| /libCEED/include/ceed/jit-source/cuda/ |
| H A D | cuda-shared-basis-tensor-at-points-templates.h | 15 template <int Q_1D> 19 …for (CeedInt i = 2; i < Q_1D; i++) chebyshev_x[i] = 2 * x * chebyshev_x[i - 1] - chebyshev_x[i - 2… in ChebyshevPolynomialsAtPoint() 22 template <int Q_1D> 30 for (CeedInt i = 2; i < Q_1D; i++) { in ChebyshevDerivativeAtPoint() 43 template <int NUM_COMP, int NUM_POINTS, int P_1D, int Q_1D> 46 CeedScalar chebyshev_x[Q_1D]; in InterpAtPoints1d() 49 ChebyshevPolynomialsAtPoint<Q_1D>(r_X[0], chebyshev_x); in InterpAtPoints1d() 52 if (data.t_id_x < Q_1D) data.slice[data.t_id_x] = r_C[comp]; in InterpAtPoints1d() 55 for (CeedInt i = 0; i < Q_1D; i++) { in InterpAtPoints1d() 64 template <int NUM_COMP, int NUM_POINTS, int P_1D, int Q_1D> [all …]
|
| H A D | cuda-shared-basis-tensor-templates.h | 19 template <int NUM_COMP, int P_1D, int Q_1D> 25 if (data.t_id_x < Q_1D) { in ContractX1d() 35 template <int NUM_COMP, int P_1D, int Q_1D> 42 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeX1d() 51 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 54 ContractX1d<NUM_COMP, P_1D, Q_1D>(data, &r_U[comp], c_B, &r_V[comp]); in Interp1d() 61 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 65 ContractTransposeX1d<NUM_COMP, P_1D, Q_1D>(data, &r_U[comp], c_B, &r_V[comp]); in InterpTranspose1d() 72 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 83 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> [all …]
|
| H A D | cuda-shared-basis-tensor-flattened-templates.h | 19 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 26 if (t_id_x < Q_1D && t_id_y < P_1D) { in ContractX2dFlattened() 36 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 43 if (t_id_x < Q_1D && t_id_y < Q_1D) { in ContractY2dFlattened() 53 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 60 if (t_id_x < Q_1D && t_id_y < P_1D) { in ContractTransposeY2dFlattened() 61 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeY2dFlattened() 70 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 78 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeX2dFlattened() 87 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> [all …]
|
| H A D | cuda-gen-templates.h | 147 template <int NUM_COMP_OUT, int NUM_COMP_FIELD, int Q_1D> 150 if (data.t_id_x < Q_1D) { in WriteLVecStandard1d_QFAssembly() 151 const CeedInt ind = data.t_id_x + elem * Q_1D; in WriteLVecStandard1d_QFAssembly() 154 … d_v[ind + (input_offset * NUM_COMP_OUT + output_offset + comp) * (Q_1D * num_elem)] = r_v[comp]; in WriteLVecStandard1d_QFAssembly() 276 template <int NUM_COMP_OUT, int NUM_COMP_FIELD, int Q_1D> 279 if (data.t_id_x < Q_1D && data.t_id_y < Q_1D) { in WriteLVecStandard2d_QFAssembly() 280 const CeedInt ind = (data.t_id_x + data.t_id_y * Q_1D) + elem * Q_1D * Q_1D; in WriteLVecStandard2d_QFAssembly() 283 …d_v[ind + (input_offset * NUM_COMP_OUT + output_offset + comp) * (Q_1D * Q_1D * num_elem)] = r_v[c… in WriteLVecStandard2d_QFAssembly() 356 template <int NUM_COMP, int COMP_STRIDE, int Q_1D> 360 if (data.t_id_x < Q_1D && data.t_id_y < Q_1D) { in ReadEVecSliceStandard3d() [all …]
|
| H A D | cuda-shared-basis-nontensor-templates.h | 15 template <int NUM_COMP, int P_1D, int Q_1D> 20 if (data.t_id_x < Q_1D) { in Contract1d() 31 template <int NUM_COMP, int P_1D, int Q_1D> 36 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTranspose1d()
|
| H A D | cuda-ref-basis-tensor-at-points.h | 15 template <int Q_1D> 19 …for (CeedInt i = 2; i < Q_1D; i++) chebyshev_x[i] = 2 * x * chebyshev_x[i - 1] - chebyshev_x[i - 2… in ChebyshevPolynomialsAtPoint() 22 template <int Q_1D> 30 for (CeedInt i = 2; i < Q_1D; i++) { in ChebyshevDerivativeAtPoint()
|
| /libCEED/include/ceed/jit-source/hip/ |
| H A D | hip-shared-basis-tensor-at-points-templates.h | 15 template <int Q_1D> 19 …for (CeedInt i = 2; i < Q_1D; i++) chebyshev_x[i] = 2 * x * chebyshev_x[i - 1] - chebyshev_x[i - 2… in ChebyshevPolynomialsAtPoint() 22 template <int Q_1D> 30 for (CeedInt i = 2; i < Q_1D; i++) { in ChebyshevDerivativeAtPoint() 43 template <int NUM_COMP, int NUM_POINTS, int P_1D, int Q_1D> 46 CeedScalar chebyshev_x[Q_1D]; in InterpAtPoints1d() 49 ChebyshevPolynomialsAtPoint<Q_1D>(r_X[0], chebyshev_x); in InterpAtPoints1d() 53 if (data.t_id_x < Q_1D) data.slice[data.t_id_x] = r_C[comp]; in InterpAtPoints1d() 56 for (CeedInt i = 0; i < Q_1D; i++) { in InterpAtPoints1d() 65 template <int NUM_COMP, int NUM_POINTS, int P_1D, int Q_1D> [all …]
|
| H A D | hip-shared-basis-tensor-templates.h | 19 template <int NUM_COMP, int P_1D, int Q_1D> 25 if (data.t_id_x < Q_1D) { in ContractX1d() 35 template <int NUM_COMP, int P_1D, int Q_1D> 42 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeX1d() 51 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 54 ContractX1d<NUM_COMP, P_1D, Q_1D>(data, &r_U[comp], c_B, &r_V[comp]); in Interp1d() 61 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 65 ContractTransposeX1d<NUM_COMP, P_1D, Q_1D>(data, &r_U[comp], c_B, &r_V[comp]); in InterpTranspose1d() 72 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 83 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> [all …]
|
| H A D | hip-shared-basis-tensor-flattened-templates.h | 19 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 26 if (t_id_x < Q_1D && t_id_y < P_1D) { in ContractX2dFlattened() 36 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 43 if (t_id_x < Q_1D && t_id_y < Q_1D) { in ContractY2dFlattened() 53 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 60 if (t_id_x < Q_1D && t_id_y < P_1D) { in ContractTransposeY2dFlattened() 61 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeY2dFlattened() 70 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> 78 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeX2dFlattened() 87 template <int NUM_COMP, int P_1D, int Q_1D, int T_1D> [all …]
|
| H A D | hip-gen-templates.h | 146 template <int NUM_COMP_OUT, int NUM_COMP_FIELD, int Q_1D> 149 if (data.t_id_x < Q_1D) { in WriteLVecStandard1d_QFAssembly() 150 const CeedInt ind = data.t_id_x + elem * Q_1D; in WriteLVecStandard1d_QFAssembly() 153 … d_v[ind + (input_offset * NUM_COMP_OUT + output_offset + comp) * (Q_1D * num_elem)] = r_v[comp]; in WriteLVecStandard1d_QFAssembly() 274 template <int NUM_COMP_OUT, int NUM_COMP_FIELD, int Q_1D> 277 if (data.t_id_x < Q_1D && data.t_id_y < Q_1D) { in WriteLVecStandard2d_QFAssembly() 278 const CeedInt ind = (data.t_id_x + data.t_id_y * Q_1D) + elem * Q_1D * Q_1D; in WriteLVecStandard2d_QFAssembly() 281 …d_v[ind + (input_offset * NUM_COMP_OUT + output_offset + comp) * (Q_1D * Q_1D * num_elem)] = r_v[c… in WriteLVecStandard2d_QFAssembly() 353 template <int NUM_COMP, int COMP_STRIDE, int Q_1D> 357 if (data.t_id_x < Q_1D && data.t_id_y < Q_1D) { in ReadEVecSliceStandard3d() [all …]
|
| H A D | hip-shared-basis-nontensor-templates.h | 15 template <int NUM_COMP, int P_1D, int Q_1D> 20 if (data.t_id_x < Q_1D) { in Contract1d() 31 template <int NUM_COMP, int P_1D, int Q_1D> 36 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTranspose1d()
|
| H A D | hip-ref-basis-tensor-at-points.h | 15 template <int Q_1D> 19 …for (CeedInt i = 2; i < Q_1D; i++) chebyshev_x[i] = 2 * x * chebyshev_x[i - 1] - chebyshev_x[i - 2… in ChebyshevPolynomialsAtPoint() 22 template <int Q_1D> 30 for (CeedInt i = 2; i < Q_1D; i++) { in ChebyshevDerivativeAtPoint()
|
| /libCEED/include/ceed/jit-source/sycl/ |
| H A D | sycl-shared-basis-tensor-templates.h | 19 inline void ContractX1d(const CeedInt P_1D, const CeedInt Q_1D, private const CeedScalar *restrict … in ContractX1d() argument 27 if (item_id_x < Q_1D) { in ContractX1d() 38 inline void ContractTransposeX1d(const CeedInt P_1D, const CeedInt Q_1D, private const CeedScalar *… in ContractTransposeX1d() argument 47 for (CeedInt i = 0; i < Q_1D; i++) { in ContractTransposeX1d() 57 inline void Interp1d(const CeedInt NUM_COMP, const CeedInt P_1D, const CeedInt Q_1D, private const … in Interp1d() argument 60 ContractX1d(P_1D, Q_1D, r_U + comp, s_B, r_V + comp, scratch); in Interp1d() 67 inline void InterpTranspose1d(const CeedInt NUM_COMP, const CeedInt P_1D, const CeedInt Q_1D, priva… in InterpTranspose1d() argument 70 ContractTransposeX1d(P_1D, Q_1D, r_U + comp, s_B, r_V + comp, scratch); in InterpTranspose1d() 77 inline void Grad1d(const CeedInt NUM_COMP, const CeedInt P_1D, const CeedInt Q_1D, private const Ce… in Grad1d() argument 80 ContractX1d(P_1D, Q_1D, r_U + comp, s_G, r_V + comp, scratch); in Grad1d() [all …]
|
| H A D | sycl-gen-templates.h | 214 …ffset3d(const CeedInt num_comp, const CeedInt strides_comp, const CeedInt Q_1D, const CeedInt num_… in readSliceQuadsOffset3d() argument 220 if (item_id_x < Q_1D && item_id_y < Q_1D && elem < num_elem) { in readSliceQuadsOffset3d() 221 const CeedInt node = item_id_x + Q_1D * (item_id_y + Q_1D * q); in readSliceQuadsOffset3d() 222 const CeedInt ind = indices[node + elem * Q_1D * Q_1D * Q_1D]; in readSliceQuadsOffset3d() 230 inline void readSliceQuadsStrided3d(const CeedInt num_comp, const CeedInt Q_1D, CeedInt strides_nod… in readSliceQuadsStrided3d() argument 237 if (item_id_x < Q_1D && item_id_y < Q_1D && elem < num_elem) { in readSliceQuadsStrided3d() 238 const CeedInt node = item_id_x + Q_1D * (item_id_y + Q_1D * q); in readSliceQuadsStrided3d() 285 inline void gradCollo3d(const CeedInt num_comp, const CeedInt Q_1D, const CeedInt q, const private … in gradCollo3d() argument 291 if (item_id_x < Q_1D && item_id_y < Q_1D) { in gradCollo3d() 292 scratch[item_id_x + item_id_y * T_1D] = r_U[q + comp * Q_1D]; in gradCollo3d() [all …]
|