41     for (CeedInt comp = 0; comp < num_comp; ++comp) {  in readDofsOffset1d()  local
42       r_u[comp] = d_u[ind + strides_comp * comp];  in readDofsOffset1d()
59     for (CeedInt comp = 0; comp < num_comp; comp++) {  in readDofsStrided1d()  local
60       r_u[comp] = d_u[ind + comp * strides_comp];  in readDofsStrided1d()
76     for (CeedInt comp = 0; comp < num_comp; ++comp)  in writeDofsOffset1d()  local
77 …atomic_fetch_add_explicit(&d_v[ind + strides_comp * comp], r_v[comp], memory_order_relaxed, memory…  in writeDofsOffset1d()
93     for (CeedInt comp = 0; comp < num_comp; comp++) {  in writeDofsStrided1d()  local
94       d_v[ind + comp * strides_comp] = r_v[comp];  in writeDofsStrided1d()
115     for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[comp] = d_u[ind + strides_comp * comp];  in readDofsOffset2d()  local
132     for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[comp] = d_u[ind + comp * strides_comp];  in readDofsStrided2d()  local
148     for (CeedInt comp = 0; comp < num_comp; ++comp)  in writeDofsOffset2d()  local
149 …atomic_fetch_add_explicit(&d_v[ind + strides_comp * comp], r_v[comp], memory_order_relaxed, memory…  in writeDofsOffset2d()
166     for (CeedInt comp = 0; comp < num_comp; ++comp) d_v[ind + comp * strides_comp] += r_v[comp];  in writeDofsStrided2d()  local
187 …for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[z + comp * P_1D] = d_u[ind + strides_comp * co…  in readDofsOffset3d()  local
206 …for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[z + comp * P_1D] = d_u[ind + comp * strides_co…  in readDofsStrided3d()  local
223     for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[comp] = d_u[ind + strides_comp * comp];  in readSliceQuadsOffset3d()  local
240     for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[comp] = d_u[ind + comp * strides_comp];  in readSliceQuadsStrided3d()  local
257       for (CeedInt comp = 0; comp < num_comp; ++comp)  in writeDofsOffset3d()  local
258 …atomic_fetch_add_explicit(&d_v[ind + strides_comp * comp], r_v[z + comp * P_1D], memory_order_rela…  in writeDofsOffset3d()
277 …for (CeedInt comp = 0; comp < num_comp; ++comp) d_v[ind + comp * strides_comp] += r_v[z + comp * P…  in writeDofsStrided3d()  local
290   for (CeedInt comp = 0; comp < num_comp; ++comp) {  in gradCollo3d()  local
292       scratch[item_id_x + item_id_y * T_1D] = r_U[q + comp * Q_1D];  in gradCollo3d()
298       r_V[comp + 0 * num_comp] = 0.0;  in gradCollo3d()
300 …r_V[comp + 0 * num_comp] += s_G[i + item_id_x * Q_1D] * scratch[i + item_id_y * T_1D];  // Contrac…  in gradCollo3d()
303       r_V[comp + 1 * num_comp] = 0.0;  in gradCollo3d()
305 …r_V[comp + 1 * num_comp] += s_G[i + item_id_y * Q_1D] * scratch[item_id_x + i * T_1D];  // Contrac…  in gradCollo3d()
308       r_V[comp + 2 * num_comp] = 0.0;  in gradCollo3d()
309 …for (CeedInt i = 0; i < Q_1D; ++i) r_V[comp + 2 * num_comp] += s_G[i + q * Q_1D] * r_U[i + comp * …  in gradCollo3d()
324   for (CeedInt comp = 0; comp < num_comp; ++comp) {  in gradColloTranspose3d()  local
327       scratch[item_id_x + item_id_y * T_1D] = r_U[comp + 0 * num_comp];  in gradColloTranspose3d()
333 …r_V[q + comp * Q_1D] += s_G[item_id_x + i * Q_1D] * scratch[i + item_id_y * T_1D];  // Contract x …  in gradColloTranspose3d()
339       scratch[item_id_x + item_id_y * T_1D] = r_U[comp + 1 * num_comp];  in gradColloTranspose3d()
345 …r_V[q + comp * Q_1D] += s_G[item_id_y + i * Q_1D] * scratch[item_id_x + i * T_1D];  // Contract y …  in gradColloTranspose3d()
352 …r_V[i + comp * Q_1D] += s_G[i + q * Q_1D] * r_U[comp + 2 * num_comp];  // PARTIAL contract z direc…  in gradColloTranspose3d()