90       CeedInt num_t_col   = 1;  in CeedBasisApplyCore_Magma()  local
97           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_1D);  in CeedBasisApplyCore_Magma()
98           shared_mem += sizeof(CeedScalar) * num_t_col * (num_comp * (1 * P + 1 * Q));  in CeedBasisApplyCore_Magma()
103           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_2D);  in CeedBasisApplyCore_Magma()
106           shared_mem += num_t_col * (P * max_P_Q * sizeof(CeedScalar));  in CeedBasisApplyCore_Magma()
110           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_3D);  in CeedBasisApplyCore_Magma()
113           shared_mem += sizeof(CeedScalar) * num_t_col * (CeedIntMax(P * P * max_P_Q, P * Q * Q));  in CeedBasisApplyCore_Magma()
116       CeedInt grid   = CeedDivUpInt(num_elem, num_t_col);  in CeedBasisApplyCore_Magma()
121                                                     num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
123 …eedRunKernelDimSharedMagma(ceed, impl->Interp, NULL, grid, num_threads, num_t_col, 1, shared_mem, …  in CeedBasisApplyCore_Magma()
167       CeedInt num_t_col   = 1;  in CeedBasisApplyCore_Magma()  local
174           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_1D);  in CeedBasisApplyCore_Magma()
175           shared_mem += sizeof(CeedScalar) * num_t_col * (num_comp * (1 * P + 1 * Q));  in CeedBasisApplyCore_Magma()
180           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_2D);  in CeedBasisApplyCore_Magma()
183           shared_mem += sizeof(CeedScalar) * num_t_col * (P * max_P_Q);  in CeedBasisApplyCore_Magma()
187           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_3D);  in CeedBasisApplyCore_Magma()
190 …   shared_mem += sizeof(CeedScalar) * num_t_col * CeedIntMax(P * P * P, (P * P * Q) + (P * Q * Q));  in CeedBasisApplyCore_Magma()
193       CeedInt grid   = CeedDivUpInt(num_elem, num_t_col);  in CeedBasisApplyCore_Magma()
199                                                     num_t_col, 1, shared_mem, args));  in CeedBasisApplyCore_Magma()
201 …(CeedRunKernelDimSharedMagma(ceed, impl->Grad, NULL, grid, num_threads, num_t_col, 1, shared_mem, …  in CeedBasisApplyCore_Magma()
209       CeedInt num_t_col      = 1;  in CeedBasisApplyCore_Magma()  local
215           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_1D);  in CeedBasisApplyCore_Magma()
217           shared_mem += sizeof(CeedScalar) * num_t_col * Q;  // for output  in CeedBasisApplyCore_Magma()
221           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_2D);  in CeedBasisApplyCore_Magma()
226           num_t_col   = MAGMA_BASIS_NTCOL(num_threads, MAGMA_MAXTHREADS_3D);  in CeedBasisApplyCore_Magma()
230       CeedInt grid   = CeedDivUpInt(num_elem, num_t_col);  in CeedBasisApplyCore_Magma()
233 …eedRunKernelDimSharedMagma(ceed, impl->Weight, NULL, grid, num_threads, num_t_col, 1, shared_mem, …  in CeedBasisApplyCore_Magma()
425       CeedInt num_t_col    = MAGMA_BASIS_NTCOL(M, MAGMA_MAXTHREADS_1D);  in CeedBasisApplyNonTensorCore_Magma()  local
426       CeedInt grid         = CeedDivUpInt(N, num_t_col * NB);  in CeedBasisApplyNonTensorCore_Magma()
428       CeedInt shared_mem_B = num_t_col * K * NB * sizeof(CeedScalar);  in CeedBasisApplyNonTensorCore_Magma()
432 …CeedCallBackend(CeedRunKernelDimSharedMagma(ceed, Kernel, NULL, grid, M, num_t_col, 1, shared_mem,…  in CeedBasisApplyNonTensorCore_Magma()
446     CeedInt num_t_col  = MAGMA_BASIS_NTCOL(Q, MAGMA_MAXTHREADS_1D);  in CeedBasisApplyNonTensorCore_Magma()  local
447     CeedInt grid       = CeedDivUpInt(num_elem, num_t_col);  in CeedBasisApplyNonTensorCore_Magma()
448     CeedInt shared_mem = Q * sizeof(CeedScalar) + num_t_col * Q * sizeof(CeedScalar);  in CeedBasisApplyNonTensorCore_Magma()
451 …CeedCallBackend(CeedRunKernelDimSharedMagma(ceed, impl->Weight, NULL, grid, Q, num_t_col, 1, share…  in CeedBasisApplyNonTensorCore_Magma()