Actual source code: mpiviennacl.cxx


  2: /*
  3:    This file contains routines for Parallel vector operations.
  4:  */
  5: #include <petscconf.h>
  6: #include <../src/vec/vec/impls/mpi/pvecimpl.h>
  7: #include <../src/vec/vec/impls/seq/seqviennacl/viennaclvecimpl.h>

  9: /*MC
 10:    VECVIENNACL - VECVIENNACL = "viennacl" - A VECSEQVIENNACL on a single-process communicator, and VECMPIVIENNACL otherwise.

 12:    Options Database Keys:
 13: . -vec_type viennacl - sets the vector type to VECVIENNACL during a call to VecSetFromOptions()

 15:   Level: beginner

 17: .seealso: VecCreate(), VecSetType(), VecSetFromOptions(), VecCreateMPIWithArray(), VECSEQVIENNACL, VECMPIVIENNACL, VECSTANDARD, VecType, VecCreateMPI(), VecCreateMPI()
 18: M*/

 20: PetscErrorCode VecDestroy_MPIViennaCL(Vec v)
 21: {
 22:   try {
 23:     if (v->spptr) {
 24:       delete ((Vec_ViennaCL*)v->spptr)->GPUarray_allocated;
 25:       delete (Vec_ViennaCL*) v->spptr;
 26:     }
 27:   } catch(std::exception const & ex) {
 28:     SETERRQ(PETSC_COMM_SELF,PETSC_ERR_LIB,"ViennaCL error: %s", ex.what());
 29:   }
 30:   VecDestroy_MPI(v);
 31:   return 0;
 32: }

 34: PetscErrorCode VecNorm_MPIViennaCL(Vec xin,NormType type,PetscReal *z)
 35: {
 36:   PetscReal sum,work = 0.0;

 38:   if (type == NORM_2 || type == NORM_FROBENIUS) {
 39:     VecNorm_SeqViennaCL(xin,NORM_2,&work);
 40:     work *= work;
 41:     MPIU_Allreduce(&work,&sum,1,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 42:     *z    = PetscSqrtReal(sum);
 43:   } else if (type == NORM_1) {
 44:     /* Find the local part */
 45:     VecNorm_SeqViennaCL(xin,NORM_1,&work);
 46:     /* Find the global max */
 47:     MPIU_Allreduce(&work,z,1,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 48:   } else if (type == NORM_INFINITY) {
 49:     /* Find the local max */
 50:     VecNorm_SeqViennaCL(xin,NORM_INFINITY,&work);
 51:     /* Find the global max */
 52:     MPIU_Allreduce(&work,z,1,MPIU_REAL,MPIU_MAX,PetscObjectComm((PetscObject)xin));
 53:   } else if (type == NORM_1_AND_2) {
 54:     PetscReal temp[2];
 55:     VecNorm_SeqViennaCL(xin,NORM_1,temp);
 56:     VecNorm_SeqViennaCL(xin,NORM_2,temp+1);
 57:     temp[1] = temp[1]*temp[1];
 58:     MPIU_Allreduce(temp,z,2,MPIU_REAL,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 59:     z[1] = PetscSqrtReal(z[1]);
 60:   }
 61:   return 0;
 62: }

 64: PetscErrorCode VecDot_MPIViennaCL(Vec xin,Vec yin,PetscScalar *z)
 65: {
 66:   PetscScalar    sum,work;

 68:   VecDot_SeqViennaCL(xin,yin,&work);
 69:   MPIU_Allreduce(&work,&sum,1,MPIU_SCALAR,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 70:   *z   = sum;
 71:   return 0;
 72: }

 74: PetscErrorCode VecTDot_MPIViennaCL(Vec xin,Vec yin,PetscScalar *z)
 75: {
 76:   PetscScalar    sum,work;

 78:   VecTDot_SeqViennaCL(xin,yin,&work);
 79:   MPIU_Allreduce(&work,&sum,1,MPIU_SCALAR,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 80:   *z   = sum;
 81:   return 0;
 82: }

 84: PetscErrorCode VecMDot_MPIViennaCL(Vec xin,PetscInt nv,const Vec y[],PetscScalar *z)
 85: {
 86:   PetscScalar    awork[128],*work = awork;

 88:   if (nv > 128) {
 89:     PetscMalloc1(nv,&work);
 90:   }
 91:   VecMDot_SeqViennaCL(xin,nv,y,work);
 92:   MPIU_Allreduce(work,z,nv,MPIU_SCALAR,MPIU_SUM,PetscObjectComm((PetscObject)xin));
 93:   if (nv > 128) {
 94:     PetscFree(work);
 95:   }
 96:   return 0;
 97: }

 99: /*MC
100:    VECMPIVIENNACL - VECMPIVIENNACL = "mpiviennacl" - The basic parallel vector, modified to use ViennaCL

102:    Options Database Keys:
103: . -vec_type mpiviennacl - sets the vector type to VECMPIVIENNACL during a call to VecSetFromOptions()

105:   Level: beginner

107: .seealso: VecCreate(), VecSetType(), VecSetFromOptions(), VecCreateMPIWithArray(), VECMPI, VecType, VecCreateMPI(), VecCreateMPI()
108: M*/

110: PetscErrorCode VecDuplicate_MPIViennaCL(Vec win,Vec *v)
111: {
112:   Vec_MPI        *vw,*w = (Vec_MPI*)win->data;
113:   PetscScalar    *array;

115:   VecCreate(PetscObjectComm((PetscObject)win),v);
116:   PetscLayoutReference(win->map,&(*v)->map);

118:   VecCreate_MPI_Private(*v,PETSC_FALSE,w->nghost,0);
119:   vw   = (Vec_MPI*)(*v)->data;
120:   PetscMemcpy((*v)->ops,win->ops,sizeof(struct _VecOps));

122:   /* save local representation of the parallel vector (and scatter) if it exists */
123:   if (w->localrep) {
124:     VecGetArray(*v,&array);
125:     VecCreateSeqWithArray(PETSC_COMM_SELF,1,win->map->n+w->nghost,array,&vw->localrep);
126:     PetscMemcpy(vw->localrep->ops,w->localrep->ops,sizeof(struct _VecOps));
127:     VecRestoreArray(*v,&array);
128:     PetscLogObjectParent((PetscObject)*v,(PetscObject)vw->localrep);
129:     vw->localupdate = w->localupdate;
130:     if (vw->localupdate) {
131:       PetscObjectReference((PetscObject)vw->localupdate);
132:     }
133:   }

135:   /* New vector should inherit stashing property of parent */
136:   (*v)->stash.donotstash   = win->stash.donotstash;
137:   (*v)->stash.ignorenegidx = win->stash.ignorenegidx;

139:   /* change type_name appropriately */
140:   PetscObjectChangeTypeName((PetscObject)(*v),VECMPIVIENNACL);

142:   PetscObjectListDuplicate(((PetscObject)win)->olist,&((PetscObject)(*v))->olist);
143:   PetscFunctionListDuplicate(((PetscObject)win)->qlist,&((PetscObject)(*v))->qlist);
144:   (*v)->map->bs   = PetscAbs(win->map->bs);
145:   (*v)->bstash.bs = win->bstash.bs;
146:   return 0;
147: }

149: PetscErrorCode VecDotNorm2_MPIViennaCL(Vec s,Vec t,PetscScalar *dp,PetscScalar *nm)
150: {
151:   PetscScalar    work[2],sum[2];

153:   VecDotNorm2_SeqViennaCL(s,t,work,work+1);
154:   MPIU_Allreduce((void*)&work,(void*)&sum,2,MPIU_SCALAR,MPIU_SUM,PetscObjectComm((PetscObject)s));
155:   *dp  = sum[0];
156:   *nm  = sum[1];
157:   return 0;
158: }

160: PetscErrorCode VecBindToCPU_MPIViennaCL(Vec vv, PetscBool bind)
161: {
162:   vv->boundtocpu = bind;

164:   if (bind) {
165:     VecViennaCLCopyFromGPU(vv);
166:     vv->offloadmask = PETSC_OFFLOAD_CPU; /* since the CPU code will likely change values in the vector */
167:     vv->ops->dotnorm2               = NULL;
168:     vv->ops->waxpy                  = VecWAXPY_Seq;
169:     vv->ops->dot                    = VecDot_MPI;
170:     vv->ops->mdot                   = VecMDot_MPI;
171:     vv->ops->tdot                   = VecTDot_MPI;
172:     vv->ops->norm                   = VecNorm_MPI;
173:     vv->ops->scale                  = VecScale_Seq;
174:     vv->ops->copy                   = VecCopy_Seq;
175:     vv->ops->set                    = VecSet_Seq;
176:     vv->ops->swap                   = VecSwap_Seq;
177:     vv->ops->axpy                   = VecAXPY_Seq;
178:     vv->ops->axpby                  = VecAXPBY_Seq;
179:     vv->ops->maxpy                  = VecMAXPY_Seq;
180:     vv->ops->aypx                   = VecAYPX_Seq;
181:     vv->ops->axpbypcz               = VecAXPBYPCZ_Seq;
182:     vv->ops->pointwisemult          = VecPointwiseMult_Seq;
183:     vv->ops->setrandom              = VecSetRandom_Seq;
184:     vv->ops->placearray             = VecPlaceArray_Seq;
185:     vv->ops->replacearray           = VecReplaceArray_Seq;
186:     vv->ops->resetarray             = VecResetArray_Seq;
187:     vv->ops->dot_local              = VecDot_Seq;
188:     vv->ops->tdot_local             = VecTDot_Seq;
189:     vv->ops->norm_local             = VecNorm_Seq;
190:     vv->ops->mdot_local             = VecMDot_Seq;
191:     vv->ops->pointwisedivide        = VecPointwiseDivide_Seq;
192:     vv->ops->getlocalvector         = NULL;
193:     vv->ops->restorelocalvector     = NULL;
194:     vv->ops->getlocalvectorread     = NULL;
195:     vv->ops->restorelocalvectorread = NULL;
196:     vv->ops->getarraywrite          = NULL;
197:   } else {
198:     vv->ops->dotnorm2        = VecDotNorm2_MPIViennaCL;
199:     vv->ops->waxpy           = VecWAXPY_SeqViennaCL;
200:     vv->ops->duplicate       = VecDuplicate_MPIViennaCL;
201:     vv->ops->dot             = VecDot_MPIViennaCL;
202:     vv->ops->mdot            = VecMDot_MPIViennaCL;
203:     vv->ops->tdot            = VecTDot_MPIViennaCL;
204:     vv->ops->norm            = VecNorm_MPIViennaCL;
205:     vv->ops->scale           = VecScale_SeqViennaCL;
206:     vv->ops->copy            = VecCopy_SeqViennaCL;
207:     vv->ops->set             = VecSet_SeqViennaCL;
208:     vv->ops->swap            = VecSwap_SeqViennaCL;
209:     vv->ops->axpy            = VecAXPY_SeqViennaCL;
210:     vv->ops->axpby           = VecAXPBY_SeqViennaCL;
211:     vv->ops->maxpy           = VecMAXPY_SeqViennaCL;
212:     vv->ops->aypx            = VecAYPX_SeqViennaCL;
213:     vv->ops->axpbypcz        = VecAXPBYPCZ_SeqViennaCL;
214:     vv->ops->pointwisemult   = VecPointwiseMult_SeqViennaCL;
215:     vv->ops->setrandom       = VecSetRandom_SeqViennaCL;
216:     vv->ops->dot_local       = VecDot_SeqViennaCL;
217:     vv->ops->tdot_local      = VecTDot_SeqViennaCL;
218:     vv->ops->norm_local      = VecNorm_SeqViennaCL;
219:     vv->ops->mdot_local      = VecMDot_SeqViennaCL;
220:     vv->ops->destroy         = VecDestroy_MPIViennaCL;
221:     vv->ops->pointwisedivide = VecPointwiseDivide_SeqViennaCL;
222:     vv->ops->placearray      = VecPlaceArray_SeqViennaCL;
223:     vv->ops->replacearray    = VecReplaceArray_SeqViennaCL;
224:     vv->ops->resetarray      = VecResetArray_SeqViennaCL;
225:     vv->ops->getarraywrite   = VecGetArrayWrite_SeqViennaCL;
226:     vv->ops->getarray        = VecGetArray_SeqViennaCL;
227:     vv->ops->restorearray    = VecRestoreArray_SeqViennaCL;
228:   }
229:   return 0;
230: }

232: PETSC_EXTERN PetscErrorCode VecCreate_MPIViennaCL(Vec vv)
233: {
234:   PetscLayoutSetUp(vv->map);
235:   VecViennaCLAllocateCheck(vv);
236:   VecCreate_MPIViennaCL_Private(vv,PETSC_FALSE,0,((Vec_ViennaCL*)(vv->spptr))->GPUarray);
237:   VecViennaCLAllocateCheckHost(vv);
238:   VecSet(vv,0.0);
239:   VecSet_Seq(vv,0.0);
240:   vv->offloadmask = PETSC_OFFLOAD_BOTH;
241:   return 0;
242: }

244: PETSC_EXTERN PetscErrorCode VecCreate_ViennaCL(Vec v)
245: {
246:   PetscMPIInt    size;

248:   MPI_Comm_size(PetscObjectComm((PetscObject)v),&size);
249:   if (size == 1) {
250:     VecSetType(v,VECSEQVIENNACL);
251:   } else {
252:     VecSetType(v,VECMPIVIENNACL);
253:   }
254:   return 0;
255: }

257: /*@C
258:    VecCreateMPIViennaCLWithArray - Creates a parallel, array-style vector,
259:    where the user provides the viennacl vector to store the vector values.

261:    Collective

263:    Input Parameters:
264: +  comm  - the MPI communicator to use
265: .  bs    - block size, same meaning as VecSetBlockSize()
266: .  n     - local vector length, cannot be PETSC_DECIDE
267: .  N     - global vector length (or PETSC_DECIDE to have calculated)
268: -  array - the user provided GPU array to store the vector values

270:    Output Parameter:
271: .  vv - the vector

273:    Notes:
274:    Use VecDuplicate() or VecDuplicateVecs() to form additional vectors of the
275:    same type as an existing vector.

277:    If the user-provided array is NULL, then VecViennaCLPlaceArray() can be used
278:    at a later stage to SET the array for storing the vector values.

280:    PETSc does NOT free the array when the vector is destroyed via VecDestroy().
281:    The user should not free the array until the vector is destroyed.

283:    Level: intermediate

285: .seealso: VecCreateSeqViennaCLWithArray(), VecCreateMPIWithArray(), VecCreateSeqWithArray(),
286:           VecCreate(), VecCreateMPI(), VecCreateGhostWithArray(), VecViennaCLPlaceArray()

288: @*/
289: PetscErrorCode  VecCreateMPIViennaCLWithArray(MPI_Comm comm,PetscInt bs,PetscInt n,PetscInt N,const ViennaCLVector *array,Vec *vv)
290: {
292:   PetscSplitOwnership(comm,&n,&N);
293:   VecCreate(comm,vv);
294:   VecSetSizes(*vv,n,N);
295:   VecSetBlockSize(*vv,bs);
296:   VecCreate_MPIViennaCL_Private(*vv,PETSC_FALSE,0,array);
297:   return 0;
298: }

300: /*@C
301:    VecCreateMPIViennaCLWithArrays - Creates a parallel, array-style vector,
302:    where the user provides the ViennaCL vector to store the vector values.

304:    Collective

306:    Input Parameters:
307: +  comm  - the MPI communicator to use
308: .  bs    - block size, same meaning as VecSetBlockSize()
309: .  n     - local vector length, cannot be PETSC_DECIDE
310: .  N     - global vector length (or PETSC_DECIDE to have calculated)
311: -  cpuarray - the user provided CPU array to store the vector values
312: -  viennaclvec - ViennaCL vector where the Vec entries are to be stored on the device.

314:    Output Parameter:
315: .  vv - the vector

317:    Notes:
318:    If both cpuarray and viennaclvec are provided, the caller must ensure that
319:    the provided arrays have identical values.

321:    Use VecDuplicate() or VecDuplicateVecs() to form additional vectors of the
322:    same type as an existing vector.

324:    PETSc does NOT free the provided arrays when the vector is destroyed via
325:    VecDestroy(). The user should not free the array until the vector is
326:    destroyed.

328:    Level: intermediate

330: .seealso: VecCreateSeqViennaCLWithArrays(), VecCreateMPIWithArray()
331:           VecCreate(), VecDuplicate(), VecDuplicateVecs(), VecCreateGhost(),
332:           VecCreateMPI(), VecCreateGhostWithArray(), VecViennaCLPlaceArray(),
333:           VecPlaceArray(), VecCreateMPICUDAWithArrays(),
334:           VecViennaCLAllocateCheckHost()
335: @*/
336: PetscErrorCode  VecCreateMPIViennaCLWithArrays(MPI_Comm comm,PetscInt bs,PetscInt n,PetscInt N,const PetscScalar cpuarray[],const ViennaCLVector *viennaclvec,Vec *vv)
337: {
338:   VecCreateMPIViennaCLWithArray(comm,bs,n,N,viennaclvec,vv);
339:   if (cpuarray && viennaclvec) {
340:     Vec_MPI *s         = (Vec_MPI*)((*vv)->data);
341:     s->array           = (PetscScalar*)cpuarray;
342:     (*vv)->offloadmask = PETSC_OFFLOAD_BOTH;
343:   } else if (cpuarray) {
344:     Vec_MPI *s         = (Vec_MPI*)((*vv)->data);
345:     s->array           = (PetscScalar*)cpuarray;
346:     (*vv)->offloadmask =  PETSC_OFFLOAD_CPU;
347:   } else if (viennaclvec) {
348:     (*vv)->offloadmask = PETSC_OFFLOAD_GPU;
349:   } else {
350:     (*vv)->offloadmask = PETSC_OFFLOAD_UNALLOCATED;
351:   }
352:   return 0;
353: }

355: PetscErrorCode VecCreate_MPIViennaCL_Private(Vec vv,PetscBool alloc,PetscInt nghost,const ViennaCLVector *array)
356: {
357:   Vec_ViennaCL   *vecviennacl;

359:   VecCreate_MPI_Private(vv,PETSC_FALSE,0,0);
360:   PetscObjectChangeTypeName((PetscObject)vv,VECMPIVIENNACL);

362:   VecBindToCPU_MPIViennaCL(vv,PETSC_FALSE);
363:   vv->ops->bindtocpu = VecBindToCPU_MPIViennaCL;

365:   if (alloc && !array) {
366:     VecViennaCLAllocateCheck(vv);
367:     VecViennaCLAllocateCheckHost(vv);
368:     VecSet(vv,0.0);
369:     VecSet_Seq(vv,0.0);
370:     vv->offloadmask = PETSC_OFFLOAD_BOTH;
371:   }
372:   if (array) {
373:     if (!vv->spptr)
374:       vv->spptr = new Vec_ViennaCL;
375:     vecviennacl = (Vec_ViennaCL*)vv->spptr;
376:     vecviennacl->GPUarray_allocated = 0;
377:     vecviennacl->GPUarray = (ViennaCLVector*)array;
378:     vv->offloadmask = PETSC_OFFLOAD_UNALLOCATED;
379:   }

381:   return 0;
382: }