| /libCEED/backends/sycl-gen/ |
| H A D | ceed-sycl-gen-operator-build.sycl.cpp | 30 extern "C" int BlockGridCalculate_Sycl_gen(const CeedInt dim, const CeedInt P_1d, const CeedInt Q_1… in BlockGridCalculate_Sycl_gen() argument
31 const CeedInt thread1d = CeedIntMax(Q_1d, P_1d); in BlockGridCalculate_Sycl_gen()
66 …CeedInt Q, P_1d = 0, Q_1d = 0, elem_size, num_input_fields, num_output_fields, n… in CeedOperatorBuildKernel_Sycl_gen() local
150 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedOperatorBuildKernel_Sycl_gen()
151 if (P_1d > impl->max_P_1d) impl->max_P_1d = P_1d; in CeedOperatorBuildKernel_Sycl_gen()
216 CeedCallBackend(BlockGridCalculate_Sycl_gen(dim, P_1d, Q_1d, block_sizes)); in CeedOperatorBuildKernel_Sycl_gen()
285 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedOperatorBuildKernel_Sycl_gen()
286 code << " const CeedInt P_in_" << i << " = " << P_1d << ";\n"; in CeedOperatorBuildKernel_Sycl_gen()
301 code << " local CeedScalar s_B_in_" << i << "[" << P_1d * Q_1d << "];\n"; in CeedOperatorBuildKernel_Sycl_gen()
307 code << " local CeedScalar s_B_in_" << i << "[" << P_1d * Q_1d << "];\n"; in CeedOperatorBuildKernel_Sycl_gen()
[all …]
|
| H A D | ceed-sycl-gen-operator-build.hpp | 9 CEED_INTERN int BlockGridCalculate_Sycl_gen(const CeedInt dim, const CeedInt P_1d, const CeedInt Q_…
|
| H A D | ceed-sycl-gen-operator.sycl.cpp | 125 const CeedInt P_1d = impl->max_P_1d; in CeedOperatorApplyAdd_Sycl_gen() local
128 CeedCallBackend(BlockGridCalculate_Sycl_gen(dim, P_1d, Q_1d, block_sizes)); in CeedOperatorApplyAdd_Sycl_gen()
|
| /libCEED/backends/hip-shared/ |
| H A D | ceed-hip-shared-basis.c | 39 static int ComputeBasisThreadBlockSizes(const CeedInt dim, const CeedInt P_1d, const CeedInt Q_1d, … in ComputeBasisThreadBlockSizes() argument
44 const CeedInt thread_1d = CeedIntMax(P_1d, Q_1d); in ComputeBasisThreadBlockSizes()
118 CeedInt P_1d, Q_1d; in CeedBasisApplyTensorCore_Hip_shared() local
122 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyTensorCore_Hip_shared()
124 CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in CeedBasisApplyTensorCore_Hip_shared()
165 CeedInt P_1d, Q_1d; in CeedBasisApplyTensorCore_Hip_shared() local
169 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyTensorCore_Hip_shared()
171 CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in CeedBasisApplyTensorCore_Hip_shared()
333 CeedInt P_1d; in CeedBasisApplyAtPointsCore_Hip_shared() local
335 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPointsCore_Hip_shared()
[all …]
|
| H A D | ceed-hip-shared.h | 41 CEED_INTERN int CeedBasisCreateTensorH1_Hip_shared(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const C…
|
| /libCEED/backends/cuda-shared/ |
| H A D | ceed-cuda-shared-basis.c | 51 CeedInt P_1d, Q_1d; in CeedBasisApplyTensorCore_Cuda_shared() local
54 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyTensorCore_Cuda_shared()
56 CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in CeedBasisApplyTensorCore_Cuda_shared()
101 CeedInt P_1d, Q_1d; in CeedBasisApplyTensorCore_Cuda_shared() local
104 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyTensorCore_Cuda_shared()
106 CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in CeedBasisApplyTensorCore_Cuda_shared()
268 CeedInt P_1d; in CeedBasisApplyAtPointsCore_Cuda_shared() local
270 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPointsCore_Cuda_shared()
278 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedBasisApplyAtPointsCore_Cuda_shared()
279 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedBasisApplyAtPointsCore_Cuda_shared()
[all …]
|
| H A D | ceed-cuda-shared.h | 40 CEED_INTERN int CeedBasisCreateTensorH1_Cuda_shared(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const …
|
| /libCEED/interface/ |
| H A D | ceed-basis.c | 400 CeedInt dim, num_comp, num_q_comp, num_nodes, P_1d = 1, Q_1d = 1, total_num_points = 0; in CeedBasisApplyAtPointsCheckDims() local
404 CeedCall(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPointsCheckDims()
478 CeedInt dim, num_comp, P_1d = 1, Q_1d = 1, total_num_points = num_points[0]; in CeedBasisApplyAtPoints_Core() local
483 CeedCall(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPoints_Core()
507 CeedCall(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedBasisApplyAtPoints_Core()
508 CeedCall(CeedCalloc(P_1d * Q_1d, &chebyshev_grad_1d)); in CeedBasisApplyAtPoints_Core()
515 …CeedCall(CeedBasisCreateTensorH1(ceed, dim, num_comp, P_1d, Q_1d, chebyshev_interp_1d, chebyshev_g… in CeedBasisApplyAtPoints_Core()
532 …CeedCall(CeedBasisCreateTensorH1Lagrange(ceed_ref, dim, num_comp, P_1d, Q_1d, CEED_GAUSS, &basis_r… in CeedBasisApplyAtPoints_Core()
735 CeedInt P_1d, Q_1d; in CeedBasisGetCollocatedGrad() local
741 CeedCall(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisGetCollocatedGrad()
[all …]
|
| H A D | ceed-preconditioning.c | 1285 …r *interp_1d, const CeedScalar *grad_1d, const CeedScalar *q_weight_1d, CeedInt P_1d, CeedInt Q_1d, in CeedBuildMassLaplace() argument
1287 for (CeedInt i = 0; i < P_1d; i++) { in CeedBuildMassLaplace()
1288 for (CeedInt j = 0; j < P_1d; j++) { in CeedBuildMassLaplace()
1290 … (CeedInt k = 0; k < Q_1d; k++) sum += interp_1d[k * P_1d + i] * q_weight_1d[k] * interp_1d[k * P_… in CeedBuildMassLaplace()
1291 mass[i + j * P_1d] = sum; in CeedBuildMassLaplace()
1295 for (CeedInt i = 0; i < P_1d; i++) { in CeedBuildMassLaplace()
1296 for (CeedInt j = 0; j < P_1d; j++) { in CeedBuildMassLaplace()
1299 …for (CeedInt k = 0; k < Q_1d; k++) sum += grad_1d[k * P_1d + i] * q_weight_1d[k] * grad_1d[k * P_1… in CeedBuildMassLaplace()
1300 laplace[i + j * P_1d] = sum; in CeedBuildMassLaplace()
1304 for (CeedInt i = 0; i < P_1d; i++) laplace[i + P_1d * i] += perturbation; in CeedBuildMassLaplace()
[all …]
|
| /libCEED/backends/sycl-ref/ |
| H A D | ceed-sycl-ref-basis.sycl.cpp | 67 const CeedInt P_1d = kh.get_specialization_constant<BASIS_P_1D_ID>(); in CeedBasisApplyInterp_Sycl() local
70 const CeedInt num_nodes = CeedIntPow(P_1d, dim); in CeedBasisApplyInterp_Sycl()
72 const CeedInt P = is_transpose ? Q_1d : P_1d; in CeedBasisApplyInterp_Sycl()
73 const CeedInt Q = is_transpose ? P_1d : Q_1d; in CeedBasisApplyInterp_Sycl()
74 const CeedInt stride_0 = is_transpose ? 1 : P_1d; in CeedBasisApplyInterp_Sycl()
75 const CeedInt stride_1 = is_transpose ? P_1d : 1; in CeedBasisApplyInterp_Sycl()
170 const CeedInt P_1d = kh.get_specialization_constant<BASIS_P_1D_ID>(); in CeedBasisApplyGrad_Sycl() local
173 const CeedInt num_nodes = CeedIntPow(P_1d, dim); in CeedBasisApplyGrad_Sycl()
175 const CeedInt P = is_transpose ? Q_1d : P_1d; in CeedBasisApplyGrad_Sycl()
176 const CeedInt Q = is_transpose ? P_1d : Q_1d; in CeedBasisApplyGrad_Sycl()
[all …]
|
| H A D | ceed-sycl-ref.hpp | 48 CeedInt P_1d; member
117 CEED_INTERN int CeedBasisCreateTensorH1_Sycl(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedSca…
|
| /libCEED/backends/ref/ |
| H A D | ceed-ref-basis.c | 51 CeedInt P_1d, Q_1d; in CeedBasisApplyCore_Ref() local
53 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyCore_Ref()
61 CeedInt P = P_1d, Q = Q_1d; in CeedBasisApplyCore_Ref()
65 Q = P_1d; in CeedBasisApplyCore_Ref()
86 CeedInt P = P_1d, Q = Q_1d; in CeedBasisApplyCore_Ref()
115 Q = P_1d; in CeedBasisApplyCore_Ref()
149 Q = P_1d; in CeedBasisApplyCore_Ref()
278 int CeedBasisCreateTensorH1_Ref(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScalar *interp_1… in CeedBasisCreateTensorH1_Ref() argument
290 if (Q_1d >= P_1d && !impl->is_collocated) { in CeedBasisCreateTensorH1_Ref()
|
| H A D | ceed-ref.h | 70 CEED_INTERN int CeedBasisCreateTensorH1_Ref(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScal…
|
| /libCEED/backends/hip-ref/ |
| H A D | ceed-hip-ref-basis.c | 159 CeedInt P_1d; in CeedBasisApplyAtPointsCore_Hip() local
161 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPointsCore_Hip()
169 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedBasisApplyAtPointsCore_Hip()
170 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedBasisApplyAtPointsCore_Hip()
183 …s_kernel_source, &data->moduleAtPoints, 9, "BASIS_Q_1D", Q_1d, "BASIS_P_1D", P_1d, "BASIS_BUF_LEN", in CeedBasisApplyAtPointsCore_Hip()
184 …Q_1d * CeedIntPow(Q_1d > P_1d ? Q_1d : P_1d, dim - 1), "BASIS_DIM", dim, "BASIS_NUM_COMP", num_com… in CeedBasisApplyAtPointsCore_Hip()
185 …"BASIS_NUM_NODES", CeedIntPow(P_1d, dim), "BASIS_NUM_QPTS", CeedIntPow(Q_1d, dim), "BASIS_NUM_PTS", in CeedBasisApplyAtPointsCore_Hip()
398 int CeedBasisCreateTensorH1_Hip(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScalar *interp_1… in CeedBasisCreateTensorH1_Hip() argument
403 const CeedInt interp_bytes = q_bytes * P_1d; in CeedBasisCreateTensorH1_Hip()
423 …ed, basis_kernel_source, &data->module, 7, "BASIS_Q_1D", Q_1d, "BASIS_P_1D", P_1d, "BASIS_BUF_LEN", in CeedBasisCreateTensorH1_Hip()
[all …]
|
| H A D | ceed-hip-ref.h | 162 CEED_INTERN int CeedBasisCreateTensorH1_Hip(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScal…
|
| /libCEED/backends/cuda-ref/ |
| H A D | ceed-cuda-ref-basis.c | 160 CeedInt P_1d; in CeedBasisApplyAtPointsCore_Cuda() local
162 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyAtPointsCore_Cuda()
170 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedBasisApplyAtPointsCore_Cuda()
171 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedBasisApplyAtPointsCore_Cuda()
184 …s_kernel_source, &data->moduleAtPoints, 9, "BASIS_Q_1D", Q_1d, "BASIS_P_1D", P_1d, "BASIS_BUF_LEN", in CeedBasisApplyAtPointsCore_Cuda()
185 …Q_1d * CeedIntPow(Q_1d > P_1d ? Q_1d : P_1d, dim - 1), "BASIS_DIM", dim, "BASIS_NUM_COMP", num_com… in CeedBasisApplyAtPointsCore_Cuda()
186 …"BASIS_NUM_NODES", CeedIntPow(P_1d, dim), "BASIS_NUM_QPTS", CeedIntPow(Q_1d, dim), "BASIS_NUM_PTS", in CeedBasisApplyAtPointsCore_Cuda()
399 int CeedBasisCreateTensorH1_Cuda(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScalar *interp_… in CeedBasisCreateTensorH1_Cuda() argument
404 const CeedInt interp_bytes = q_bytes * P_1d; in CeedBasisCreateTensorH1_Cuda()
424 …ed, basis_kernel_source, &data->module, 7, "BASIS_Q_1D", Q_1d, "BASIS_P_1D", P_1d, "BASIS_BUF_LEN", in CeedBasisCreateTensorH1_Cuda()
[all …]
|
| H A D | ceed-cuda-ref.h | 157 CEED_INTERN int CeedBasisCreateTensorH1_Cuda(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedSca…
|
| /libCEED/backends/sycl-shared/ |
| H A D | ceed-sycl-shared-basis.sycl.cpp | 168 int CeedBasisCreateTensorH1_Sycl_shared(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScalar *… in CeedBasisCreateTensorH1_Sycl_shared() argument
182 const CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in CeedBasisCreateTensorH1_Sycl_shared()
183 const CeedInt num_nodes = CeedIntPow(P_1d, dim); in CeedBasisCreateTensorH1_Sycl_shared()
210 const CeedInt interp_length = Q_1d * P_1d; in CeedBasisCreateTensorH1_Sycl_shared()
223 const bool has_collocated_grad = (dim == 3) && (Q_1d >= P_1d); in CeedBasisCreateTensorH1_Sycl_shared()
241 jit_constants["BASIS_P_1D"] = P_1d; in CeedBasisCreateTensorH1_Sycl_shared()
|
| H A D | ceed-sycl-shared.hpp | 33 CEED_INTERN int CeedBasisCreateTensorH1_Sycl_shared(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const …
|
| /libCEED/backends/magma/ |
| H A D | ceed-magma-basis.c | 33 CeedInt dim, num_comp, num_nodes, P_1d, Q_1d, P, Q; in CeedBasisApplyCore_Magma() local
44 CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedBasisApplyCore_Magma()
46 P = P_1d; in CeedBasisApplyCore_Magma()
50 Q = P_1d; in CeedBasisApplyCore_Magma()
64 CeedInt elem_dofs_size = CeedIntPow(P_1d, dim); in CeedBasisApplyCore_Magma()
129 CeedInt elem_dofs_size = CeedIntPow(P_1d, dim); in CeedBasisApplyCore_Magma()
531 int CeedBasisCreateTensorH1_Magma(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedScalar *interp… in CeedBasisCreateTensorH1_Magma() argument
551 CeedCallBackend(magma_malloc((void **)&impl->d_interp_1d, Q_1d * P_1d * sizeof(interp_1d[0]))); in CeedBasisCreateTensorH1_Magma()
552 …magma_setvector(Q_1d * P_1d, sizeof(interp_1d[0]), interp_1d, 1, impl->d_interp_1d, 1, data->queue… in CeedBasisCreateTensorH1_Magma()
553 CeedCallBackend(magma_malloc((void **)&impl->d_grad_1d, Q_1d * P_1d * sizeof(grad_1d[0]))); in CeedBasisCreateTensorH1_Magma()
[all …]
|
| H A D | ceed-magma.h | 78 CEED_INTERN int CeedBasisCreateTensorH1_Magma(CeedInt dim, CeedInt P_1d, CeedInt Q_1d, const CeedSc…
|
| /libCEED/backends/hip-gen/ |
| H A D | ceed-hip-gen-operator-build.h | 9 CEED_INTERN int BlockGridCalculate_Hip_gen(CeedInt dim, CeedInt num_elem, CeedInt P_1d, CeedInt Q_1…
|
| H A D | ceed-hip-gen-operator-build.cpp | 34 …alculate_Hip_gen(const CeedInt dim, const CeedInt num_elem, const CeedInt P_1d, const CeedInt Q_1d… in BlockGridCalculate_Hip_gen() argument
35 const CeedInt thread_1d = CeedIntMax(Q_1d, P_1d); in BlockGridCalculate_Hip_gen()
229 CeedInt elem_size = 0, num_comp = 0, dim = max_dim, P_1d = 0; in CeedOperatorBuildKernelFieldData_Hip_gen() local
249 if (is_tensor) CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedOperatorBuildKernelFieldData_Hip_gen()
250 else CeedCallBackend(CeedBasisGetNumNodes(basis, &P_1d)); in CeedOperatorBuildKernelFieldData_Hip_gen()
262 …code << tab << "const CeedInt " << P_name << " = " << (basis == CEED_BASIS_NONE ? Q_1d : P_1d) << … in CeedOperatorBuildKernelFieldData_Hip_gen()
279 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedOperatorBuildKernelFieldData_Hip_gen()
280 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedOperatorBuildKernelFieldData_Hip_gen()
317 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedOperatorBuildKernelFieldData_Hip_gen()
318 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedOperatorBuildKernelFieldData_Hip_gen()
[all …]
|
| /libCEED/backends/cuda-gen/ |
| H A D | ceed-cuda-gen-operator-build.cpp | 202 CeedInt elem_size = 0, num_comp = 0, dim = max_dim, P_1d = 0; in CeedOperatorBuildKernelFieldData_Cuda_gen() local
222 if (is_tensor) CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedOperatorBuildKernelFieldData_Cuda_gen()
223 else CeedCallBackend(CeedBasisGetNumNodes(basis, &P_1d)); in CeedOperatorBuildKernelFieldData_Cuda_gen()
235 …code << tab << "const CeedInt " << P_name << " = " << (basis == CEED_BASIS_NONE ? Q_1d : P_1d) << … in CeedOperatorBuildKernelFieldData_Cuda_gen()
252 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedOperatorBuildKernelFieldData_Cuda_gen()
253 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedOperatorBuildKernelFieldData_Cuda_gen()
290 interp_bytes = P_1d * Q_1d * sizeof(CeedScalar); in CeedOperatorBuildKernelFieldData_Cuda_gen()
291 CeedCallBackend(CeedCalloc(P_1d * Q_1d, &chebyshev_interp_1d)); in CeedOperatorBuildKernelFieldData_Cuda_gen()
532 CeedInt dim = max_dim, elem_size = 0, num_comp = 0, P_1d = 0; in CeedOperatorBuildKernelBasis_Cuda_gen() local
544 if (is_tensor) CeedCallBackend(CeedBasisGetNumNodes1D(basis, &P_1d)); in CeedOperatorBuildKernelBasis_Cuda_gen()
[all …]
|
| /libCEED/examples/fluids/src/ |
| H A D | dm_utils.c | 393 CeedInt P_1d = (CeedInt)round(pow(P, 1.0 / dim)); in CreateBasisFromPlex() local
396 …PetscCallCeed(ceed, CeedBasisCreateTensorH1Lagrange(ceed, dim, num_comp, P_1d, Q_1d, CEED_GAUSS, b… in CreateBasisFromPlex()
|