91       CeedInt shared_mem  = 0;  in CeedBasisApplyCore_Magma()  local
98           shared_mem += sizeof(CeedScalar) * num_t_col * (num_comp * (1 * P + 1 * Q));  in CeedBasisApplyCore_Magma()
99           shared_mem += sizeof(CeedScalar) * (P * Q);  in CeedBasisApplyCore_Magma()
104           shared_mem += P * Q * sizeof(CeedScalar);  // for sT  in CeedBasisApplyCore_Magma()
106           shared_mem += num_t_col * (P * max_P_Q * sizeof(CeedScalar));  in CeedBasisApplyCore_Magma()
111           shared_mem += sizeof(CeedScalar) * (P * Q);  // for sT  in CeedBasisApplyCore_Magma()
113           shared_mem += sizeof(CeedScalar) * num_t_col * (CeedIntMax(P * P * max_P_Q, P * Q * Q));  in CeedBasisApplyCore_Magma()
121                                                     num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
123 …KernelDimSharedMagma(ceed, impl->Interp, NULL, grid, num_threads, num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
168       CeedInt shared_mem  = 0;  in CeedBasisApplyCore_Magma()  local
175           shared_mem += sizeof(CeedScalar) * num_t_col * (num_comp * (1 * P + 1 * Q));  in CeedBasisApplyCore_Magma()
176           shared_mem += sizeof(CeedScalar) * (P * Q);  in CeedBasisApplyCore_Magma()
181           shared_mem += sizeof(CeedScalar) * 2 * P * Q;  // for sTinterp and sTgrad  in CeedBasisApplyCore_Magma()
183           shared_mem += sizeof(CeedScalar) * num_t_col * (P * max_P_Q);  in CeedBasisApplyCore_Magma()
188           shared_mem += sizeof(CeedScalar) * 2 * P * Q;  // for sTinterp and sTgrad  in CeedBasisApplyCore_Magma()
190 …   shared_mem += sizeof(CeedScalar) * num_t_col * CeedIntMax(P * P * P, (P * P * Q) + (P * Q * Q));  in CeedBasisApplyCore_Magma()
199                                                     num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
201 …unKernelDimSharedMagma(ceed, impl->Grad, NULL, grid, num_threads, num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
210       CeedInt shared_mem     = 0;  in CeedBasisApplyCore_Magma()  local
216           shared_mem += sizeof(CeedScalar) * Q;              // for d_q_weight_1d  in CeedBasisApplyCore_Magma()
217           shared_mem += sizeof(CeedScalar) * num_t_col * Q;  // for output  in CeedBasisApplyCore_Magma()
222           shared_mem += sizeof(CeedScalar) * Q;  // for d_q_weight_1d  in CeedBasisApplyCore_Magma()
227           shared_mem += sizeof(CeedScalar) * Q;  // for d_q_weight_1d  in CeedBasisApplyCore_Magma()
233 …KernelDimSharedMagma(ceed, impl->Weight, NULL, grid, num_threads, num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
429 …CeedInt shared_mem   = (t_mode != CEED_TRANSPOSE && q_comp > 1) ? (shared_mem_A + shared_mem_B) : …  in CeedBasisApplyNonTensorCore_Magma()  local
432 …lBackend(CeedRunKernelDimSharedMagma(ceed, Kernel, NULL, grid, M, num_t_col, 1, shared_mem, args));  in CeedBasisApplyNonTensorCore_Magma()
448     CeedInt shared_mem = Q * sizeof(CeedScalar) + num_t_col * Q * sizeof(CeedScalar);  in CeedBasisApplyNonTensorCore_Magma()  local
451 …nd(CeedRunKernelDimSharedMagma(ceed, impl->Weight, NULL, grid, Q, num_t_col, 1, shared_mem, args));  in CeedBasisApplyNonTensorCore_Magma()