Actual source code: vpbjacobi_cuda.cu
1: #include <petscdevice_cuda.h>
2: #include <../src/ksp/pc/impls/vpbjacobi/vpbjacobi.h>
4: /* A class that manages helper arrays assisting parallel PCApply() with CUDA */
5: struct PC_VPBJacobi_CUDA {
6: /* Cache the old sizes to check if we need realloc */
7: PetscInt n; /* number of rows of the local matrix */
8: PetscInt nblocks; /* number of point blocks */
9: PetscInt nsize; /* sum of sizes of the point blocks */
11: /* Helper arrays that are pre-computed on host and then copied to device.
12: bs: [nblocks+1], "csr" version of bsizes[], with bs[0] = 0, bs[nblocks] = n.
13: bs2: [nblocks+1], "csr" version of squares of bsizes[], with bs2[0] = 0, bs2[nblocks] = nsize.
14: matIdx: [n], row i of the local matrix belongs to the matIdx_d[i] block
15: */
16: PetscInt *bs_h, *bs2_h, *matIdx_h;
17: PetscInt *bs_d, *bs2_d, *matIdx_d;
19: MatScalar *diag_d; /* [nsize], store inverse of the point blocks on device */
21: PC_VPBJacobi_CUDA(PetscInt n, PetscInt nblocks, PetscInt nsize, const PetscInt *bsizes, MatScalar *diag_h) : n(n), nblocks(nblocks), nsize(nsize)
22: {
23: /* malloc memory on host and device, and then update */
24: PetscCallVoid(PetscMalloc3(nblocks + 1, &bs_h, nblocks + 1, &bs2_h, n, &matIdx_h));
25: PetscCallCUDAVoid(cudaMalloc(&bs_d, sizeof(PetscInt) * (nblocks + 1)));
26: PetscCallCUDAVoid(cudaMalloc(&bs2_d, sizeof(PetscInt) * (nblocks + 1)));
27: PetscCallCUDAVoid(cudaMalloc(&matIdx_d, sizeof(PetscInt) * n));
28: PetscCallCUDAVoid(cudaMalloc(&diag_d, sizeof(MatScalar) * nsize));
29: PetscCallVoid(UpdateOffsetsOnDevice(bsizes, diag_h));
30: }
32: PetscErrorCode UpdateOffsetsOnDevice(const PetscInt *bsizes, MatScalar *diag_h)
33: {
34: PetscFunctionBegin;
35: PetscCall(ComputeOffsetsOnHost(bsizes));
36: PetscCallCUDA(cudaMemcpy(bs_d, bs_h, sizeof(PetscInt) * (nblocks + 1), cudaMemcpyHostToDevice));
37: PetscCallCUDA(cudaMemcpy(bs2_d, bs2_h, sizeof(PetscInt) * (nblocks + 1), cudaMemcpyHostToDevice));
38: PetscCallCUDA(cudaMemcpy(matIdx_d, matIdx_h, sizeof(PetscInt) * n, cudaMemcpyHostToDevice));
39: PetscCallCUDA(cudaMemcpy(diag_d, diag_h, sizeof(MatScalar) * nsize, cudaMemcpyHostToDevice));
40: PetscCall(PetscLogCpuToGpu(sizeof(PetscInt) * (2 * nblocks + 2 + n) + sizeof(MatScalar) * nsize));
41: PetscFunctionReturn(PETSC_SUCCESS);
42: }
44: ~PC_VPBJacobi_CUDA()
45: {
46: PetscCallVoid(PetscFree3(bs_h, bs2_h, matIdx_h));
47: PetscCallCUDAVoid(cudaFree(bs_d));
48: PetscCallCUDAVoid(cudaFree(bs2_d));
49: PetscCallCUDAVoid(cudaFree(matIdx_d));
50: PetscCallCUDAVoid(cudaFree(diag_d));
51: }
53: private:
54: PetscErrorCode ComputeOffsetsOnHost(const PetscInt *bsizes)
55: {
56: PetscFunctionBegin;
57: bs_h[0] = bs2_h[0] = 0;
58: for (PetscInt i = 0; i < nblocks; i++) {
59: bs_h[i + 1] = bs_h[i] + bsizes[i];
60: bs2_h[i + 1] = bs2_h[i] + bsizes[i] * bsizes[i];
61: for (PetscInt j = 0; j < bsizes[i]; j++) matIdx_h[bs_h[i] + j] = i;
62: }
63: PetscFunctionReturn(PETSC_SUCCESS);
64: }
65: };
67: /* Like cublasDgemvBatched() but with variable-sized blocks
69: Input Parameters:
70: + n - number of rows of the local matrix
71: . bs - [nblocks+1], prefix sum of bsizes[]
72: . bs2 - [nblocks+1], prefix sum of squares of bsizes[]
73: . matIdx - [n], store block/matrix index for each row
74: . A - blocks of the matrix back to back in column-major order
75: . x - the input vector
76: - transpose - whether it is MatMult for Ax (false) or MatMultTranspose for A^Tx (true)
78: Output Parameter:
79: . y - the output vector
80: */
81: __global__ static void MatMultBatched(PetscInt n, const PetscInt *bs, const PetscInt *bs2, const PetscInt *matIdx, const MatScalar *A, const PetscScalar *x, PetscScalar *y, PetscBool transpose)
82: {
83: const PetscInt gridSize = gridDim.x * blockDim.x;
84: PetscInt tid = blockIdx.x * blockDim.x + threadIdx.x;
85: PetscInt i, j, k, m;
87: /* One row per thread. The blocks/matrices are stored in column-major order */
88: for (; tid < n; tid += gridSize) {
89: k = matIdx[tid]; /* k-th block */
90: m = bs[k + 1] - bs[k]; /* block size of the k-th block */
91: i = tid - bs[k]; /* i-th row of the block */
92: A += bs2[k] + i * (transpose ? m : 1); /* advance A to the first entry of i-th row */
93: x += bs[k];
94: y += bs[k];
96: y[i] = 0.0;
97: for (j = 0; j < m; j++) {
98: y[i] += A[0] * x[j];
99: A += (transpose ? 1 : m);
100: }
101: }
102: }
104: static PetscErrorCode PCApplyOrTranspose_VPBJacobi_CUDA(PC pc, Vec x, Vec y, PetscBool transpose)
105: {
106: PC_VPBJacobi *jac = (PC_VPBJacobi *)pc->data;
107: PC_VPBJacobi_CUDA *pcuda = static_cast<PC_VPBJacobi_CUDA *>(jac->spptr);
108: const PetscScalar *xx;
109: PetscScalar *yy;
110: PetscInt n;
112: PetscFunctionBegin;
113: PetscCall(PetscLogGpuTimeBegin());
114: if (PetscDefined(USE_DEBUG)) {
115: PetscBool isCuda;
116: PetscCall(PetscObjectTypeCompareAny((PetscObject)x, &isCuda, VECSEQCUDA, VECMPICUDA, ""));
117: if (isCuda) PetscCall(PetscObjectTypeCompareAny((PetscObject)y, &isCuda, VECSEQCUDA, VECMPICUDA, ""));
118: PetscCheck(isCuda, PETSC_COMM_SELF, PETSC_ERR_SUP, "PC: applying a CUDA pmat to non-cuda vectors");
119: }
121: PetscCall(MatGetLocalSize(pc->pmat, &n, NULL));
122: if (n) {
123: PetscInt gridSize = PetscMin((n + 255) / 256, 2147483647); /* <= 2^31-1 */
124: PetscCall(VecCUDAGetArrayRead(x, &xx));
125: PetscCall(VecCUDAGetArrayWrite(y, &yy));
126: MatMultBatched<<<gridSize, 256>>>(n, pcuda->bs_d, pcuda->bs2_d, pcuda->matIdx_d, pcuda->diag_d, xx, yy, transpose);
127: PetscCallCUDA(cudaGetLastError());
128: PetscCall(VecCUDARestoreArrayRead(x, &xx));
129: PetscCall(VecCUDARestoreArrayWrite(y, &yy));
130: }
131: PetscCall(PetscLogGpuFlops(pcuda->nsize * 2)); /* FMA on entries in all blocks */
132: PetscCall(PetscLogGpuTimeEnd());
133: PetscFunctionReturn(PETSC_SUCCESS);
134: }
136: static PetscErrorCode PCApply_VPBJacobi_CUDA(PC pc, Vec x, Vec y)
137: {
138: PetscFunctionBegin;
139: PetscCall(PCApplyOrTranspose_VPBJacobi_CUDA(pc, x, y, PETSC_FALSE));
140: PetscFunctionReturn(PETSC_SUCCESS);
141: }
143: static PetscErrorCode PCApplyTranspose_VPBJacobi_CUDA(PC pc, Vec x, Vec y)
144: {
145: PetscFunctionBegin;
146: PetscCall(PCApplyOrTranspose_VPBJacobi_CUDA(pc, x, y, PETSC_TRUE));
147: PetscFunctionReturn(PETSC_SUCCESS);
148: }
150: static PetscErrorCode PCDestroy_VPBJacobi_CUDA(PC pc)
151: {
152: PC_VPBJacobi *jac = (PC_VPBJacobi *)pc->data;
154: PetscFunctionBegin;
155: PetscCallCXX(delete static_cast<PC_VPBJacobi_CUDA *>(jac->spptr));
156: PetscCall(PCDestroy_VPBJacobi(pc));
157: PetscFunctionReturn(PETSC_SUCCESS);
158: }
160: PETSC_INTERN PetscErrorCode PCSetUp_VPBJacobi_CUDA(PC pc)
161: {
162: PC_VPBJacobi *jac = (PC_VPBJacobi *)pc->data;
163: PC_VPBJacobi_CUDA *pcuda = static_cast<PC_VPBJacobi_CUDA *>(jac->spptr);
164: PetscInt i, n, nblocks, nsize = 0;
165: const PetscInt *bsizes;
167: PetscFunctionBegin;
168: PetscCall(PCSetUp_VPBJacobi_Host(pc)); /* Compute the inverse on host now. Might worth doing it on device directly */
169: PetscCall(MatGetVariableBlockSizes(pc->pmat, &nblocks, &bsizes));
170: for (i = 0; i < nblocks; i++) nsize += bsizes[i] * bsizes[i];
171: PetscCall(MatGetLocalSize(pc->pmat, &n, NULL));
173: /* If one calls MatSetVariableBlockSizes() multiple times and sizes have been changed (is it allowed?), we delete the old and rebuild anyway */
174: if (pcuda && (pcuda->n != n || pcuda->nblocks != nblocks || pcuda->nsize != nsize)) {
175: PetscCallCXX(delete pcuda);
176: pcuda = nullptr;
177: }
179: if (!pcuda) { /* allocate the struct along with the helper arrays from the scratch */
180: PetscCallCXX(jac->spptr = new PC_VPBJacobi_CUDA(n, nblocks, nsize, bsizes, jac->diag));
181: } else { /* update the value only */
182: PetscCall(pcuda->UpdateOffsetsOnDevice(bsizes, jac->diag));
183: }
185: pc->ops->apply = PCApply_VPBJacobi_CUDA;
186: pc->ops->applytranspose = PCApplyTranspose_VPBJacobi_CUDA;
187: pc->ops->destroy = PCDestroy_VPBJacobi_CUDA;
188: PetscFunctionReturn(PETSC_SUCCESS);
189: }