libceed-sys 0.12.0

Low-level bindings for libCEED library.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162

// Copyright (c) 2017-2022, Lawrence Livermore National Security, LLC and other CEED contributors.
// All Rights Reserved. See the top-level LICENSE and NOTICE files for details.
//
// SPDX-License-Identifier: BSD-2-Clause
//
// This file is part of CEED:  http://github.com/ceed

#include "../../kernel-defines.hpp"

const char *occa_tensor_basis_2d_gpu_source = STRINGIFY_SOURCE(

    typedef CeedScalar * dofArray @dim(P1D, P1D, BASIS_COMPONENT_COUNT, elementCount);
    typedef const CeedScalar *const_dofArray @dim(P1D, P1D, BASIS_COMPONENT_COUNT, elementCount);

    typedef CeedScalar * quadArray @dim(Q1D, Q1D, elementCount, BASIS_COMPONENT_COUNT, 2);
    typedef const CeedScalar *const_quadArray @dim(Q1D, Q1D, elementCount, BASIS_COMPONENT_COUNT, 2);

    typedef CeedScalar * sharedBufferArray @dim(MAX_PQ, MAX_PQ, ELEMENTS_PER_BLOCK); typedef const CeedScalar *quadToDof @dim(P1D, Q1D);
    typedef CeedScalar * elementWeightArray @dim(Q1D, Q1D, elementCount);

    //---[ Utility Methods ]----------------
    inline void readDofs(const int element, const int component, const int px, const int py, const_dofArray U, CeedScalar *Up) {
      // Zero out extra entries
      *Up = ((px < P1D) && (py < P1D) ? U(px, py, component, element) : 0.0);
    }

    inline void writeDofs(const int element, const int component, const int px, const int py, const CeedScalar Vp, dofArray V) {
      if ((px < P1D) && (py < P1D)) {
        V(px, py, component, element) = Vp;
      }
    }

    inline void readQuads(const int elementCount, const int element, const int component, const int qx, const int qy, const int dim,
                          const_quadArray U, CeedScalar *Uq) { *Uq = U(qx, qy, element, component, dim); }

    inline void writeQuads(const int elementCount, const int element, const int component, const int qx, const int qy, const int dim,
                           const CeedScalar Vq, quadArray V) { V(qx, qy, element, component, dim) = Vq; }

    inline void contractX(const int qx, const int qy, const int localElement, sharedBufferArray sharedBuffer, quadToDof B, const CeedScalar U,
                          CeedScalar *V) {
      sharedBuffer(qx, qy, localElement) = U;
      *V                                 = 0.0;
      @barrier();
      for (int p = 0; p < P1D; ++p) {
        *V += B(p, qx) * sharedBuffer(p, qy, localElement);
      }
      @barrier();
    }

    inline void contractY(const int qx, const int qy, const int localElement, sharedBufferArray sharedBuffer, quadToDof B, const CeedScalar U,
                          CeedScalar *V) {
      sharedBuffer(qx, qy, localElement) = U;
      *V                                 = 0.0;
      @barrier();
      for (int p = 0; p < P1D; ++p) {
        *V += B(p, qy) * sharedBuffer(qx, p, localElement);
      }
      @barrier();
    }

    inline void contractTransposeX(const int px, const int py, const int localElement, sharedBufferArray sharedBuffer, quadToDof B,
                                   const CeedScalar U, CeedScalar *V) {
      sharedBuffer(px, py, localElement) = U;
      *V                                 = 0.0;
      @barrier();
      for (int q = 0; q < Q1D; ++q) {
        *V += B(px, q) * sharedBuffer(q, py, localElement);
      }
      @barrier();
    }

    inline void contractTransposeY(const int px, const int py, const int localElement, sharedBufferArray sharedBuffer, quadToDof B,
                                   const CeedScalar U, CeedScalar *V) {
      sharedBuffer(px, py, localElement) = U;
      *V                                 = 0.0;
      @barrier();
      for (int q = 0; q < Q1D; ++q) {
        *V += B(py, q) * sharedBuffer(px, q, localElement);
      }
      @barrier();
    }

    //---[ Kernels ]------------------------
    @kernel void interp(const CeedInt elementCount, quadToDof B, const CeedScalar *U, CeedScalar *V) {
      for (int elementOffset = 0; elementOffset < elementCount; elementOffset += ELEMENTS_PER_BLOCK; @outer) {
        @shared CeedScalar sharedBuffer[MAX_PQ * MAX_PQ * ELEMENTS_PER_BLOCK];

        for (int localElement = 0; localElement < ELEMENTS_PER_BLOCK; ++localElement; @inner) {
          const int element = elementOffset + localElement;
          for (int qy = 0; qy < Q1D; ++qy; @inner) {
            for (int qx = 0; qx < Q1D; ++qx; @inner) {
              if (element < elementCount) {
                for (int component = 0; component < BASIS_COMPONENT_COUNT; ++component) {
                  CeedScalar r1, r2;
                  if (!TRANSPOSE) {
                    readDofs(element, component, qx, qy, U, &r1);
                    contractX(qx, qy, localElement, sharedBuffer, B, r1, &r2);
                    contractY(qx, qy, localElement, sharedBuffer, B, r2, &r1);
                    writeQuads(elementCount, element, component, qx, qy, 0, r1, V);
                  } else {
                    readQuads(elementCount, element, component, qx, qy, 0, U, &r1);
                    contractTransposeY(qx, qy, localElement, sharedBuffer, B, r1, &r2);
                    contractTransposeX(qx, qy, localElement, sharedBuffer, B, r2, &r1);
                    writeDofs(element, component, qx, qy, r1, V);
                  }
                }
              }
            }
          }
        }
      }
    }

    @kernel void grad(const CeedInt elementCount, quadToDof B, quadToDof Bx, const CeedScalar *U, CeedScalar *V) {
      for (int elementOffset = 0; elementOffset < elementCount; elementOffset += ELEMENTS_PER_BLOCK; @outer) {
        @shared CeedScalar sharedBuffer[MAX_PQ * MAX_PQ * ELEMENTS_PER_BLOCK];

        for (int localElement = 0; localElement < ELEMENTS_PER_BLOCK; ++localElement; @inner) {
          const int element = elementOffset + localElement;
          for (int qy = 0; qy < Q1D; ++qy; @inner) {
            for (int qx = 0; qx < Q1D; ++qx; @inner) {
              if (element < elementCount) {
                for (int component = 0; component < BASIS_COMPONENT_COUNT; ++component) {
                  CeedScalar r1, r2, r3;
                  if (!TRANSPOSE) {
                    readDofs(element, component, qx, qy, U, &r1);
                    contractX(qx, qy, localElement, sharedBuffer, Bx, r1, &r2);
                    contractY(qx, qy, localElement, sharedBuffer, B, r2, &r3);
                    writeQuads(elementCount, element, component, qx, qy, 0, r3, V);
                    contractX(qx, qy, localElement, sharedBuffer, B, r1, &r2);
                    contractY(qx, qy, localElement, sharedBuffer, Bx, r2, &r3);
                    writeQuads(elementCount, element, component, qx, qy, 1, r3, V);
                  } else {
                    readQuads(elementCount, element, component, qx, qy, 0, U, &r1);
                    contractTransposeY(qx, qy, localElement, sharedBuffer, B, r1, &r2);
                    contractTransposeX(qx, qy, localElement, sharedBuffer, Bx, r2, &r3);
                    readQuads(elementCount, element, component, qx, qy, 1, U, &r1);
                    contractTransposeY(qx, qy, localElement, sharedBuffer, Bx, r1, &r2);
                    contractTransposeX(qx, qy, localElement, sharedBuffer, B, r2, &r1);
                    writeDofs(element, component, qx, qy, r1 + r3, V);
                  }
                }
              }
            }
          }
        }
      }
    }

    @kernel void weight(const CeedInt elementCount, const CeedScalar *qWeights1D, elementWeightArray W) {
      for (int elementOffset = 0; elementOffset < elementCount; elementOffset += ELEMENTS_PER_BLOCK; @outer) {
        for (int element = elementOffset; element < (elementOffset + ELEMENTS_PER_BLOCK); ++element; @outer) {
          for (int qy = 0; qy < Q1D; ++qy; @inner) {
            for (int qx = 0; qx < Q1D; ++qx; @inner) {
              W(qx, qy, element) = qWeights1D[qx] * qWeights1D[qy];
            }
          }
        }
      }
    }

);