Home
last modified time | relevance | path

Searched refs:node (Results 1 – 25 of 34) sorted by relevance

12

/libCEED/include/ceed/jit-source/hip/
H A Dhip-ref-restriction-curl-oriented.h17 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in CurlOrientedNoTranspose() local
18 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in CurlOrientedNoTranspose()
19 const CeedInt elem = node / RSTR_ELEM_SIZE; in CurlOrientedNoTranspose()
20 const CeedInt ind_dl = loc_node > 0 ? indices[node - 1] : 0; in CurlOrientedNoTranspose()
21 const CeedInt ind_d = indices[node]; in CurlOrientedNoTranspose()
22 const CeedInt ind_du = loc_node < (RSTR_ELEM_SIZE - 1) ? indices[node + 1] : 0; in CurlOrientedNoTranspose()
23 const CeedInt8 curl_orient_dl = curl_orients[3 * node + 0]; in CurlOrientedNoTranspose()
24 const CeedInt8 curl_orient_d = curl_orients[3 * node + 1]; in CurlOrientedNoTranspose()
25 const CeedInt8 curl_orient_du = curl_orients[3 * node + 2]; in CurlOrientedNoTranspose()
42 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in CurlOrientedUnsignedNoTranspose() local
[all …]
H A Dhip-ref-restriction-strided.h16 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in StridedNoTranspose() local
17 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in StridedNoTranspose()
18 const CeedInt elem = node / RSTR_ELEM_SIZE; in StridedNoTranspose()
31 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in StridedTranspose() local
32 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in StridedTranspose()
33 const CeedInt elem = node / RSTR_ELEM_SIZE; in StridedTranspose()
H A Dhip-ref-restriction-oriented.h17 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OrientedNoTranspose() local
18 const CeedInt ind = indices[node]; in OrientedNoTranspose()
19 const bool orient = orients[node]; in OrientedNoTranspose()
20 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OrientedNoTranspose()
21 const CeedInt elem = node / RSTR_ELEM_SIZE; in OrientedNoTranspose()
35 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OrientedTranspose() local
36 const CeedInt ind = indices[node]; in OrientedTranspose()
37 const bool orient = orients[node]; in OrientedTranspose()
38 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OrientedTranspose()
39 const CeedInt elem = node / RSTR_ELEM_SIZE; in OrientedTranspose()
H A Dhip-ref-restriction-offset.h16 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OffsetNoTranspose() local
17 const CeedInt ind = indices[node]; in OffsetNoTranspose()
18 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OffsetNoTranspose()
19 const CeedInt elem = node / RSTR_ELEM_SIZE; in OffsetNoTranspose()
32 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OffsetTranspose() local
33 const CeedInt ind = indices[node]; in OffsetTranspose()
34 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OffsetTranspose()
35 const CeedInt elem = node / RSTR_ELEM_SIZE; in OffsetTranspose()
H A Dhip-shared-basis-read-write-templates.h31 const CeedInt node = data.t_id_x; in ReadElementStrided1d() local
32 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d()
47 const CeedInt node = data.t_id_x; in WriteElementStrided1d() local
48 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d()
60 const CeedInt node = data.t_id_x; in SumElementStrided1d() local
61 const CeedInt ind = node * strides_node + elem * strides_elem; in SumElementStrided1d()
80 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in ReadElementStrided2d() local
81 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d()
96 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in WriteElementStrided2d() local
97 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d()
[all …]
H A Dhip-ref-restriction-at-points.h18 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in AtPointsTranspose() local
19 const CeedInt ind = indices[node]; in AtPointsTranspose()
20 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in AtPointsTranspose()
21 const CeedInt elem = node / RSTR_ELEM_SIZE; in AtPointsTranspose()
H A Dhip-gen-templates.h76 const CeedInt node = data.t_id_x; in ReadLVecStandard1d() local
77 const CeedInt ind = indices[node + elem * P_1D]; in ReadLVecStandard1d()
89 const CeedInt node = data.t_id_x; in ReadLVecStrided1d() local
90 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in ReadLVecStrided1d()
103 const CeedInt node = data.t_id_x; in WriteLVecStandard1d() local
104 const CeedInt ind = indices[node + elem * P_1D]; in WriteLVecStandard1d()
165 const CeedInt node = data.t_id_x; in WriteLVecStrided1d() local
166 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in WriteLVecStrided1d()
197 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in ReadLVecStandard2d() local
198 const CeedInt ind = indices[node + elem * P_1D * P_1D]; in ReadLVecStandard2d()
[all …]
/libCEED/include/ceed/jit-source/cuda/
H A Dcuda-ref-restriction-curl-oriented.h17 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in CurlOrientedNoTranspose() local
18 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in CurlOrientedNoTranspose()
19 const CeedInt elem = node / RSTR_ELEM_SIZE; in CurlOrientedNoTranspose()
20 const CeedInt ind_dl = loc_node > 0 ? indices[node - 1] : 0; in CurlOrientedNoTranspose()
21 const CeedInt ind_d = indices[node]; in CurlOrientedNoTranspose()
22 const CeedInt ind_du = loc_node < (RSTR_ELEM_SIZE - 1) ? indices[node + 1] : 0; in CurlOrientedNoTranspose()
23 const CeedInt8 curl_orient_dl = curl_orients[3 * node + 0]; in CurlOrientedNoTranspose()
24 const CeedInt8 curl_orient_d = curl_orients[3 * node + 1]; in CurlOrientedNoTranspose()
25 const CeedInt8 curl_orient_du = curl_orients[3 * node + 2]; in CurlOrientedNoTranspose()
42 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in CurlOrientedUnsignedNoTranspose() local
[all …]
H A Dcuda-ref-restriction-strided.h16 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in StridedNoTranspose() local
17 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in StridedNoTranspose()
18 const CeedInt elem = node / RSTR_ELEM_SIZE; in StridedNoTranspose()
31 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in StridedTranspose() local
32 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in StridedTranspose()
33 const CeedInt elem = node / RSTR_ELEM_SIZE; in StridedTranspose()
H A Dcuda-ref-restriction-oriented.h17 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OrientedNoTranspose() local
18 const CeedInt ind = indices[node]; in OrientedNoTranspose()
19 const bool orient = orients[node]; in OrientedNoTranspose()
20 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OrientedNoTranspose()
21 const CeedInt elem = node / RSTR_ELEM_SIZE; in OrientedNoTranspose()
35 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OrientedTranspose() local
36 const CeedInt ind = indices[node]; in OrientedTranspose()
37 const bool orient = orients[node]; in OrientedTranspose()
38 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OrientedTranspose()
39 const CeedInt elem = node / RSTR_ELEM_SIZE; in OrientedTranspose()
H A Dcuda-ref-restriction-offset.h16 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OffsetNoTranspose() local
17 const CeedInt ind = indices[node]; in OffsetNoTranspose()
18 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OffsetNoTranspose()
19 const CeedInt elem = node / RSTR_ELEM_SIZE; in OffsetNoTranspose()
32 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in OffsetTranspose() local
33 const CeedInt ind = indices[node]; in OffsetTranspose()
34 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in OffsetTranspose()
35 const CeedInt elem = node / RSTR_ELEM_SIZE; in OffsetTranspose()
H A Dcuda-shared-basis-read-write-templates.h31 const CeedInt node = data.t_id_x; in ReadElementStrided1d() local
32 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d()
47 const CeedInt node = data.t_id_x; in WriteElementStrided1d() local
48 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d()
60 const CeedInt node = data.t_id_x; in SumElementStrided1d() local
61 const CeedInt ind = node * strides_node + elem * strides_elem; in SumElementStrided1d()
80 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in ReadElementStrided2d() local
81 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d()
96 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in WriteElementStrided2d() local
97 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d()
[all …]
H A Dcuda-ref-restriction-at-points.h18 …for (CeedInt node = blockIdx.x * blockDim.x + threadIdx.x; node < RSTR_NUM_ELEM * RSTR_ELEM_SIZE; … in AtPointsTranspose() local
19 const CeedInt ind = indices[node]; in AtPointsTranspose()
20 const CeedInt loc_node = node % RSTR_ELEM_SIZE; in AtPointsTranspose()
21 const CeedInt elem = node / RSTR_ELEM_SIZE; in AtPointsTranspose()
H A Dcuda-gen-templates.h76 const CeedInt node = data.t_id_x; in ReadLVecStandard1d() local
77 const CeedInt ind = indices[node + elem * P_1D]; in ReadLVecStandard1d()
90 const CeedInt node = data.t_id_x; in ReadLVecStrided1d() local
91 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in ReadLVecStrided1d()
104 const CeedInt node = data.t_id_x; in WriteLVecStandard1d() local
105 const CeedInt ind = indices[node + elem * P_1D]; in WriteLVecStandard1d()
166 const CeedInt node = data.t_id_x; in WriteLVecStrided1d() local
167 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in WriteLVecStrided1d()
198 const CeedInt node = data.t_id_x + data.t_id_y * P_1D; in ReadLVecStandard2d() local
199 const CeedInt ind = indices[node + elem * P_1D * P_1D]; in ReadLVecStandard2d()
[all …]
/libCEED/doc/img/tex/
H A DlibCEEDBackends.tex20 \node at (0.8,6.1) {\large Application};
27 node[pos=.5,align=center,color=black] {Ratel};
36 node[pos=.5,align=center,color=black] {PETSc};
45 node[pos=.5,align=center,color=black] {Nek5000};
54 node[pos=.5,align=center,color=black] {MFEM};
60 \node at (0.8,6.1) {\large Library};
66 node[pos=.5,align=center,color=black] {libCEED};
80 \node at (0.95,6.1) {\large Backends};
88 node[pos=.5,align=center,color=black] {Pure C};
97 node[pos=.5,align=center,color=black] {AVX};
[all …]
/libCEED/doc/img/
H A DlibCEEDBackends.tex22 \node at (1.0,6.1) {\large Application};
30 node[pos=.5,align=center,color=black] {PETSc};
39 node[pos=.5,align=center,color=black] {Ratel};
48 node[pos=.5,align=center,color=black] {HONEE};
57 node[pos=.5,align=center,color=black] {MFEM};
63 \node at (0.8,6.1) {\large Library};
69 node[pos=.5,align=center,color=black] {libCEED};
83 \node at (0.95,6.1) {\large Backends};
91 node[pos=.5,align=center,color=black] {Pure C};
100 node[pos=.5,align=center,color=black] {AVX};
[all …]
/libCEED/include/ceed/jit-source/sycl/
H A Dsycl-shared-basis-read-write-templates.h35 const CeedInt node = item_id_x; in ReadElementStrided1d() local
36 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d()
53 const CeedInt node = item_id_x; in WriteElementStrided1d() local
54 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d()
76 const CeedInt node = item_id_x + item_id_y * P_1D; in ReadElementStrided2d() local
77 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d()
95 const CeedInt node = item_id_x + item_id_y * P_1D; in WriteElementStrided2d() local
96 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d()
119 const CeedInt node = item_id_x + item_id_y * P_1D + z * P_1D * P_1D; in ReadElementStrided3d() local
120 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided3d()
[all …]
H A Dsycl-gen-templates.h39 const CeedInt node = item_id_x; in readDofsOffset1d() local
40 const CeedInt ind = indices[node + elem * P_1D]; in readDofsOffset1d()
57 const CeedInt node = item_id_x; in readDofsStrided1d() local
58 const CeedInt ind = node * strides_node + elem * strides_elem; in readDofsStrided1d()
74 const CeedInt node = item_id_x; in writeDofsOffset1d() local
75 const CeedInt ind = indices[node + elem * P_1D]; in writeDofsOffset1d()
91 const CeedInt node = item_id_x; in writeDofsStrided1d() local
92 const CeedInt ind = node * strides_node + elem * strides_elem; in writeDofsStrided1d()
113 const CeedInt node = item_id_x + item_id_y * P_1D; in readDofsOffset2d() local
114 const CeedInt ind = indices[node + elem * P_1D * P_1D]; in readDofsOffset2d()
[all …]
/libCEED/backends/sycl-ref/
H A Dceed-sycl-restriction.sycl.cpp40 sycl_queue.parallel_for<CeedElemRestrSyclStridedNT>(kernel_range, e, [=](sycl::id<1> node) { in CeedElemRestrictionStridedNoTranspose_Sycl() argument
41 const CeedInt loc_node = node % elem_size; in CeedElemRestrictionStridedNoTranspose_Sycl()
42 const CeedInt elem = node / elem_size; in CeedElemRestrictionStridedNoTranspose_Sycl()
67 sycl_queue.parallel_for<CeedElemRestrSyclOffsetNT>(kernel_range, e, [=](sycl::id<1> node) { in CeedElemRestrictionOffsetNoTranspose_Sycl() argument
68 const CeedInt ind = indices[node]; in CeedElemRestrictionOffsetNoTranspose_Sycl()
69 const CeedInt loc_node = node % elem_size; in CeedElemRestrictionOffsetNoTranspose_Sycl()
70 const CeedInt elem = node / elem_size; in CeedElemRestrictionOffsetNoTranspose_Sycl()
96 sycl_queue.parallel_for<CeedElemRestrSyclStridedT>(kernel_range, e, [=](sycl::id<1> node) { in CeedElemRestrictionStridedTranspose_Sycl() argument
97 const CeedInt loc_node = node % elem_size; in CeedElemRestrictionStridedTranspose_Sycl()
98 const CeedInt elem = node / elem_size; in CeedElemRestrictionStridedTranspose_Sycl()
[all …]
/libCEED/julia/LibCEED.jl/src/
H A DElemRestriction.jl57 - `ncomp`: Number of field components per interpolation node (1 for scalar fields)
58 - `compstride`: Stride between components for the same L-vector "node". Data for node $i$,
121 - `ncomp`: Number of field components per interpolation node (1 for scalar fields)
122 - `compstride`: Stride between components for the same L-vector "node". Data for node $i$,
189 - `ncomp`: Number of field components per interpolation node (1 for scalar fields)
190 - `compstride`: Stride between components for the same L-vector "node". Data for node $i$,
248 - `ncomp`: Number of field components per interpolation node (1 for scalar fields)
251 - `strides`: Array for strides between [nodes, components, elements]. Data for node $i$,
/libCEED/benchmarks/
H A DREADME.md15 of processors per node.
30 bound of the problem sizes, per compute node; the default value is 3*2^20.
H A Dbenchmark.sh170 -p|--proc-node)
/libCEED/examples/fluids/src/
H A Dsetupts.c86 for (PetscInt node = 0; node < dof / num_comp; node++) { in Surface_Forces_NS() local
88 reaction_force[w * dim + j] -= r[node].momentum[j]; in Surface_Forces_NS()
/libCEED/interface/
H A Dceed-basis.c2343 for (CeedInt node = 0; node < basis->P; node++) { in CeedBasisGetInterp() local
2344 CeedInt p = (node / CeedIntPow(basis->P_1d, d)) % basis->P_1d; in CeedBasisGetInterp()
2347 basis->interp[qpt * (basis->P) + node] *= basis->interp_1d[q * basis->P_1d + p]; in CeedBasisGetInterp()
2397 for (CeedInt node = 0; node < basis->P; node++) { in CeedBasisGetGrad() local
2398 CeedInt p = (node / CeedIntPow(basis->P_1d, d)) % basis->P_1d; in CeedBasisGetGrad()
2401 …if (i == d) basis->grad[(i * basis->Q + qpt) * (basis->P) + node] *= basis->grad_1d[q * basis->P_1… in CeedBasisGetGrad()
2402 …else basis->grad[(i * basis->Q + qpt) * (basis->P) + node] *= basis->interp_1d[q * basis->P_1d + p… in CeedBasisGetGrad()
/libCEED/doc/sphinx/source/
H A DlibCEEDdev.md97 …Data for node `i`, component `j`, element `k` can be found in the L-vector at index `offsets[i + k…
99 …Data for node `i`, component `j`, element `k` can be found in the L-vector at index `i*strides[0] …
104 …If the backend uses a strided E-vector layout, then the data for node `i`, component `j`, element …

12