strides[1] + CeedIntMin(e + j, num_elem - 1) * (CeedSize)strides[2]]; 65 } 66 } 67 } 68 } 69 return CEED_ERROR_SUCCESS; 70 } 71 72 static inline int CeedElemRestrictionApplyOffsetNoTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 73 const CeedInt comp_stride, CeedInt start, CeedInt stop, CeedInt num_elem, 74 CeedInt elem_size, CeedSize v_offset, const CeedScalar *__restrict__ uu, 75 CeedScalar *__restrict__ vv) { 76 // Default restriction with offsets 77 CeedElemRestriction_Memcheck *impl; 78 79 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 80 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 81 for (CeedSize k = 0; k < num_comp; k++) { 82 CeedPragmaSIMD for (CeedSize i = 0; i < elem_size * block_size; i++) { 83 vv[elem_size * (k * block_size + e * num_comp) + i - v_offset] = uu[impl->offsets[i + e * elem_size] + k * comp_stride]; 84 } 85 } 86 } 87 return CEED_ERROR_SUCCESS; 88 } 89 90 static inline int CeedElemRestrictionApplyOrientedNoTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, 91 const CeedInt block_size, const CeedInt comp_stride, CeedInt start, 92 CeedInt stop, CeedInt num_elem, CeedInt elem_size, CeedSize v_offset, 93 const CeedScalar *__restrict__ uu, CeedScalar *__restrict__ vv) { 94 // Restriction with orientations 95 CeedElemRestriction_Memcheck *impl; 96 97 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 98 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 99 for (CeedSize k = 0; k < num_comp; k++) { 100 CeedPragmaSIMD for (CeedSize i = 0; i < elem_size * block_size; i++) { 101 vv[elem_size * (k * block_size + e * num_comp) + i - v_offset] = 102 uu[impl->offsets[i + e * elem_size] + k * comp_stride] * (impl->orients[i + e * elem_size] ? -1.0 : 1.0); 103 } 104 } 105 } 106 return CEED_ERROR_SUCCESS; 107 } 108 109 static inline int CeedElemRestrictionApplyCurlOrientedNoTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, 110 const CeedInt block_size, const CeedInt comp_stride, CeedInt start, 111 CeedInt stop, CeedInt num_elem, CeedInt elem_size, CeedSize v_offset, 112 const CeedScalar *__restrict__ uu, CeedScalar *__restrict__ vv) { 113 // Restriction with tridiagonal transformation 114 CeedElemRestriction_Memcheck *impl; 115 116 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 117 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 118 for (CeedSize k = 0; k < num_comp; k++) { 119 CeedSize n = 0; 120 121 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 122 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 123 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 124 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size] + 125 uu[impl->offsets[j + (n + 1) * block_size + e * elem_size] + k * comp_stride] * 126 impl->curl_orients[j + (3 * n + 2) * block_size + e * 3 * elem_size]; 127 } 128 for (n = 1; n < elem_size - 1; n++) { 129 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 130 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 131 uu[impl->offsets[j + (n - 1) * block_size + e * elem_size] + k * comp_stride] * 132 impl->curl_orients[j + (3 * n + 0) * block_size + e * 3 * elem_size] + 133 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 134 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size] + 135 uu[impl->offsets[j + (n + 1) * block_size + e * elem_size] + k * comp_stride] * 136 impl->curl_orients[j + (3 * n + 2) * block_size + e * 3 * elem_size]; 137 } 138 } 139 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 140 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 141 uu[impl->offsets[j + (n - 1) * block_size + e * elem_size] + k * comp_stride] * 142 impl->curl_orients[j + (3 * n + 0) * block_size + e * 3 * elem_size] + 143 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 144 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]; 145 } 146 } 147 } 148 return CEED_ERROR_SUCCESS; 149 } 150 151 static inline int CeedElemRestrictionApplyCurlOrientedUnsignedNoTranspose_Memcheck_Core( 152 CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, const CeedInt comp_stride, CeedInt start, CeedInt stop, 153 CeedInt num_elem, CeedInt elem_size, CeedSize v_offset, const CeedScalar *__restrict__ uu, CeedScalar *__restrict__ vv) { 154 // Restriction with (unsigned) tridiagonal transformation 155 CeedElemRestriction_Memcheck *impl; 156 157 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 158 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 159 for (CeedSize k = 0; k < num_comp; k++) { 160 CeedSize n = 0; 161 162 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 163 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 164 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 165 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]) + 166 uu[impl->offsets[j + (n + 1) * block_size + e * elem_size] + k * comp_stride] * 167 abs(impl->curl_orients[j + (3 * n + 2) * block_size + e * 3 * elem_size]); 168 } 169 for (n = 1; n < elem_size - 1; n++) { 170 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 171 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 172 uu[impl->offsets[j + (n - 1) * block_size + e * elem_size] + k * comp_stride] * 173 abs(impl->curl_orients[j + (3 * n + 0) * block_size + e * 3 * elem_size]) + 174 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 175 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]) + 176 uu[impl->offsets[j + (n + 1) * block_size + e * elem_size] + k * comp_stride] * 177 abs(impl->curl_orients[j + (3 * n + 2) * block_size + e * 3 * elem_size]); 178 } 179 } 180 CeedPragmaSIMD for (CeedSize j = 0; j < block_size; j++) { 181 vv[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] = 182 uu[impl->offsets[j + (n - 1) * block_size + e * elem_size] + k * comp_stride] * 183 abs(impl->curl_orients[j + (3 * n + 0) * block_size + e * 3 * elem_size]) + 184 uu[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] * 185 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]); 186 } 187 } 188 } 189 return CEED_ERROR_SUCCESS; 190 } 191 192 static inline int CeedElemRestrictionApplyStridedTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 193 CeedInt start, CeedInt stop, CeedInt num_elem, CeedInt elem_size, 194 CeedSize v_offset, const CeedScalar *__restrict__ uu, 195 CeedScalar *__restrict__ vv) { 196 // Get strides 197 bool has_backend_strides; 198 CeedInt strides[3] = {0}; 199 200 CeedCallBackend(CeedElemRestrictionHasBackendStrides(rstr, &has_backend_strides)); 201 if (has_backend_strides) CeedCallBackend(CeedElemRestrictionGetBackendStrides_Memcheck(rstr, strides)); 202 else CeedCallBackend(CeedElemRestrictionGetStrides(rstr, strides)); 203 204 // Apply restriction 205 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 206 for (CeedSize k = 0; k < num_comp; k++) { 207 for (CeedSize n = 0; n < elem_size; n++) { 208 CeedPragmaSIMD for (CeedSize j = 0; j < CeedIntMin(block_size, num_elem - e); j++) { 209 vv[n * strides[0] + k * strides[1] + (e + j) * strides[2]] += 210 uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset]; 211 } 212 } 213 } 214 } 215 return CEED_ERROR_SUCCESS; 216 } 217 218 static inline int CeedElemRestrictionApplyOffsetTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 219 const CeedInt comp_stride, CeedInt start, CeedInt stop, CeedInt num_elem, 220 CeedInt elem_size, CeedSize v_offset, const CeedScalar *__restrict__ uu, 221 CeedScalar *__restrict__ vv) { 222 // Default restriction with offsets 223 CeedElemRestriction_Memcheck *impl; 224 225 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 226 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 227 for (CeedSize k = 0; k < num_comp; k++) { 228 for (CeedSize i = 0; i < elem_size * block_size; i += block_size) { 229 // Iteration bound set to discard padding elements 230 for (CeedSize j = i; j < i + CeedIntMin(block_size, num_elem - e); j++) { 231 CeedScalar vv_loc; 232 233 vv_loc = uu[elem_size * (k * block_size + e * num_comp) + j - v_offset]; 234 CeedPragmaAtomic vv[impl->offsets[j + e * elem_size] + k * comp_stride] += vv_loc; 235 } 236 } 237 } 238 } 239 return CEED_ERROR_SUCCESS; 240 } 241 242 static inline int CeedElemRestrictionApplyOrientedTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 243 const CeedInt comp_stride, CeedInt start, CeedInt stop, CeedInt num_elem, 244 CeedInt elem_size, CeedSize v_offset, const CeedScalar *__restrict__ uu, 245 CeedScalar *__restrict__ vv) { 246 // Restriction with orientations 247 CeedElemRestriction_Memcheck *impl; 248 249 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 250 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 251 for (CeedSize k = 0; k < num_comp; k++) { 252 for (CeedSize i = 0; i < elem_size * block_size; i += block_size) { 253 // Iteration bound set to discard padding elements 254 for (CeedSize j = i; j < i + CeedIntMin(block_size, num_elem - e); j++) { 255 CeedScalar vv_loc; 256 257 vv_loc = uu[elem_size * (k * block_size + e * num_comp) + j - v_offset] * (impl->orients[j + e * elem_size] ? -1.0 : 1.0); 258 CeedPragmaAtomic vv[impl->offsets[j + e * elem_size] + k * comp_stride] += vv_loc; 259 } 260 } 261 } 262 } 263 return CEED_ERROR_SUCCESS; 264 } 265 266 static inline int CeedElemRestrictionApplyCurlOrientedTranspose_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, 267 const CeedInt block_size, const CeedInt comp_stride, CeedInt start, 268 CeedInt stop, CeedInt num_elem, CeedInt elem_size, CeedSize v_offset, 269 const CeedScalar *__restrict__ uu, CeedScalar *__restrict__ vv) { 270 // Restriction with tridiagonal transformation 271 CeedElemRestriction_Memcheck *impl; 272 CeedScalar vv_loc[block_size]; 273 274 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 275 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 276 for (CeedSize k = 0; k < num_comp; k++) { 277 // Iteration bound set to discard padding elements 278 const CeedSize block_end = CeedIntMin(block_size, num_elem - e); 279 CeedSize n = 0; 280 281 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 282 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 283 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size] + 284 uu[e * elem_size * num_comp + (k * elem_size + n + 1) * block_size + j - v_offset] * 285 impl->curl_orients[j + (3 * n + 3) * block_size + e * 3 * elem_size]; 286 } 287 for (CeedSize j = 0; j < block_end; j++) { 288 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 289 } 290 for (n = 1; n < elem_size - 1; n++) { 291 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 292 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n - 1) * block_size + j - v_offset] * 293 impl->curl_orients[j + (3 * n - 1) * block_size + e * 3 * elem_size] + 294 uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 295 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size] + 296 uu[e * elem_size * num_comp + (k * elem_size + n + 1) * block_size + j - v_offset] * 297 impl->curl_orients[j + (3 * n + 3) * block_size + e * 3 * elem_size]; 298 } 299 for (CeedSize j = 0; j < block_end; j++) { 300 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 301 } 302 } 303 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 304 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n - 1) * block_size + j - v_offset] * 305 impl->curl_orients[j + (3 * n - 1) * block_size + e * 3 * elem_size] + 306 uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 307 impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]; 308 } 309 for (CeedSize j = 0; j < block_end; j++) { 310 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 311 } 312 } 313 } 314 return CEED_ERROR_SUCCESS; 315 } 316 317 static inline int CeedElemRestrictionApplyCurlOrientedUnsignedTranspose_Memcheck_Core( 318 CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, const CeedInt comp_stride, CeedInt start, CeedInt stop, 319 CeedInt num_elem, CeedInt elem_size, CeedSize v_offset, const CeedScalar *__restrict__ uu, CeedScalar *__restrict__ vv) { 320 // Restriction with (unsigned) tridiagonal transformation 321 CeedElemRestriction_Memcheck *impl; 322 CeedScalar vv_loc[block_size]; 323 324 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 325 for (CeedSize e = start * block_size; e < stop * block_size; e += block_size) { 326 for (CeedSize k = 0; k < num_comp; k++) { 327 // Iteration bound set to discard padding elements 328 const CeedSize block_end = CeedIntMin(block_size, num_elem - e); 329 CeedSize n = 0; 330 331 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 332 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 333 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]) + 334 uu[e * elem_size * num_comp + (k * elem_size + n + 1) * block_size + j - v_offset] * 335 abs(impl->curl_orients[j + (3 * n + 3) * block_size + e * 3 * elem_size]); 336 } 337 for (CeedSize j = 0; j < block_end; j++) { 338 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 339 } 340 for (n = 1; n < elem_size - 1; n++) { 341 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 342 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n - 1) * block_size + j - v_offset] * 343 abs(impl->curl_orients[j + (3 * n - 1) * block_size + e * 3 * elem_size]) + 344 uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 345 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]) + 346 uu[e * elem_size * num_comp + (k * elem_size + n + 1) * block_size + j - v_offset] * 347 abs(impl->curl_orients[j + (3 * n + 3) * block_size + e * 3 * elem_size]); 348 } 349 for (CeedSize j = 0; j < block_end; j++) { 350 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 351 } 352 } 353 CeedPragmaSIMD for (CeedSize j = 0; j < block_end; j++) { 354 vv_loc[j] = uu[e * elem_size * num_comp + (k * elem_size + n - 1) * block_size + j - v_offset] * 355 abs(impl->curl_orients[j + (3 * n - 1) * block_size + e * 3 * elem_size]) + 356 uu[e * elem_size * num_comp + (k * elem_size + n) * block_size + j - v_offset] * 357 abs(impl->curl_orients[j + (3 * n + 1) * block_size + e * 3 * elem_size]); 358 } 359 for (CeedSize j = 0; j < block_end; j++) { 360 CeedPragmaAtomic vv[impl->offsets[j + n * block_size + e * elem_size] + k * comp_stride] += vv_loc[j]; 361 } 362 } 363 } 364 return CEED_ERROR_SUCCESS; 365 } 366 367 static inline int CeedElemRestrictionApplyAtPointsInElement_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, CeedInt start, 368 CeedInt stop, CeedTransposeMode t_mode, const CeedScalar *__restrict__ uu, 369 CeedScalar *__restrict__ vv) { 370 CeedInt num_points, l_vec_offset; 371 CeedSize e_vec_offset = 0; 372 CeedElemRestriction_Memcheck *impl; 373 374 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 375 for (CeedSize e = start; e < stop; e++) { 376 l_vec_offset = impl->offsets[e]; 377 CeedCallBackend(CeedElemRestrictionGetNumPointsInElement(rstr, e, &num_points)); 378 if (t_mode == CEED_NOTRANSPOSE) { 379 for (CeedSize i = 0; i < num_points; i++) { 380 for (CeedSize j = 0; j < num_comp; j++) vv[j * num_points + i + e_vec_offset] = uu[impl->offsets[i + l_vec_offset] * num_comp + j]; 381 } 382 } else { 383 for (CeedSize i = 0; i < num_points; i++) { 384 for (CeedSize j = 0; j < num_comp; j++) vv[impl->offsets[i + l_vec_offset] * num_comp + j] += uu[j * num_points + i + e_vec_offset]; 385 } 386 } 387 e_vec_offset += num_points * (CeedSize)num_comp; 388 } 389 return CEED_ERROR_SUCCESS; 390 } 391 392 static inline int CeedElemRestrictionApply_Memcheck_Core(CeedElemRestriction rstr, const CeedInt num_comp, const CeedInt block_size, 393 const CeedInt comp_stride, CeedInt start, CeedInt stop, CeedTransposeMode t_mode, 394 bool use_signs, bool use_orients, CeedVector u, CeedVector v, CeedRequest *request) { 395 CeedInt num_elem, elem_size; 396 CeedSize v_offset; 397 CeedRestrictionType rstr_type; 398 const CeedScalar *uu; 399 CeedScalar *vv; 400 401 CeedCallBackend(CeedElemRestrictionGetNumElements(rstr, &num_elem)); 402 CeedCallBackend(CeedElemRestrictionGetElementSize(rstr, &elem_size)); 403 v_offset = start * block_size * elem_size * (CeedSize)num_comp; 404 CeedCallBackend(CeedElemRestrictionGetType(rstr, &rstr_type)); 405 CeedCallBackend(CeedVectorGetArrayRead(u, CEED_MEM_HOST, &uu)); 406 407 if (t_mode == CEED_TRANSPOSE) { 408 // Sum into for transpose mode, E-vector to L-vector 409 CeedCallBackend(CeedVectorGetArray(v, CEED_MEM_HOST, &vv)); 410 } else { 411 // Overwrite for notranspose mode, L-vector to E-vector 412 CeedCallBackend(CeedVectorGetArrayWrite(v, CEED_MEM_HOST, &vv)); 413 } 414 415 if (t_mode == CEED_TRANSPOSE) { 416 // Restriction from E-vector to L-vector 417 // Performing v += r^T * u 418 // uu has shape [elem_size, num_comp, num_elem], row-major 419 // vv has shape [nnodes, num_comp] 420 // Sum into for transpose mode 421 switch (rstr_type) { 422 case CEED_RESTRICTION_STRIDED: 423 CeedCallBackend(CeedElemRestrictionApplyStridedTranspose_Memcheck_Core(rstr, num_comp, block_size, start, stop, num_elem, elem_size, v_offset, 424 uu, vv)); 425 break; 426 case CEED_RESTRICTION_STANDARD: 427 CeedCallBackend(CeedElemRestrictionApplyOffsetTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 428 elem_size, v_offset, uu, vv)); 429 break; 430 case CEED_RESTRICTION_ORIENTED: 431 if (use_signs) { 432 CeedCallBackend(CeedElemRestrictionApplyOrientedTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 433 elem_size, v_offset, uu, vv)); 434 } else { 435 CeedCallBackend(CeedElemRestrictionApplyOffsetTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 436 elem_size, v_offset, uu, vv)); 437 } 438 break; 439 case CEED_RESTRICTION_CURL_ORIENTED: 440 if (use_signs && use_orients) { 441 CeedCallBackend(CeedElemRestrictionApplyCurlOrientedTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 442 elem_size, v_offset, uu, vv)); 443 } else if (use_orients) { 444 CeedCallBackend(CeedElemRestrictionApplyCurlOrientedUnsignedTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, 445 num_elem, elem_size, v_offset, uu, vv)); 446 } else { 447 CeedCallBackend(CeedElemRestrictionApplyOffsetTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 448 elem_size, v_offset, uu, vv)); 449 } 450 break; 451 case CEED_RESTRICTION_POINTS: 452 CeedCallBackend(CeedElemRestrictionApplyAtPointsInElement_Memcheck_Core(rstr, num_comp, start, stop, t_mode, uu, vv)); 453 break; 454 } 455 } else { 456 // Restriction from L-vector to E-vector 457 // Perform: v = r * u 458 // vv has shape [elem_size, num_comp, num_elem], row-major 459 // uu has shape [nnodes, num_comp] 460 // Overwrite for notranspose mode 461 switch (rstr_type) { 462 case CEED_RESTRICTION_STRIDED: 463 CeedCallBackend(CeedElemRestrictionApplyStridedNoTranspose_Memcheck_Core(rstr, num_comp, block_size, start, stop, num_elem, elem_size, 464 v_offset, uu, vv)); 465 break; 466 case CEED_RESTRICTION_STANDARD: 467 CeedCallBackend(CeedElemRestrictionApplyOffsetNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 468 elem_size, v_offset, uu, vv)); 469 break; 470 case CEED_RESTRICTION_ORIENTED: 471 if (use_signs) { 472 CeedCallBackend(CeedElemRestrictionApplyOrientedNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 473 elem_size, v_offset, uu, vv)); 474 } else { 475 CeedCallBackend(CeedElemRestrictionApplyOffsetNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 476 elem_size, v_offset, uu, vv)); 477 } 478 break; 479 case CEED_RESTRICTION_CURL_ORIENTED: 480 if (use_signs && use_orients) { 481 CeedCallBackend(CeedElemRestrictionApplyCurlOrientedNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, 482 num_elem, elem_size, v_offset, uu, vv)); 483 } else if (use_orients) { 484 CeedCallBackend(CeedElemRestrictionApplyCurlOrientedUnsignedNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, 485 num_elem, elem_size, v_offset, uu, vv)); 486 } else { 487 CeedCallBackend(CeedElemRestrictionApplyOffsetNoTranspose_Memcheck_Core(rstr, num_comp, block_size, comp_stride, start, stop, num_elem, 488 elem_size, v_offset, uu, vv)); 489 } 490 break; 491 case CEED_RESTRICTION_POINTS: 492 CeedCallBackend(CeedElemRestrictionApplyAtPointsInElement_Memcheck_Core(rstr, num_comp, start, stop, t_mode, uu, vv)); 493 break; 494 } 495 } 496 CeedCallBackend(CeedVectorRestoreArrayRead(u, &uu)); 497 CeedCallBackend(CeedVectorRestoreArray(v, &vv)); 498 if (request != CEED_REQUEST_IMMEDIATE && request != CEED_REQUEST_ORDERED) *request = NULL; 499 return CEED_ERROR_SUCCESS; 500 } 501 502 //------------------------------------------------------------------------------ 503 // ElemRestriction Apply 504 //------------------------------------------------------------------------------ 505 static int CeedElemRestrictionApply_Memcheck(CeedElemRestriction rstr, CeedTransposeMode t_mode, CeedVector u, CeedVector v, CeedRequest *request) { 506 CeedInt num_block, block_size, num_comp, comp_stride; 507 CeedElemRestriction_Memcheck *impl; 508 509 CeedCallBackend(CeedElemRestrictionGetNumBlocks(rstr, &num_block)); 510 CeedCallBackend(CeedElemRestrictionGetBlockSize(rstr, &block_size)); 511 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 512 CeedCallBackend(CeedElemRestrictionGetCompStride(rstr, &comp_stride)); 513 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 514 CeedCallBackend(impl->Apply(rstr, num_comp, block_size, comp_stride, 0, num_block, t_mode, true, true, u, v, request)); 515 return CEED_ERROR_SUCCESS; 516 } 517 518 //------------------------------------------------------------------------------ 519 // ElemRestriction Apply Unsigned 520 //------------------------------------------------------------------------------ 521 static int CeedElemRestrictionApplyUnsigned_Memcheck(CeedElemRestriction rstr, CeedTransposeMode t_mode, CeedVector u, CeedVector v, 522 CeedRequest *request) { 523 CeedInt num_block, block_size, num_comp, comp_stride; 524 CeedElemRestriction_Memcheck *impl; 525 526 CeedCallBackend(CeedElemRestrictionGetNumBlocks(rstr, &num_block)); 527 CeedCallBackend(CeedElemRestrictionGetBlockSize(rstr, &block_size)); 528 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 529 CeedCallBackend(CeedElemRestrictionGetCompStride(rstr, &comp_stride)); 530 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 531 CeedCallBackend(impl->Apply(rstr, num_comp, block_size, comp_stride, 0, num_block, t_mode, false, true, u, v, request)); 532 return CEED_ERROR_SUCCESS; 533 } 534 535 //------------------------------------------------------------------------------ 536 // ElemRestriction Apply Unoriented 537 //------------------------------------------------------------------------------ 538 static int CeedElemRestrictionApplyUnoriented_Memcheck(CeedElemRestriction rstr, CeedTransposeMode t_mode, CeedVector u, CeedVector v, 539 CeedRequest *request) { 540 CeedInt num_block, block_size, num_comp, comp_stride; 541 CeedElemRestriction_Memcheck *impl; 542 543 CeedCallBackend(CeedElemRestrictionGetNumBlocks(rstr, &num_block)); 544 CeedCallBackend(CeedElemRestrictionGetBlockSize(rstr, &block_size)); 545 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 546 CeedCallBackend(CeedElemRestrictionGetCompStride(rstr, &comp_stride)); 547 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 548 CeedCallBackend(impl->Apply(rstr, num_comp, block_size, comp_stride, 0, num_block, t_mode, false, false, u, v, request)); 549 return CEED_ERROR_SUCCESS; 550 } 551 552 //------------------------------------------------------------------------------ 553 // ElemRestriction Apply Points 554 //------------------------------------------------------------------------------ 555 static int CeedElemRestrictionApplyAtPointsInElement_Memcheck(CeedElemRestriction rstr, CeedInt elem, CeedTransposeMode t_mode, CeedVector u, 556 CeedVector v, CeedRequest *request) { 557 CeedInt num_comp; 558 CeedElemRestriction_Memcheck *impl; 559 560 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 561 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 562 return impl->Apply(rstr, num_comp, 0, 1, elem, elem + 1, t_mode, false, false, u, v, request); 563 } 564 565 //------------------------------------------------------------------------------ 566 // ElemRestriction Apply Block 567 //------------------------------------------------------------------------------ 568 static int CeedElemRestrictionApplyBlock_Memcheck(CeedElemRestriction rstr, CeedInt block, CeedTransposeMode t_mode, CeedVector u, CeedVector v, 569 CeedRequest *request) { 570 CeedInt block_size, num_comp, comp_stride; 571 CeedElemRestriction_Memcheck *impl; 572 573 CeedCallBackend(CeedElemRestrictionGetBlockSize(rstr, &block_size)); 574 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 575 CeedCallBackend(CeedElemRestrictionGetCompStride(rstr, &comp_stride)); 576 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 577 CeedCallBackend(impl->Apply(rstr, num_comp, block_size, comp_stride, block, block + 1, t_mode, true, true, u, v, request)); 578 return CEED_ERROR_SUCCESS; 579 } 580 581 //------------------------------------------------------------------------------ 582 // ElemRestriction Get Offsets 583 //------------------------------------------------------------------------------ 584 static int CeedElemRestrictionGetOffsets_Memcheck(CeedElemRestriction rstr, CeedMemType mem_type, const CeedInt **offsets) { 585 CeedElemRestriction_Memcheck *impl; 586 587 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 588 589 CeedCheck(mem_type == CEED_MEM_HOST, CeedElemRestrictionReturnCeed(rstr), CEED_ERROR_BACKEND, "Can only provide to HOST memory"); 590 591 *offsets = impl->offsets; 592 return CEED_ERROR_SUCCESS; 593 } 594 595 //------------------------------------------------------------------------------ 596 // ElemRestriction Get Orientations 597 //------------------------------------------------------------------------------ 598 static int CeedElemRestrictionGetOrientations_Memcheck(CeedElemRestriction rstr, CeedMemType mem_type, const bool **orients) { 599 CeedElemRestriction_Memcheck *impl; 600 601 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 602 603 CeedCheck(mem_type == CEED_MEM_HOST, CeedElemRestrictionReturnCeed(rstr), CEED_ERROR_BACKEND, "Can only provide to HOST memory"); 604 605 *orients = impl->orients; 606 return CEED_ERROR_SUCCESS; 607 } 608 609 //------------------------------------------------------------------------------ 610 // ElemRestriction Get Curl-Conforming Orientations 611 //------------------------------------------------------------------------------ 612 static int CeedElemRestrictionGetCurlOrientations_Memcheck(CeedElemRestriction rstr, CeedMemType mem_type, const CeedInt8 **curl_orients) { 613 CeedElemRestriction_Memcheck *impl; 614 615 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 616 617 CeedCheck(mem_type == CEED_MEM_HOST, CeedElemRestrictionReturnCeed(rstr), CEED_ERROR_BACKEND, "Can only provide to HOST memory"); 618 619 *curl_orients = impl->curl_orients; 620 return CEED_ERROR_SUCCESS; 621 } 622 623 //------------------------------------------------------------------------------ 624 // ElemRestriction Destroy 625 //------------------------------------------------------------------------------ 626 static int CeedElemRestrictionDestroy_Memcheck(CeedElemRestriction rstr) { 627 CeedElemRestriction_Memcheck *impl; 628 629 CeedCallBackend(CeedElemRestrictionGetData(rstr, &impl)); 630 CeedCallBackend(CeedFree(&impl->offsets_allocated)); 631 CeedCallBackend(CeedFree(&impl->orients_allocated)); 632 CeedCallBackend(CeedFree(&impl->curl_orients_allocated)); 633 CeedCallBackend(CeedFree(&impl)); 634 return CEED_ERROR_SUCCESS; 635 } 636 637 //------------------------------------------------------------------------------ 638 // ElemRestriction Create 639 //------------------------------------------------------------------------------ 640 int CeedElemRestrictionCreate_Memcheck(CeedMemType mem_type, CeedCopyMode copy_mode, const CeedInt *offsets, const bool *orients, 641 const CeedInt8 *curl_orients, CeedElemRestriction rstr) { 642 Ceed ceed; 643 CeedInt num_elem, elem_size, num_block, block_size, num_comp, comp_stride, num_points = 0, num_offsets; 644 CeedRestrictionType rstr_type; 645 CeedElemRestriction_Memcheck *impl; 646 647 CeedCallBackend(CeedElemRestrictionGetCeed(rstr, &ceed)); 648 CeedCallBackend(CeedElemRestrictionGetNumElements(rstr, &num_elem)); 649 CeedCallBackend(CeedElemRestrictionGetElementSize(rstr, &elem_size)); 650 CeedCallBackend(CeedElemRestrictionGetNumBlocks(rstr, &num_block)); 651 CeedCallBackend(CeedElemRestrictionGetBlockSize(rstr, &block_size)); 652 CeedCallBackend(CeedElemRestrictionGetNumComponents(rstr, &num_comp)); 653 CeedCallBackend(CeedElemRestrictionGetCompStride(rstr, &comp_stride)); 654 CeedCallBackend(CeedElemRestrictionGetType(rstr, &rstr_type)); 655 656 CeedCheck(mem_type == CEED_MEM_HOST, ceed, CEED_ERROR_BACKEND, "Only MemType = HOST supported"); 657 658 CeedCallBackend(CeedCalloc(1, &impl)); 659 CeedCallBackend(CeedElemRestrictionSetData(rstr, impl)); 660 661 // Set layouts 662 { 663 bool has_backend_strides; 664 CeedInt e_layout[3] = {1, elem_size, elem_size * num_comp}, l_layout[3] = {0}; 665 666 CeedCallBackend(CeedElemRestrictionSetELayout(rstr, e_layout)); 667 if (rstr_type == CEED_RESTRICTION_STRIDED) { 668 CeedCallBackend(CeedElemRestrictionHasBackendStrides(rstr, &has_backend_strides)); 669 if (has_backend_strides) { 670 CeedCallBackend(CeedElemRestrictionGetBackendStrides_Memcheck(rstr, l_layout)); 671 CeedCallBackend(CeedElemRestrictionSetLLayout(rstr, l_layout)); 672 } 673 } 674 } 675 676 // Expand E-vector size for AtPoints 677 if (rstr_type == CEED_RESTRICTION_POINTS) { 678 CeedSize max_points = 0, num_points_total = 0; 679 680 for (CeedInt i = 0; i < num_elem; i++) { 681 CeedInt num_points = offsets[i + 1] - offsets[i]; 682 683 max_points = CeedIntMax(max_points, num_points); 684 num_points_total += num_points; 685 } 686 // -- Increase size for last element 687 num_points_total += (max_points - (offsets[num_elem] - offsets[num_elem - 1])); 688 CeedCallBackend(CeedElemRestrictionSetAtPointsEVectorSize(rstr, num_points_total * num_comp)); 689 } 690 691 // Offsets data 692 if (rstr_type != CEED_RESTRICTION_STRIDED) { 693 // Check indices 694 { 695 CeedSize l_size; 696 697 CeedCallBackend(CeedElemRestrictionGetLVectorSize(rstr, &l_size)); 698 for (CeedInt i = 0; i < num_elem * elem_size; i++) { 699 CeedCheck(offsets[i] >= 0 && offsets[i] + (num_comp - 1) * comp_stride < l_size, ceed, CEED_ERROR_BACKEND, 700 "Restriction offset %" CeedInt_FMT " (%" CeedInt_FMT ") out of range [0, %" CeedInt_FMT "]", i, offsets[i], l_size); 701 } 702 } 703 704 // Copy data 705 if (rstr_type == CEED_RESTRICTION_POINTS) CeedCallBackend(CeedElemRestrictionGetNumPoints(rstr, &num_points)); 706 num_offsets = rstr_type == CEED_RESTRICTION_POINTS ? (num_elem + 1 + num_points) : (num_elem * elem_size); 707 switch (copy_mode) { 708 case CEED_COPY_VALUES: 709 CeedCallBackend(CeedMalloc(num_offsets, &impl->offsets_allocated)); 710 memcpy(impl->offsets_allocated, offsets, num_offsets * sizeof(offsets[0])); 711 impl->offsets = impl->offsets_allocated; 712 break; 713 case CEED_OWN_POINTER: 714 impl->offsets_allocated = (CeedInt *)offsets; 715 impl->offsets = impl->offsets_allocated; 716 break; 717 case CEED_USE_POINTER: 718 impl->offsets = offsets; 719 } 720 721 // Orientation data 722 if (rstr_type == CEED_RESTRICTION_ORIENTED) { 723 CeedCheck(orients != NULL, ceed, CEED_ERROR_BACKEND, "No orients array provided for oriented restriction"); 724 switch (copy_mode) { 725 case CEED_COPY_VALUES: 726 CeedCallBackend(CeedMalloc(num_offsets, &impl->orients_allocated)); 727 memcpy(impl->orients_allocated, orients, num_offsets * sizeof(orients[0])); 728 impl->orients = impl->orients_allocated; 729 break; 730 case CEED_OWN_POINTER: 731 impl->orients_allocated = (bool *)orients; 732 impl->orients = impl->orients_allocated; 733 break; 734 case CEED_USE_POINTER: 735 impl->orients = orients; 736 } 737 } else if (rstr_type == CEED_RESTRICTION_CURL_ORIENTED) { 738 CeedCheck(curl_orients != NULL, ceed, CEED_ERROR_BACKEND, "No curl_orients array provided for oriented restriction"); 739 switch (copy_mode) { 740 case CEED_COPY_VALUES: 741 CeedCallBackend(CeedMalloc(3 * num_offsets, &impl->curl_orients_allocated)); 742 memcpy(impl->curl_orients_allocated, curl_orients, 3 * num_offsets * sizeof(curl_orients[0])); 743 impl->curl_orients = impl->curl_orients_allocated; 744 break; 745 case CEED_OWN_POINTER: 746 impl->curl_orients_allocated = (CeedInt8 *)curl_orients; 747 impl->curl_orients = impl->curl_orients_allocated; 748 break; 749 case CEED_USE_POINTER: 750 impl->curl_orients = curl_orients; 751 } 752 } 753 } 754 755 // Set apply function 756 impl->Apply = CeedElemRestrictionApply_Memcheck_Core; 757 758 // Register backend functions 759 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "Apply", CeedElemRestrictionApply_Memcheck)); 760 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "ApplyUnsigned", CeedElemRestrictionApplyUnsigned_Memcheck)); 761 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "ApplyUnoriented", CeedElemRestrictionApplyUnoriented_Memcheck)); 762 if (rstr_type == CEED_RESTRICTION_POINTS) { 763 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "ApplyAtPointsInElement", 764 CeedElemRestrictionApplyAtPointsInElement_Memcheck)); 765 } 766 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "ApplyBlock", CeedElemRestrictionApplyBlock_Memcheck)); 767 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "GetOffsets", CeedElemRestrictionGetOffsets_Memcheck)); 768 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "GetOrientations", CeedElemRestrictionGetOrientations_Memcheck)); 769 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "GetCurlOrientations", CeedElemRestrictionGetCurlOrientations_Memcheck)); 770 CeedCallBackend(CeedSetBackendFunction(ceed, "ElemRestriction", rstr, "Destroy", CeedElemRestrictionDestroy_Memcheck)); 771 CeedCallBackend(CeedDestroy(&ceed)); 772 return CEED_ERROR_SUCCESS; 773 } 774 775 //------------------------------------------------------------------------------ 776