| /libCEED/python/tests/ |
| H A D | test-2-elemrestriction.py | 30 ind = np.zeros(2 * num_elem, dtype="int32") 32 ind[2 * i + 0] = i 33 ind[2 * i + 1] = i + 1 34 r = ceed.ElemRestriction(num_elem, 2, 1, 1, num_elem + 1, ind, 89 ind = np.zeros(2 * num_elem, dtype="int32") 91 ind[2 * i + 0] = i 92 ind[2 * i + 1] = i + 1 94 num_elem + 1, ind, cmode=libceed.COPY_VALUES) 111 assert y_array[indx] == a[ind[k * elem_size + i]] 137 ind = np.zeros(2 * num_elem, dtype="int32") [all …]
|
| /libCEED/include/ceed/jit-source/cuda/ |
| H A D | cuda-shared-basis-read-write-templates.h | 32 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d() local 35 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided1d() 48 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d() local 51 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided1d() 61 const CeedInt ind = node * strides_node + elem * strides_elem; in SumElementStrided1d() local 64 d_v[ind + comp * strides_comp] += r_v[comp]; in SumElementStrided1d() 81 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d() local 84 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided2d() 97 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d() local 100 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided2d() [all …]
|
| H A D | cuda-gen-templates.h | 30 const CeedInt ind = indices[p + elem * NUM_PTS]; in ReadPoint() local 33 r_u[comp] = d_u[ind + comp * COMP_STRIDE]; in ReadPoint() 44 const CeedInt ind = indices[p + elem * NUM_PTS]; in WritePoint() local 47 d_u[ind + comp * COMP_STRIDE] += r_u[comp]; in WritePoint() 77 const CeedInt ind = indices[node + elem * P_1D]; in ReadLVecStandard1d() local 79 for (CeedInt comp = 0; comp < NUM_COMP; comp++) r_u[comp] = d_u[ind + COMP_STRIDE * comp]; in ReadLVecStandard1d() 91 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in ReadLVecStrided1d() local 93 for (CeedInt comp = 0; comp < NUM_COMP; comp++) r_u[comp] = d_u[ind + comp * STRIDES_COMP]; in ReadLVecStrided1d() 105 const CeedInt ind = indices[node + elem * P_1D]; in WriteLVecStandard1d() local 107 …for (CeedInt comp = 0; comp < NUM_COMP; comp++) atomicAdd(&d_v[ind + COMP_STRIDE * comp], r_v[comp… in WriteLVecStandard1d() [all …]
|
| H A D | cuda-ref-restriction-offset.h | 17 const CeedInt ind = indices[node]; in OffsetNoTranspose() local 22 …v[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + elem * RSTR_ELEM_SIZE] = u[ind + comp * RSTR_… in OffsetNoTranspose() 33 const CeedInt ind = indices[node]; in OffsetTranspose() local 38 …atomicAdd(&v[ind + comp * RSTR_COMP_STRIDE], u[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + … in OffsetTranspose() 48 const CeedInt ind = l_vec_indices[i]; in OffsetTranspose() local 64 …for (CeedInt comp = 0; comp < RSTR_NUM_COMP; comp++) v[ind + comp * RSTR_COMP_STRIDE] += value[com… in OffsetTranspose()
|
| H A D | cuda-ref-restriction-oriented.h | 18 const CeedInt ind = indices[node]; in OrientedNoTranspose() local 24 …v[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + elem * RSTR_ELEM_SIZE] = u[ind + comp * RSTR_… in OrientedNoTranspose() 36 const CeedInt ind = indices[node]; in OrientedTranspose() local 42 atomicAdd(&v[ind + comp * RSTR_COMP_STRIDE], in OrientedTranspose() 54 const CeedInt ind = l_vec_indices[i]; in OrientedTranspose() local 71 …for (CeedInt comp = 0; comp < RSTR_NUM_COMP; comp++) v[ind + comp * RSTR_COMP_STRIDE] += value[com… in OrientedTranspose()
|
| /libCEED/include/ceed/jit-source/hip/ |
| H A D | hip-shared-basis-read-write-templates.h | 32 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d() local 35 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided1d() 48 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d() local 51 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided1d() 61 const CeedInt ind = node * strides_node + elem * strides_elem; in SumElementStrided1d() local 64 d_v[ind + comp * strides_comp] += r_v[comp]; in SumElementStrided1d() 81 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d() local 84 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided2d() 97 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d() local 100 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided2d() [all …]
|
| H A D | hip-gen-templates.h | 30 const CeedInt ind = indices[p + elem * NUM_PTS]; in ReadPoint() local 33 r_u[comp] = d_u[ind + comp * COMP_STRIDE]; in ReadPoint() 44 const CeedInt ind = indices[p + elem * NUM_PTS]; in WritePoint() local 47 d_u[ind + comp * COMP_STRIDE] += r_u[comp]; in WritePoint() 77 const CeedInt ind = indices[node + elem * P_1D]; in ReadLVecStandard1d() local 79 for (CeedInt comp = 0; comp < NUM_COMP; comp++) r_u[comp] = d_u[ind + COMP_STRIDE * comp]; in ReadLVecStandard1d() 90 const CeedInt ind = node * STRIDES_NODE + elem * STRIDES_ELEM; in ReadLVecStrided1d() local 92 for (CeedInt comp = 0; comp < NUM_COMP; comp++) r_u[comp] = d_u[ind + comp * STRIDES_COMP]; in ReadLVecStrided1d() 104 const CeedInt ind = indices[node + elem * P_1D]; in WriteLVecStandard1d() local 106 …for (CeedInt comp = 0; comp < NUM_COMP; comp++) atomicAdd(&d_v[ind + COMP_STRIDE * comp], r_v[comp… in WriteLVecStandard1d() [all …]
|
| H A D | hip-ref-restriction-offset.h | 17 const CeedInt ind = indices[node]; in OffsetNoTranspose() local 22 …v[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + elem * RSTR_ELEM_SIZE] = u[ind + comp * RSTR_… in OffsetNoTranspose() 33 const CeedInt ind = indices[node]; in OffsetTranspose() local 38 …atomicAdd(&v[ind + comp * RSTR_COMP_STRIDE], u[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + … in OffsetTranspose() 48 const CeedInt ind = l_vec_indices[i]; in OffsetTranspose() local 64 …for (CeedInt comp = 0; comp < RSTR_NUM_COMP; comp++) v[ind + comp * RSTR_COMP_STRIDE] += value[com… in OffsetTranspose()
|
| H A D | hip-ref-restriction-oriented.h | 18 const CeedInt ind = indices[node]; in OrientedNoTranspose() local 24 …v[loc_node + comp * RSTR_ELEM_SIZE * RSTR_NUM_ELEM + elem * RSTR_ELEM_SIZE] = u[ind + comp * RSTR_… in OrientedNoTranspose() 36 const CeedInt ind = indices[node]; in OrientedTranspose() local 42 atomicAdd(&v[ind + comp * RSTR_COMP_STRIDE], in OrientedTranspose() 54 const CeedInt ind = l_vec_indices[i]; in OrientedTranspose() local 71 …for (CeedInt comp = 0; comp < RSTR_NUM_COMP; comp++) v[ind + comp * RSTR_COMP_STRIDE] += value[com… in OrientedTranspose()
|
| /libCEED/tests/ |
| H A D | t209-elemrestriction.c | 11 CeedInt ind[4 * num_elem]; in main() local 19 ind[4 * i + 0] = i * 3 + 0; in main() 20 ind[4 * i + 1] = i * 3 + 1; in main() 21 ind[4 * i + 2] = i * 3 + 2; in main() 22 ind[4 * i + 3] = i * 3 + 3; in main() 24 …eed, num_elem, 4, 1, 1, 3 * num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t209-elemrestriction-f.f90 | 15 integer*4 ind(4*ne) variable 29 ind(4*i-3)=3*i-3 30 ind(4*i-2)=3*i-2 31 ind(4*i-1)=3*i-1 32 ind(4*i-0)=3*i-0 35 & ceed_use_pointer,ind,r,err)
|
| H A D | t216-elemrestriction.c | 10 CeedInt ind[2 * num_elem]; in main() local 16 ind[2 * i + 0] = i; in main() 17 ind[2 * i + 1] = i + 1; in main() 19 …um_elem, 2, 1, comp_stride, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main() 20 …lem, 2, 1, comp_stride + 1, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction_… in main()
|
| H A D | t230-elemrestriction.c | 9 CeedInt ind[(num_elem + 1) + num_points]; in main() local 21 ind[i] = offset; in main() 23 ind[offset + j] = point_index; in main() 28 ind[num_elem] = offset; in main() 30 …eed, num_elem, num_points, 1, num_points, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t210-elemrestriction-f.f90 | 13 integer*4 ind(2*ne) variable 21 ind(2*i-1)=i-1 22 ind(2*i )=i 26 & ceed_use_pointer,ind,r,err)
|
| H A D | t205-elemrestriction.c | 12 CeedInt ind[2 * num_elem]; in main() local 29 ind[2 * i + 0] = 2 * i; in main() 30 ind[2 * i + 1] = 2 * (i + 1); in main() 32 …d, num_elem, 2, 2, 1, 2 * (num_elem + 1), CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main() 46 … if (y_array[i * e_layout[0] + j * e_layout[1] + k * e_layout[2]] != x_array[ind[i + k * 2] + j]) { in main() 49 …alar)y_array[i * e_layout[0] + j * e_layout[1] + k * e_layout[2]], x_array[ind[i + k * 2] + j * (n… in main()
|
| H A D | t204-elemrestriction.c | 12 CeedInt ind[2 * num_elem]; in main() local 29 ind[2 * i + 0] = i; in main() 30 ind[2 * i + 1] = i + 1; in main() 32 …, 2, 2, num_elem + 1, 2 * (num_elem + 1), CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main() 46 …if (y_array[i * e_layout[0] + j * e_layout[1] + k * e_layout[2]] != x_array[ind[i + k * 2] + j * (… in main() 49 …alar)y_array[i * e_layout[0] + j * e_layout[1] + k * e_layout[2]], x_array[ind[i + k * 2] + j * (n… in main()
|
| H A D | t215-elemrestriction.c | 9 CeedInt ind[2 * num_elem]; in main() local 16 ind[2 * i + 0] = i; in main() 17 ind[2 * i + 1] = i + 1; in main() 19 …te(ceed, num_elem, 2, 1, 1, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t214-elemrestriction.c | 9 CeedInt ind[2 * num_elem]; in main() local 16 ind[2 * i + 0] = i; in main() 17 ind[2 * i + 1] = i + 1; in main() 19 …te(ceed, num_elem, 2, 1, 1, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t210-elemrestriction.c | 9 CeedInt ind[2 * num_elem]; in main() local 15 ind[2 * i + 0] = i; in main() 16 ind[2 * i + 1] = i + 1; in main() 18 …te(ceed, num_elem, 2, 1, 1, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t234-elemrestriction.c | 11 CeedInt ind[(num_elem + 1) + num_points]; in main() local 24 ind[i] = offset; in main() 26 ind[offset + j] = point_index; in main() 31 ind[num_elem] = offset; in main() 33 …eed, num_elem, num_points, 1, num_points, CEED_MEM_HOST, CEED_COPY_VALUES, ind, &elem_restriction); in main()
|
| H A D | t232-elemrestriction.c | 11 CeedInt ind[(num_elem + 1) + num_points]; in main() local 24 ind[i] = offset; in main() 26 ind[offset + j] = point_index; in main() 31 ind[num_elem] = offset; in main() 33 …eed, num_elem, num_points, 1, num_points, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t200-elemrestriction.c | 11 CeedInt ind[2 * num_elem]; in main() local 26 ind[2 * i + 0] = i; in main() 27 ind[2 * i + 1] = i + 1; in main() 29 …te(ceed, num_elem, 2, 1, 1, num_elem + 1, CEED_MEM_HOST, CEED_USE_POINTER, ind, &elem_restriction); in main()
|
| H A D | t200-elemrestriction-f.f90 | 15 integer*4 ind(2*ne) variable 35 ind(2*i-1)=i-1 36 ind(2*i )=i 40 & ceed_use_pointer,ind,r,err)
|
| /libCEED/include/ceed/jit-source/sycl/ |
| H A D | sycl-shared-basis-read-write-templates.h | 36 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided1d() local 38 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided1d() 54 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided1d() local 56 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided1d() 77 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided2d() local 79 r_u[comp] = d_u[ind + comp * strides_comp]; in ReadElementStrided2d() 96 const CeedInt ind = node * strides_node + elem * strides_elem; in WriteElementStrided2d() local 98 d_v[ind + comp * strides_comp] = r_v[comp]; in WriteElementStrided2d() 120 const CeedInt ind = node * strides_node + elem * strides_elem; in ReadElementStrided3d() local 122 r_u[z + comp * P_1D] = d_u[ind + comp * strides_comp]; in ReadElementStrided3d() [all …]
|
| H A D | sycl-gen-templates.h | 40 const CeedInt ind = indices[node + elem * P_1D]; in readDofsOffset1d() local 42 r_u[comp] = d_u[ind + strides_comp * comp]; in readDofsOffset1d() 58 const CeedInt ind = node * strides_node + elem * strides_elem; in readDofsStrided1d() local 60 r_u[comp] = d_u[ind + comp * strides_comp]; in readDofsStrided1d() 75 const CeedInt ind = indices[node + elem * P_1D]; in writeDofsOffset1d() local 77 …atomic_fetch_add_explicit(&d_v[ind + strides_comp * comp], r_v[comp], memory_order_relaxed, memory… in writeDofsOffset1d() 92 const CeedInt ind = node * strides_node + elem * strides_elem; in writeDofsStrided1d() local 94 d_v[ind + comp * strides_comp] = r_v[comp]; in writeDofsStrided1d() 114 const CeedInt ind = indices[node + elem * P_1D * P_1D]; in readDofsOffset2d() local 115 for (CeedInt comp = 0; comp < num_comp; ++comp) r_u[comp] = d_u[ind + strides_comp * comp]; in readDofsOffset2d() [all …]
|