Line data Source code
1 : /*
2 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
3 : SLEPc - Scalable Library for Eigenvalue Problem Computations
4 : Copyright (c) 2002-, Universitat Politecnica de Valencia, Spain
5 :
6 : This file is part of SLEPc.
7 : SLEPc is distributed under a 2-clause BSD license (see LICENSE).
8 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
9 : */
10 : /*
11 : Square root function sqrt(x)
12 : */
13 :
14 : #include <slepc/private/fnimpl.h> /*I "slepcfn.h" I*/
15 : #include <slepcblaslapack.h>
16 :
17 952 : static PetscErrorCode FNEvaluateFunction_Sqrt(FN fn,PetscScalar x,PetscScalar *y)
18 : {
19 952 : PetscFunctionBegin;
20 : #if !defined(PETSC_USE_COMPLEX)
21 952 : PetscCheck(x>=0.0,PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Function not defined in the requested value");
22 : #endif
23 952 : *y = PetscSqrtScalar(x);
24 952 : PetscFunctionReturn(PETSC_SUCCESS);
25 : }
26 :
27 20 : static PetscErrorCode FNEvaluateDerivative_Sqrt(FN fn,PetscScalar x,PetscScalar *y)
28 : {
29 20 : PetscFunctionBegin;
30 20 : PetscCheck(x!=0.0,PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Derivative not defined in the requested value");
31 : #if !defined(PETSC_USE_COMPLEX)
32 20 : PetscCheck(x>0.0,PETSC_COMM_SELF,PETSC_ERR_ARG_OUTOFRANGE,"Derivative not defined in the requested value");
33 : #endif
34 20 : *y = 1.0/(2.0*PetscSqrtScalar(x));
35 20 : PetscFunctionReturn(PETSC_SUCCESS);
36 : }
37 :
38 19 : static PetscErrorCode FNEvaluateFunctionMat_Sqrt_Schur(FN fn,Mat A,Mat B)
39 : {
40 19 : PetscBLASInt n=0;
41 19 : PetscScalar *T;
42 19 : PetscInt m;
43 :
44 19 : PetscFunctionBegin;
45 19 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
46 19 : PetscCall(MatDenseGetArray(B,&T));
47 19 : PetscCall(MatGetSize(A,&m,NULL));
48 19 : PetscCall(PetscBLASIntCast(m,&n));
49 19 : PetscCall(FNSqrtmSchur(fn,n,T,n,PETSC_FALSE));
50 19 : PetscCall(MatDenseRestoreArray(B,&T));
51 19 : PetscFunctionReturn(PETSC_SUCCESS);
52 : }
53 :
54 14 : static PetscErrorCode FNEvaluateFunctionMatVec_Sqrt_Schur(FN fn,Mat A,Vec v)
55 : {
56 14 : PetscBLASInt n=0;
57 14 : PetscScalar *T;
58 14 : PetscInt m;
59 14 : Mat B;
60 :
61 14 : PetscFunctionBegin;
62 14 : PetscCall(FN_AllocateWorkMat(fn,A,&B));
63 14 : PetscCall(MatDenseGetArray(B,&T));
64 14 : PetscCall(MatGetSize(A,&m,NULL));
65 14 : PetscCall(PetscBLASIntCast(m,&n));
66 14 : PetscCall(FNSqrtmSchur(fn,n,T,n,PETSC_TRUE));
67 14 : PetscCall(MatDenseRestoreArray(B,&T));
68 14 : PetscCall(MatGetColumnVector(B,v,0));
69 14 : PetscCall(FN_FreeWorkMat(fn,&B));
70 14 : PetscFunctionReturn(PETSC_SUCCESS);
71 : }
72 :
73 12 : static PetscErrorCode FNEvaluateFunctionMat_Sqrt_DBP(FN fn,Mat A,Mat B)
74 : {
75 12 : PetscBLASInt n=0;
76 12 : PetscScalar *T;
77 12 : PetscInt m;
78 :
79 12 : PetscFunctionBegin;
80 12 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
81 12 : PetscCall(MatDenseGetArray(B,&T));
82 12 : PetscCall(MatGetSize(A,&m,NULL));
83 12 : PetscCall(PetscBLASIntCast(m,&n));
84 12 : PetscCall(FNSqrtmDenmanBeavers(fn,n,T,n,PETSC_FALSE));
85 12 : PetscCall(MatDenseRestoreArray(B,&T));
86 12 : PetscFunctionReturn(PETSC_SUCCESS);
87 : }
88 :
89 12 : static PetscErrorCode FNEvaluateFunctionMat_Sqrt_NS(FN fn,Mat A,Mat B)
90 : {
91 12 : PetscBLASInt n=0;
92 12 : PetscScalar *Ba;
93 12 : PetscInt m;
94 :
95 12 : PetscFunctionBegin;
96 12 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
97 12 : PetscCall(MatDenseGetArray(B,&Ba));
98 12 : PetscCall(MatGetSize(A,&m,NULL));
99 12 : PetscCall(PetscBLASIntCast(m,&n));
100 12 : PetscCall(FNSqrtmNewtonSchulz(fn,n,Ba,n,PETSC_FALSE));
101 12 : PetscCall(MatDenseRestoreArray(B,&Ba));
102 12 : PetscFunctionReturn(PETSC_SUCCESS);
103 : }
104 :
105 : #define MAXIT 50
106 :
107 : /*
108 : Computes the principal square root of the matrix A using the
109 : Sadeghi iteration. A is overwritten with sqrtm(A).
110 : */
111 24 : PetscErrorCode FNSqrtmSadeghi(FN fn,PetscBLASInt n,PetscScalar *A,PetscBLASInt ld)
112 : {
113 24 : PetscScalar *M,*M2,*G,*X=A,*work,work1,sqrtnrm;
114 24 : PetscScalar szero=0.0,sone=1.0,smfive=-5.0,s1d16=1.0/16.0;
115 24 : PetscReal tol,Mres=0.0,nrm,rwork[1],done=1.0;
116 24 : PetscInt i,it;
117 24 : PetscBLASInt N,*piv=NULL,info,lwork=0,query=-1,one=1,zero=0;
118 24 : PetscBool converged=PETSC_FALSE;
119 24 : unsigned int ftz;
120 :
121 24 : PetscFunctionBegin;
122 24 : N = n*n;
123 24 : tol = PetscSqrtReal((PetscReal)n)*PETSC_MACHINE_EPSILON/2;
124 24 : PetscCall(SlepcSetFlushToZero(&ftz));
125 :
126 : /* query work size */
127 24 : PetscCallBLAS("LAPACKgetri",LAPACKgetri_(&n,A,&ld,piv,&work1,&query,&info));
128 24 : PetscCall(PetscBLASIntCast((PetscInt)PetscRealPart(work1),&lwork));
129 :
130 24 : PetscCall(PetscMalloc5(N,&M,N,&M2,N,&G,lwork,&work,n,&piv));
131 24 : PetscCall(PetscArraycpy(M,A,N));
132 :
133 : /* scale M */
134 24 : nrm = LAPACKlange_("fro",&n,&n,M,&n,rwork);
135 24 : if (nrm>1.0) {
136 24 : sqrtnrm = PetscSqrtReal(nrm);
137 24 : PetscCallBLAS("LAPACKlascl",LAPACKlascl_("G",&zero,&zero,&nrm,&done,&N,&one,M,&N,&info));
138 24 : SlepcCheckLapackInfo("lascl",info);
139 24 : tol *= nrm;
140 : }
141 24 : PetscCall(PetscInfo(fn,"||A||_F = %g, new tol: %g\n",(double)nrm,(double)tol));
142 :
143 : /* X = I */
144 24 : PetscCall(PetscArrayzero(X,N));
145 1344 : for (i=0;i<n;i++) X[i+i*ld] = 1.0;
146 :
147 124 : for (it=0;it<MAXIT && !converged;it++) {
148 :
149 : /* G = (5/16)*I + (1/16)*M*(15*I-5*M+M*M) */
150 100 : PetscCallBLAS("BLASgemm",BLASgemm_("N","N",&n,&n,&n,&sone,M,&ld,M,&ld,&szero,M2,&ld));
151 100 : PetscCallBLAS("BLASaxpy",BLASaxpy_(&N,&smfive,M,&one,M2,&one));
152 6140 : for (i=0;i<n;i++) M2[i+i*ld] += 15.0;
153 100 : PetscCallBLAS("BLASgemm",BLASgemm_("N","N",&n,&n,&n,&s1d16,M,&ld,M2,&ld,&szero,G,&ld));
154 6140 : for (i=0;i<n;i++) G[i+i*ld] += 5.0/16.0;
155 :
156 : /* X = X*G */
157 100 : PetscCall(PetscArraycpy(M2,X,N));
158 100 : PetscCallBLAS("BLASgemm",BLASgemm_("N","N",&n,&n,&n,&sone,M2,&ld,G,&ld,&szero,X,&ld));
159 :
160 : /* M = M*inv(G*G) */
161 100 : PetscCallBLAS("BLASgemm",BLASgemm_("N","N",&n,&n,&n,&sone,G,&ld,G,&ld,&szero,M2,&ld));
162 100 : PetscCallBLAS("LAPACKgetrf",LAPACKgetrf_(&n,&n,M2,&ld,piv,&info));
163 100 : SlepcCheckLapackInfo("getrf",info);
164 100 : PetscCallBLAS("LAPACKgetri",LAPACKgetri_(&n,M2,&ld,piv,work,&lwork,&info));
165 100 : SlepcCheckLapackInfo("getri",info);
166 :
167 100 : PetscCall(PetscArraycpy(G,M,N));
168 100 : PetscCallBLAS("BLASgemm",BLASgemm_("N","N",&n,&n,&n,&sone,G,&ld,M2,&ld,&szero,M,&ld));
169 :
170 : /* check ||I-M|| */
171 100 : PetscCall(PetscArraycpy(M2,M,N));
172 6140 : for (i=0;i<n;i++) M2[i+i*ld] -= 1.0;
173 100 : Mres = LAPACKlange_("fro",&n,&n,M2,&n,rwork);
174 100 : PetscCheck(!PetscIsNanReal(Mres),PETSC_COMM_SELF,PETSC_ERR_FP,"The computed norm is not-a-number");
175 100 : if (Mres<=tol) converged = PETSC_TRUE;
176 100 : PetscCall(PetscInfo(fn,"it: %" PetscInt_FMT " res: %g\n",it,(double)Mres));
177 100 : PetscCall(PetscLogFlops(8.0*n*n*n+2.0*n*n+2.0*n*n*n/3.0+4.0*n*n*n/3.0+2.0*n*n*n+2.0*n*n));
178 : }
179 :
180 24 : PetscCheck(Mres<=tol,PETSC_COMM_SELF,PETSC_ERR_LIB,"SQRTM not converged after %d iterations",MAXIT);
181 :
182 : /* undo scaling */
183 24 : if (nrm>1.0) PetscCallBLAS("BLASscal",BLASscal_(&N,&sqrtnrm,A,&one));
184 :
185 24 : PetscCall(PetscFree5(M,M2,G,work,piv));
186 24 : PetscCall(SlepcResetFlushToZero(&ftz));
187 24 : PetscFunctionReturn(PETSC_SUCCESS);
188 : }
189 :
190 : #if defined(PETSC_HAVE_CUDA)
191 : #include "../src/sys/classes/fn/impls/cuda/fnutilcuda.h"
192 : #include <slepccupmblas.h>
193 :
194 : #if defined(PETSC_HAVE_MAGMA)
195 : #include <slepcmagma.h>
196 :
197 : /*
198 : * Matrix square root by Sadeghi iteration. CUDA version.
199 : * Computes the principal square root of the matrix A using the
200 : * Sadeghi iteration. A is overwritten with sqrtm(A).
201 : */
202 : PetscErrorCode FNSqrtmSadeghi_CUDAm(FN fn,PetscBLASInt n,PetscScalar *d_A,PetscBLASInt ld)
203 : {
204 : PetscScalar *d_M,*d_M2,*d_G,*d_work,alpha;
205 : const PetscScalar szero=0.0,sone=1.0,smfive=-5.0,s15=15.0,s1d16=1.0/16.0;
206 : PetscReal tol,Mres=0.0,nrm,sqrtnrm=1.0;
207 : PetscInt it,nb,lwork;
208 : PetscBLASInt *piv,N;
209 : const PetscBLASInt one=1;
210 : PetscBool converged=PETSC_FALSE;
211 : cublasHandle_t cublasv2handle;
212 :
213 : PetscFunctionBegin;
214 : PetscCall(PetscDeviceInitialize(PETSC_DEVICE_CUDA)); /* For CUDA event timers */
215 : PetscCall(PetscCUBLASGetHandle(&cublasv2handle));
216 : PetscCall(SlepcMagmaInit());
217 : N = n*n;
218 : tol = PetscSqrtReal((PetscReal)n)*PETSC_MACHINE_EPSILON/2;
219 :
220 : PetscCall(PetscMalloc1(n,&piv));
221 : PetscCallCUDA(cudaMalloc((void **)&d_M,sizeof(PetscScalar)*N));
222 : PetscCallCUDA(cudaMalloc((void **)&d_M2,sizeof(PetscScalar)*N));
223 : PetscCallCUDA(cudaMalloc((void **)&d_G,sizeof(PetscScalar)*N));
224 :
225 : nb = magma_get_xgetri_nb(n);
226 : lwork = nb*n;
227 : PetscCallCUDA(cudaMalloc((void **)&d_work,sizeof(PetscScalar)*lwork));
228 : PetscCall(PetscLogGpuTimeBegin());
229 :
230 : /* M = A */
231 : PetscCallCUDA(cudaMemcpy(d_M,d_A,sizeof(PetscScalar)*N,cudaMemcpyDeviceToDevice));
232 :
233 : /* scale M */
234 : PetscCallCUBLAS(cublasXnrm2(cublasv2handle,N,d_M,one,&nrm));
235 : if (nrm>1.0) {
236 : sqrtnrm = PetscSqrtReal(nrm);
237 : alpha = 1.0/nrm;
238 : PetscCallCUBLAS(cublasXscal(cublasv2handle,N,&alpha,d_M,one));
239 : tol *= nrm;
240 : }
241 : PetscCall(PetscInfo(fn,"||A||_F = %g, new tol: %g\n",(double)nrm,(double)tol));
242 :
243 : /* X = I */
244 : PetscCallCUDA(cudaMemset(d_A,0,sizeof(PetscScalar)*N));
245 : PetscCall(set_diagonal(n,d_A,ld,sone));
246 :
247 : for (it=0;it<MAXIT && !converged;it++) {
248 :
249 : /* G = (5/16)*I + (1/16)*M*(15*I-5*M+M*M) */
250 : PetscCallCUBLAS(cublasXgemm(cublasv2handle,CUBLAS_OP_N,CUBLAS_OP_N,n,n,n,&sone,d_M,ld,d_M,ld,&szero,d_M2,ld));
251 : PetscCallCUBLAS(cublasXaxpy(cublasv2handle,N,&smfive,d_M,one,d_M2,one));
252 : PetscCall(shift_diagonal(n,d_M2,ld,s15));
253 : PetscCallCUBLAS(cublasXgemm(cublasv2handle,CUBLAS_OP_N,CUBLAS_OP_N,n,n,n,&s1d16,d_M,ld,d_M2,ld,&szero,d_G,ld));
254 : PetscCall(shift_diagonal(n,d_G,ld,5.0/16.0));
255 :
256 : /* X = X*G */
257 : PetscCallCUDA(cudaMemcpy(d_M2,d_A,sizeof(PetscScalar)*N,cudaMemcpyDeviceToDevice));
258 : PetscCallCUBLAS(cublasXgemm(cublasv2handle,CUBLAS_OP_N,CUBLAS_OP_N,n,n,n,&sone,d_M2,ld,d_G,ld,&szero,d_A,ld));
259 :
260 : /* M = M*inv(G*G) */
261 : PetscCallCUBLAS(cublasXgemm(cublasv2handle,CUBLAS_OP_N,CUBLAS_OP_N,n,n,n,&sone,d_G,ld,d_G,ld,&szero,d_M2,ld));
262 : /* magma */
263 : PetscCallMAGMA(magma_xgetrf_gpu,n,n,d_M2,ld,piv);
264 : PetscCallMAGMA(magma_xgetri_gpu,n,d_M2,ld,piv,d_work,lwork);
265 : /* magma */
266 : PetscCallCUDA(cudaMemcpy(d_G,d_M,sizeof(PetscScalar)*N,cudaMemcpyDeviceToDevice));
267 : PetscCallCUBLAS(cublasXgemm(cublasv2handle,CUBLAS_OP_N,CUBLAS_OP_N,n,n,n,&sone,d_G,ld,d_M2,ld,&szero,d_M,ld));
268 :
269 : /* check ||I-M|| */
270 : PetscCallCUDA(cudaMemcpy(d_M2,d_M,sizeof(PetscScalar)*N,cudaMemcpyDeviceToDevice));
271 : PetscCall(shift_diagonal(n,d_M2,ld,-1.0));
272 : PetscCallCUBLAS(cublasXnrm2(cublasv2handle,N,d_M2,one,&Mres));
273 : PetscCheck(!PetscIsNanReal(Mres),PETSC_COMM_SELF,PETSC_ERR_FP,"The computed norm is not-a-number");
274 : if (Mres<=tol) converged = PETSC_TRUE;
275 : PetscCall(PetscInfo(fn,"it: %" PetscInt_FMT " res: %g\n",it,(double)Mres));
276 : PetscCall(PetscLogGpuFlops(8.0*n*n*n+2.0*n*n+2.0*n*n*n/3.0+4.0*n*n*n/3.0+2.0*n*n*n+2.0*n*n));
277 : }
278 :
279 : PetscCheck(Mres<=tol,PETSC_COMM_SELF,PETSC_ERR_LIB,"SQRTM not converged after %d iterations", MAXIT);
280 :
281 : if (nrm>1.0) {
282 : alpha = sqrtnrm;
283 : PetscCallCUBLAS(cublasXscal(cublasv2handle,N,&alpha,d_A,one));
284 : }
285 : PetscCall(PetscLogGpuTimeEnd());
286 :
287 : PetscCallCUDA(cudaFree(d_M));
288 : PetscCallCUDA(cudaFree(d_M2));
289 : PetscCallCUDA(cudaFree(d_G));
290 : PetscCallCUDA(cudaFree(d_work));
291 : PetscCall(PetscFree(piv));
292 : PetscFunctionReturn(PETSC_SUCCESS);
293 : }
294 : #endif /* PETSC_HAVE_MAGMA */
295 : #endif /* PETSC_HAVE_CUDA */
296 :
297 12 : static PetscErrorCode FNEvaluateFunctionMat_Sqrt_Sadeghi(FN fn,Mat A,Mat B)
298 : {
299 12 : PetscBLASInt n=0;
300 12 : PetscScalar *Ba;
301 12 : PetscInt m;
302 :
303 12 : PetscFunctionBegin;
304 12 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
305 12 : PetscCall(MatDenseGetArray(B,&Ba));
306 12 : PetscCall(MatGetSize(A,&m,NULL));
307 12 : PetscCall(PetscBLASIntCast(m,&n));
308 12 : PetscCall(FNSqrtmSadeghi(fn,n,Ba,n));
309 12 : PetscCall(MatDenseRestoreArray(B,&Ba));
310 12 : PetscFunctionReturn(PETSC_SUCCESS);
311 : }
312 :
313 : #if defined(PETSC_HAVE_CUDA)
314 : PetscErrorCode FNEvaluateFunctionMat_Sqrt_NS_CUDA(FN fn,Mat A,Mat B)
315 : {
316 : PetscBLASInt n=0;
317 : PetscScalar *Ba;
318 : PetscInt m;
319 :
320 : PetscFunctionBegin;
321 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
322 : PetscCall(MatDenseCUDAGetArray(B,&Ba));
323 : PetscCall(MatGetSize(A,&m,NULL));
324 : PetscCall(PetscBLASIntCast(m,&n));
325 : PetscCall(FNSqrtmNewtonSchulz_CUDA(fn,n,Ba,n,PETSC_FALSE));
326 : PetscCall(MatDenseCUDARestoreArray(B,&Ba));
327 : PetscFunctionReturn(PETSC_SUCCESS);
328 : }
329 :
330 : #if defined(PETSC_HAVE_MAGMA)
331 : PetscErrorCode FNEvaluateFunctionMat_Sqrt_DBP_CUDAm(FN fn,Mat A,Mat B)
332 : {
333 : PetscBLASInt n=0;
334 : PetscScalar *T;
335 : PetscInt m;
336 :
337 : PetscFunctionBegin;
338 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
339 : PetscCall(MatDenseCUDAGetArray(B,&T));
340 : PetscCall(MatGetSize(A,&m,NULL));
341 : PetscCall(PetscBLASIntCast(m,&n));
342 : PetscCall(FNSqrtmDenmanBeavers_CUDAm(fn,n,T,n,PETSC_FALSE));
343 : PetscCall(MatDenseCUDARestoreArray(B,&T));
344 : PetscFunctionReturn(PETSC_SUCCESS);
345 : }
346 :
347 : PetscErrorCode FNEvaluateFunctionMat_Sqrt_Sadeghi_CUDAm(FN fn,Mat A,Mat B)
348 : {
349 : PetscBLASInt n=0;
350 : PetscScalar *Ba;
351 : PetscInt m;
352 :
353 : PetscFunctionBegin;
354 : if (A!=B) PetscCall(MatCopy(A,B,SAME_NONZERO_PATTERN));
355 : PetscCall(MatDenseCUDAGetArray(B,&Ba));
356 : PetscCall(MatGetSize(A,&m,NULL));
357 : PetscCall(PetscBLASIntCast(m,&n));
358 : PetscCall(FNSqrtmSadeghi_CUDAm(fn,n,Ba,n));
359 : PetscCall(MatDenseCUDARestoreArray(B,&Ba));
360 : PetscFunctionReturn(PETSC_SUCCESS);
361 : }
362 : #endif /* PETSC_HAVE_MAGMA */
363 : #endif /* PETSC_HAVE_CUDA */
364 :
365 15 : static PetscErrorCode FNView_Sqrt(FN fn,PetscViewer viewer)
366 : {
367 15 : PetscBool isascii;
368 15 : char str[50];
369 15 : const char *methodname[] = {
370 : "Schur method for the square root",
371 : "Denman-Beavers (product form)",
372 : "Newton-Schulz iteration",
373 : "Sadeghi iteration"
374 : };
375 15 : const int nmeth=PETSC_STATIC_ARRAY_LENGTH(methodname);
376 :
377 15 : PetscFunctionBegin;
378 15 : PetscCall(PetscObjectTypeCompare((PetscObject)viewer,PETSCVIEWERASCII,&isascii));
379 15 : if (isascii) {
380 15 : if (fn->beta==(PetscScalar)1.0) {
381 1 : if (fn->alpha==(PetscScalar)1.0) PetscCall(PetscViewerASCIIPrintf(viewer," square root: sqrt(x)\n"));
382 : else {
383 0 : PetscCall(SlepcSNPrintfScalar(str,sizeof(str),fn->alpha,PETSC_TRUE));
384 0 : PetscCall(PetscViewerASCIIPrintf(viewer," square root: sqrt(%s*x)\n",str));
385 : }
386 : } else {
387 14 : PetscCall(SlepcSNPrintfScalar(str,sizeof(str),fn->beta,PETSC_TRUE));
388 14 : if (fn->alpha==(PetscScalar)1.0) PetscCall(PetscViewerASCIIPrintf(viewer," square root: %s*sqrt(x)\n",str));
389 : else {
390 14 : PetscCall(PetscViewerASCIIPrintf(viewer," square root: %s",str));
391 14 : PetscCall(PetscViewerASCIIUseTabs(viewer,PETSC_FALSE));
392 14 : PetscCall(SlepcSNPrintfScalar(str,sizeof(str),fn->alpha,PETSC_TRUE));
393 14 : PetscCall(PetscViewerASCIIPrintf(viewer,"*sqrt(%s*x)\n",str));
394 14 : PetscCall(PetscViewerASCIIUseTabs(viewer,PETSC_TRUE));
395 : }
396 : }
397 15 : if (fn->method<nmeth) PetscCall(PetscViewerASCIIPrintf(viewer," computing matrix functions with: %s\n",methodname[fn->method]));
398 : }
399 15 : PetscFunctionReturn(PETSC_SUCCESS);
400 : }
401 :
402 27 : SLEPC_EXTERN PetscErrorCode FNCreate_Sqrt(FN fn)
403 : {
404 27 : PetscFunctionBegin;
405 27 : fn->ops->evaluatefunction = FNEvaluateFunction_Sqrt;
406 27 : fn->ops->evaluatederivative = FNEvaluateDerivative_Sqrt;
407 27 : fn->ops->evaluatefunctionmat[0] = FNEvaluateFunctionMat_Sqrt_Schur;
408 27 : fn->ops->evaluatefunctionmat[1] = FNEvaluateFunctionMat_Sqrt_DBP;
409 27 : fn->ops->evaluatefunctionmat[2] = FNEvaluateFunctionMat_Sqrt_NS;
410 27 : fn->ops->evaluatefunctionmat[3] = FNEvaluateFunctionMat_Sqrt_Sadeghi;
411 : #if defined(PETSC_HAVE_CUDA)
412 : fn->ops->evaluatefunctionmatcuda[2] = FNEvaluateFunctionMat_Sqrt_NS_CUDA;
413 : #if defined(PETSC_HAVE_MAGMA)
414 : fn->ops->evaluatefunctionmatcuda[1] = FNEvaluateFunctionMat_Sqrt_DBP_CUDAm;
415 : fn->ops->evaluatefunctionmatcuda[3] = FNEvaluateFunctionMat_Sqrt_Sadeghi_CUDAm;
416 : #endif /* PETSC_HAVE_MAGMA */
417 : #endif /* PETSC_HAVE_CUDA */
418 27 : fn->ops->evaluatefunctionmatvec[0] = FNEvaluateFunctionMatVec_Sqrt_Schur;
419 27 : fn->ops->view = FNView_Sqrt;
420 27 : PetscFunctionReturn(PETSC_SUCCESS);
421 : }
|
