|
LAPACK 3.3.0
|
|
LAPACK 3.3.0
|
00001 SUBROUTINE DSPTRS( UPLO, N, NRHS, AP, IPIV, B, LDB, INFO ) 00002 * 00003 * -- LAPACK routine (version 3.2) -- 00004 * -- LAPACK is a software package provided by Univ. of Tennessee, -- 00005 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- 00006 * November 2006 00007 * 00008 * .. Scalar Arguments .. 00009 CHARACTER UPLO 00010 INTEGER INFO, LDB, N, NRHS 00011 * .. 00012 * .. Array Arguments .. 00013 INTEGER IPIV( * ) 00014 DOUBLE PRECISION AP( * ), B( LDB, * ) 00015 * .. 00016 * 00017 * Purpose 00018 * ======= 00019 * 00020 * DSPTRS solves a system of linear equations A*X = B with a real 00021 * symmetric matrix A stored in packed format using the factorization 00022 * A = U*D*U**T or A = L*D*L**T computed by DSPTRF. 00023 * 00024 * Arguments 00025 * ========= 00026 * 00027 * UPLO (input) CHARACTER*1 00028 * Specifies whether the details of the factorization are stored 00029 * as an upper or lower triangular matrix. 00030 * = 'U': Upper triangular, form is A = U*D*U**T; 00031 * = 'L': Lower triangular, form is A = L*D*L**T. 00032 * 00033 * N (input) INTEGER 00034 * The order of the matrix A. N >= 0. 00035 * 00036 * NRHS (input) INTEGER 00037 * The number of right hand sides, i.e., the number of columns 00038 * of the matrix B. NRHS >= 0. 00039 * 00040 * AP (input) DOUBLE PRECISION array, dimension (N*(N+1)/2) 00041 * The block diagonal matrix D and the multipliers used to 00042 * obtain the factor U or L as computed by DSPTRF, stored as a 00043 * packed triangular matrix. 00044 * 00045 * IPIV (input) INTEGER array, dimension (N) 00046 * Details of the interchanges and the block structure of D 00047 * as determined by DSPTRF. 00048 * 00049 * B (input/output) DOUBLE PRECISION array, dimension (LDB,NRHS) 00050 * On entry, the right hand side matrix B. 00051 * On exit, the solution matrix X. 00052 * 00053 * LDB (input) INTEGER 00054 * The leading dimension of the array B. LDB >= max(1,N). 00055 * 00056 * INFO (output) INTEGER 00057 * = 0: successful exit 00058 * < 0: if INFO = -i, the i-th argument had an illegal value 00059 * 00060 * ===================================================================== 00061 * 00062 * .. Parameters .. 00063 DOUBLE PRECISION ONE 00064 PARAMETER ( ONE = 1.0D+0 ) 00065 * .. 00066 * .. Local Scalars .. 00067 LOGICAL UPPER 00068 INTEGER J, K, KC, KP 00069 DOUBLE PRECISION AK, AKM1, AKM1K, BK, BKM1, DENOM 00070 * .. 00071 * .. External Functions .. 00072 LOGICAL LSAME 00073 EXTERNAL LSAME 00074 * .. 00075 * .. External Subroutines .. 00076 EXTERNAL DGEMV, DGER, DSCAL, DSWAP, XERBLA 00077 * .. 00078 * .. Intrinsic Functions .. 00079 INTRINSIC MAX 00080 * .. 00081 * .. Executable Statements .. 00082 * 00083 INFO = 0 00084 UPPER = LSAME( UPLO, 'U' ) 00085 IF( .NOT.UPPER .AND. .NOT.LSAME( UPLO, 'L' ) ) THEN 00086 INFO = -1 00087 ELSE IF( N.LT.0 ) THEN 00088 INFO = -2 00089 ELSE IF( NRHS.LT.0 ) THEN 00090 INFO = -3 00091 ELSE IF( LDB.LT.MAX( 1, N ) ) THEN 00092 INFO = -7 00093 END IF 00094 IF( INFO.NE.0 ) THEN 00095 CALL XERBLA( 'DSPTRS', -INFO ) 00096 RETURN 00097 END IF 00098 * 00099 * Quick return if possible 00100 * 00101 IF( N.EQ.0 .OR. NRHS.EQ.0 ) 00102 $ RETURN 00103 * 00104 IF( UPPER ) THEN 00105 * 00106 * Solve A*X = B, where A = U*D*U'. 00107 * 00108 * First solve U*D*X = B, overwriting B with X. 00109 * 00110 * K is the main loop index, decreasing from N to 1 in steps of 00111 * 1 or 2, depending on the size of the diagonal blocks. 00112 * 00113 K = N 00114 KC = N*( N+1 ) / 2 + 1 00115 10 CONTINUE 00116 * 00117 * If K < 1, exit from loop. 00118 * 00119 IF( K.LT.1 ) 00120 $ GO TO 30 00121 * 00122 KC = KC - K 00123 IF( IPIV( K ).GT.0 ) THEN 00124 * 00125 * 1 x 1 diagonal block 00126 * 00127 * Interchange rows K and IPIV(K). 00128 * 00129 KP = IPIV( K ) 00130 IF( KP.NE.K ) 00131 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00132 * 00133 * Multiply by inv(U(K)), where U(K) is the transformation 00134 * stored in column K of A. 00135 * 00136 CALL DGER( K-1, NRHS, -ONE, AP( KC ), 1, B( K, 1 ), LDB, 00137 $ B( 1, 1 ), LDB ) 00138 * 00139 * Multiply by the inverse of the diagonal block. 00140 * 00141 CALL DSCAL( NRHS, ONE / AP( KC+K-1 ), B( K, 1 ), LDB ) 00142 K = K - 1 00143 ELSE 00144 * 00145 * 2 x 2 diagonal block 00146 * 00147 * Interchange rows K-1 and -IPIV(K). 00148 * 00149 KP = -IPIV( K ) 00150 IF( KP.NE.K-1 ) 00151 $ CALL DSWAP( NRHS, B( K-1, 1 ), LDB, B( KP, 1 ), LDB ) 00152 * 00153 * Multiply by inv(U(K)), where U(K) is the transformation 00154 * stored in columns K-1 and K of A. 00155 * 00156 CALL DGER( K-2, NRHS, -ONE, AP( KC ), 1, B( K, 1 ), LDB, 00157 $ B( 1, 1 ), LDB ) 00158 CALL DGER( K-2, NRHS, -ONE, AP( KC-( K-1 ) ), 1, 00159 $ B( K-1, 1 ), LDB, B( 1, 1 ), LDB ) 00160 * 00161 * Multiply by the inverse of the diagonal block. 00162 * 00163 AKM1K = AP( KC+K-2 ) 00164 AKM1 = AP( KC-1 ) / AKM1K 00165 AK = AP( KC+K-1 ) / AKM1K 00166 DENOM = AKM1*AK - ONE 00167 DO 20 J = 1, NRHS 00168 BKM1 = B( K-1, J ) / AKM1K 00169 BK = B( K, J ) / AKM1K 00170 B( K-1, J ) = ( AK*BKM1-BK ) / DENOM 00171 B( K, J ) = ( AKM1*BK-BKM1 ) / DENOM 00172 20 CONTINUE 00173 KC = KC - K + 1 00174 K = K - 2 00175 END IF 00176 * 00177 GO TO 10 00178 30 CONTINUE 00179 * 00180 * Next solve U'*X = B, overwriting B with X. 00181 * 00182 * K is the main loop index, increasing from 1 to N in steps of 00183 * 1 or 2, depending on the size of the diagonal blocks. 00184 * 00185 K = 1 00186 KC = 1 00187 40 CONTINUE 00188 * 00189 * If K > N, exit from loop. 00190 * 00191 IF( K.GT.N ) 00192 $ GO TO 50 00193 * 00194 IF( IPIV( K ).GT.0 ) THEN 00195 * 00196 * 1 x 1 diagonal block 00197 * 00198 * Multiply by inv(U'(K)), where U(K) is the transformation 00199 * stored in column K of A. 00200 * 00201 CALL DGEMV( 'Transpose', K-1, NRHS, -ONE, B, LDB, AP( KC ), 00202 $ 1, ONE, B( K, 1 ), LDB ) 00203 * 00204 * Interchange rows K and IPIV(K). 00205 * 00206 KP = IPIV( K ) 00207 IF( KP.NE.K ) 00208 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00209 KC = KC + K 00210 K = K + 1 00211 ELSE 00212 * 00213 * 2 x 2 diagonal block 00214 * 00215 * Multiply by inv(U'(K+1)), where U(K+1) is the transformation 00216 * stored in columns K and K+1 of A. 00217 * 00218 CALL DGEMV( 'Transpose', K-1, NRHS, -ONE, B, LDB, AP( KC ), 00219 $ 1, ONE, B( K, 1 ), LDB ) 00220 CALL DGEMV( 'Transpose', K-1, NRHS, -ONE, B, LDB, 00221 $ AP( KC+K ), 1, ONE, B( K+1, 1 ), LDB ) 00222 * 00223 * Interchange rows K and -IPIV(K). 00224 * 00225 KP = -IPIV( K ) 00226 IF( KP.NE.K ) 00227 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00228 KC = KC + 2*K + 1 00229 K = K + 2 00230 END IF 00231 * 00232 GO TO 40 00233 50 CONTINUE 00234 * 00235 ELSE 00236 * 00237 * Solve A*X = B, where A = L*D*L'. 00238 * 00239 * First solve L*D*X = B, overwriting B with X. 00240 * 00241 * K is the main loop index, increasing from 1 to N in steps of 00242 * 1 or 2, depending on the size of the diagonal blocks. 00243 * 00244 K = 1 00245 KC = 1 00246 60 CONTINUE 00247 * 00248 * If K > N, exit from loop. 00249 * 00250 IF( K.GT.N ) 00251 $ GO TO 80 00252 * 00253 IF( IPIV( K ).GT.0 ) THEN 00254 * 00255 * 1 x 1 diagonal block 00256 * 00257 * Interchange rows K and IPIV(K). 00258 * 00259 KP = IPIV( K ) 00260 IF( KP.NE.K ) 00261 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00262 * 00263 * Multiply by inv(L(K)), where L(K) is the transformation 00264 * stored in column K of A. 00265 * 00266 IF( K.LT.N ) 00267 $ CALL DGER( N-K, NRHS, -ONE, AP( KC+1 ), 1, B( K, 1 ), 00268 $ LDB, B( K+1, 1 ), LDB ) 00269 * 00270 * Multiply by the inverse of the diagonal block. 00271 * 00272 CALL DSCAL( NRHS, ONE / AP( KC ), B( K, 1 ), LDB ) 00273 KC = KC + N - K + 1 00274 K = K + 1 00275 ELSE 00276 * 00277 * 2 x 2 diagonal block 00278 * 00279 * Interchange rows K+1 and -IPIV(K). 00280 * 00281 KP = -IPIV( K ) 00282 IF( KP.NE.K+1 ) 00283 $ CALL DSWAP( NRHS, B( K+1, 1 ), LDB, B( KP, 1 ), LDB ) 00284 * 00285 * Multiply by inv(L(K)), where L(K) is the transformation 00286 * stored in columns K and K+1 of A. 00287 * 00288 IF( K.LT.N-1 ) THEN 00289 CALL DGER( N-K-1, NRHS, -ONE, AP( KC+2 ), 1, B( K, 1 ), 00290 $ LDB, B( K+2, 1 ), LDB ) 00291 CALL DGER( N-K-1, NRHS, -ONE, AP( KC+N-K+2 ), 1, 00292 $ B( K+1, 1 ), LDB, B( K+2, 1 ), LDB ) 00293 END IF 00294 * 00295 * Multiply by the inverse of the diagonal block. 00296 * 00297 AKM1K = AP( KC+1 ) 00298 AKM1 = AP( KC ) / AKM1K 00299 AK = AP( KC+N-K+1 ) / AKM1K 00300 DENOM = AKM1*AK - ONE 00301 DO 70 J = 1, NRHS 00302 BKM1 = B( K, J ) / AKM1K 00303 BK = B( K+1, J ) / AKM1K 00304 B( K, J ) = ( AK*BKM1-BK ) / DENOM 00305 B( K+1, J ) = ( AKM1*BK-BKM1 ) / DENOM 00306 70 CONTINUE 00307 KC = KC + 2*( N-K ) + 1 00308 K = K + 2 00309 END IF 00310 * 00311 GO TO 60 00312 80 CONTINUE 00313 * 00314 * Next solve L'*X = B, overwriting B with X. 00315 * 00316 * K is the main loop index, decreasing from N to 1 in steps of 00317 * 1 or 2, depending on the size of the diagonal blocks. 00318 * 00319 K = N 00320 KC = N*( N+1 ) / 2 + 1 00321 90 CONTINUE 00322 * 00323 * If K < 1, exit from loop. 00324 * 00325 IF( K.LT.1 ) 00326 $ GO TO 100 00327 * 00328 KC = KC - ( N-K+1 ) 00329 IF( IPIV( K ).GT.0 ) THEN 00330 * 00331 * 1 x 1 diagonal block 00332 * 00333 * Multiply by inv(L'(K)), where L(K) is the transformation 00334 * stored in column K of A. 00335 * 00336 IF( K.LT.N ) 00337 $ CALL DGEMV( 'Transpose', N-K, NRHS, -ONE, B( K+1, 1 ), 00338 $ LDB, AP( KC+1 ), 1, ONE, B( K, 1 ), LDB ) 00339 * 00340 * Interchange rows K and IPIV(K). 00341 * 00342 KP = IPIV( K ) 00343 IF( KP.NE.K ) 00344 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00345 K = K - 1 00346 ELSE 00347 * 00348 * 2 x 2 diagonal block 00349 * 00350 * Multiply by inv(L'(K-1)), where L(K-1) is the transformation 00351 * stored in columns K-1 and K of A. 00352 * 00353 IF( K.LT.N ) THEN 00354 CALL DGEMV( 'Transpose', N-K, NRHS, -ONE, B( K+1, 1 ), 00355 $ LDB, AP( KC+1 ), 1, ONE, B( K, 1 ), LDB ) 00356 CALL DGEMV( 'Transpose', N-K, NRHS, -ONE, B( K+1, 1 ), 00357 $ LDB, AP( KC-( N-K ) ), 1, ONE, B( K-1, 1 ), 00358 $ LDB ) 00359 END IF 00360 * 00361 * Interchange rows K and -IPIV(K). 00362 * 00363 KP = -IPIV( K ) 00364 IF( KP.NE.K ) 00365 $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) 00366 KC = KC - ( N-K+2 ) 00367 K = K - 2 00368 END IF 00369 * 00370 GO TO 90 00371 100 CONTINUE 00372 END IF 00373 * 00374 RETURN 00375 * 00376 * End of DSPTRS 00377 * 00378 END
1.7.3 
