/* -- translated by f2c (version 19940927). You must link the resulting object file with the libraries: -lf2c -lm (in that order) */ #include #include "f2c.h" /* Table of constant values */ static integer c__1 = 1; static doublereal c_b22 = 0.; static logical c_true = TRUE_; static logical c_false = FALSE_; int dlatms_slu(integer *m, integer *n, char *dist, integer * iseed, char *sym, doublereal *d, integer *mode, doublereal *cond, doublereal *dmax__, integer *kl, integer *ku, char *pack, doublereal *a, integer *lda, doublereal *work, integer *info) { /* System generated locals */ integer a_dim1, a_offset, i__1, i__2, i__3, i__4, i__5, i__6; doublereal d__1, d__2, d__3; logical L__1; /* Builtin functions */ double cos(doublereal), sin(doublereal); /* Local variables */ static integer ilda, icol; static doublereal temp; static integer irow, isym; static doublereal c; static integer i, j, k; static doublereal s, alpha, angle; static integer ipack; extern /* Subroutine */ int dscal_(integer *, doublereal *, doublereal *, integer *); static integer ioffg; static integer iinfo, idist, mnmin; extern /* Subroutine */ int dcopy_(integer *, doublereal *, integer *, doublereal *, integer *); static integer iskew; static doublereal extra, dummy; extern /* Subroutine */ int dlatm1_slu(integer *, doublereal *, integer *, integer *, integer *, doublereal *, integer *, integer *); static integer ic, jc, nc; extern /* Subroutine */ int dlagge_slu(integer *, integer *, integer *, integer *, doublereal *, doublereal *, integer *, integer *, doublereal *, integer *); static integer il, iendch, ir, jr, ipackg, mr, minlda; extern doublereal dlarnd_slu(integer *, integer *); extern /* Subroutine */ int dlaset_slu(char *, integer *, integer *, doublereal *, doublereal *, doublereal *, integer *), dlartg_slu(doublereal *, doublereal *, doublereal *, doublereal *, doublereal *), dlagsy_slu(integer *, integer *, doublereal *, doublereal *, integer *, integer *, doublereal *, integer *), dlarot_slu(logical *, logical *, logical *, integer *, doublereal *, doublereal *, doublereal *, integer *, doublereal *, doublereal *); extern int input_error(char *, int *); static logical iltemp, givens; static integer ioffst, irsign; static logical ilextr, topdwn; static integer ir1, ir2, isympk, jch, llb, jkl, jku, uub; /* -- LAPACK test routine (version 2.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University September 30, 1994 Purpose ======= DLATMS generates random matrices with specified singular values (or symmetric/hermitian with specified eigenvalues) for testing LAPACK programs. DLATMS operates by applying the following sequence of operations: Set the diagonal to D, where D may be input or computed according to MODE, COND, DMAX, and SYM as described below. Generate a matrix with the appropriate band structure, by one of two methods: Method A: Generate a dense M x N matrix by multiplying D on the left and the right by random unitary matrices, then: Reduce the bandwidth according to KL and KU, using Householder transformations. Method B: Convert the bandwidth-0 (i.e., diagonal) matrix to a bandwidth-1 matrix using Givens rotations, "chasing" out-of-band elements back, much as in QR; then convert the bandwidth-1 to a bandwidth-2 matrix, etc. Note that for reasonably small bandwidths (relative to M and N) this requires less storage, as a dense matrix is not generated. Also, for symmetric matrices, only one triangle is generated. Method A is chosen if the bandwidth is a large fraction of the order of the matrix, and LDA is at least M (so a dense matrix can be stored.) Method B is chosen if the bandwidth is small (< 1/2 N for symmetric, < .3 N+M for non-symmetric), or LDA is less than M and not less than the bandwidth. Pack the matrix if desired. Options specified by PACK are: no packing zero out upper half (if symmetric) zero out lower half (if symmetric) store the upper half columnwise (if symmetric or upper triangular) store the lower half columnwise (if symmetric or lower triangular) store the lower triangle in banded format (if symmetric or lower triangular) store the upper triangle in banded format (if symmetric or upper triangular) store the entire matrix in banded format If Method B is chosen, and band format is specified, then the matrix will be generated in the band format, so no repacking will be necessary. Arguments ========= M - INTEGER The number of rows of A. Not modified. N - INTEGER The number of columns of A. Not modified. DIST - CHARACTER*1 On entry, DIST specifies the type of distribution to be used to generate the random eigen-/singular values. 'U' => UNIFORM( 0, 1 ) ( 'U' for uniform ) 'S' => UNIFORM( -1, 1 ) ( 'S' for symmetric ) 'N' => NORMAL( 0, 1 ) ( 'N' for normal ) Not modified. ISEED - INTEGER array, dimension ( 4 ) On entry ISEED specifies the seed of the random number generator. They should lie between 0 and 4095 inclusive, and ISEED(4) should be odd. The random number generator uses a linear congruential sequence limited to small integers, and so should produce machine independent random numbers. The values of ISEED are changed on exit, and can be used in the next call to DLATMS to continue the same random number sequence. Changed on exit. SYM - CHARACTER*1 If SYM='S' or 'H', the generated matrix is symmetric, with eigenvalues specified by D, COND, MODE, and DMAX; they may be positive, negative, or zero. If SYM='P', the generated matrix is symmetric, with eigenvalues (= singular values) specified by D, COND, MODE, and DMAX; they will not be negative. If SYM='N', the generated matrix is nonsymmetric, with singular values specified by D, COND, MODE, and DMAX; they will not be negative. Not modified. D - DOUBLE PRECISION array, dimension ( MIN( M , N ) ) This array is used to specify the singular values or eigenvalues of A (see SYM, above.) If MODE=0, then D is assumed to contain the singular/eigenvalues, otherwise they will be computed according to MODE, COND, and DMAX, and placed in D. Modified if MODE is nonzero. MODE - INTEGER On entry this describes how the singular/eigenvalues are to be specified: MODE = 0 means use D as input MODE = 1 sets D(1)=1 and D(2:N)=1.0/COND MODE = 2 sets D(1:N-1)=1 and D(N)=1.0/COND MODE = 3 sets D(I)=COND**(-(I-1)/(N-1)) MODE = 4 sets D(i)=1 - (i-1)/(N-1)*(1 - 1/COND) MODE = 5 sets D to random numbers in the range ( 1/COND , 1 ) such that their logarithms are uniformly distributed. MODE = 6 set D to random numbers from same distribution as the rest of the matrix. MODE < 0 has the same meaning as ABS(MODE), except that the order of the elements of D is reversed. Thus if MODE is positive, D has entries ranging from 1 to 1/COND, if negative, from 1/COND to 1, If SYM='S' or 'H', and MODE is neither 0, 6, nor -6, then the elements of D will also be multiplied by a random sign (i.e., +1 or -1.) Not modified. COND - DOUBLE PRECISION On entry, this is used as described under MODE above. If used, it must be >= 1. Not modified. DMAX - DOUBLE PRECISION If MODE is neither -6, 0 nor 6, the contents of D, as computed according to MODE and COND, will be scaled by DMAX / max(abs(D(i))); thus, the maximum absolute eigen- or singular value (which is to say the norm) will be abs(DMAX). Note that DMAX need not be positive: if DMAX is negative (or zero), D will be scaled by a negative number (or zero). Not modified. KL - INTEGER This specifies the lower bandwidth of the matrix. For example, KL=0 implies upper triangular, KL=1 implies upper Hessenberg, and KL being at least M-1 means that the matrix has full lower bandwidth. KL must equal KU if the matrix is symmetric. Not modified. KU - INTEGER This specifies the upper bandwidth of the matrix. For example, KU=0 implies lower triangular, KU=1 implies lower Hessenberg, and KU being at least N-1 means that the matrix has full upper bandwidth. KL must equal KU if the matrix is symmetric. Not modified. PACK - CHARACTER*1 This specifies packing of matrix as follows: 'N' => no packing 'U' => zero out all subdiagonal entries (if symmetric) 'L' => zero out all superdiagonal entries (if symmetric) 'C' => store the upper triangle columnwise (only if the matrix is symmetric or upper triangular) 'R' => store the lower triangle columnwise (only if the matrix is symmetric or lower triangular) 'B' => store the lower triangle in band storage scheme (only if matrix symmetric or lower triangular) 'Q' => store the upper triangle in band storage scheme (only if matrix symmetric or upper triangular) 'Z' => store the entire matrix in band storage scheme (pivoting can be provided for by using this option to store A in the trailing rows of the allocated storage) Using these options, the various LAPACK packed and banded storage schemes can be obtained: GB - use 'Z' PB, SB or TB - use 'B' or 'Q' PP, SP or TP - use 'C' or 'R' If two calls to DLATMS differ only in the PACK parameter, they will generate mathematically equivalent matrices. Not modified. A - DOUBLE PRECISION array, dimension ( LDA, N ) On exit A is the desired test matrix. A is first generated in full (unpacked) form, and then packed, if so specified by PACK. Thus, the first M elements of the first N columns will always be modified. If PACK specifies a packed or banded storage scheme, all LDA elements of the first N columns will be modified; the elements of the array which do not correspond to elements of the generated matrix are set to zero. Modified. LDA - INTEGER LDA specifies the first dimension of A as declared in the calling program. If PACK='N', 'U', 'L', 'C', or 'R', then LDA must be at least M. If PACK='B' or 'Q', then LDA must be at least MIN( KL, M-1) (which is equal to MIN(KU,N-1)). If PACK='Z', LDA must be large enough to hold the packed array: MIN( KU, N-1) + MIN( KL, M-1) + 1. Not modified. WORK - DOUBLE PRECISION array, dimension ( 3*MAX( N , M ) ) Workspace. Modified. INFO - INTEGER Error code. On exit, INFO will be set to one of the following values: 0 => normal return -1 => M negative or unequal to N and SYM='S', 'H', or 'P' -2 => N negative -3 => DIST illegal string -5 => SYM illegal string -7 => MODE not in range -6 to 6 -8 => COND less than 1.0, and MODE neither -6, 0 nor 6 -10 => KL negative -11 => KU negative, or SYM='S' or 'H' and KU not equal to KL -12 => PACK illegal string, or PACK='U' or 'L', and SYM='N'; or PACK='C' or 'Q' and SYM='N' and KL is not zero; or PACK='R' or 'B' and SYM='N' and KU is not zero; or PACK='U', 'L', 'C', 'R', 'B', or 'Q', and M is not N. -14 => LDA is less than M, or PACK='Z' and LDA is less than MIN(KU,N-1) + MIN(KL,M-1) + 1. 1 => Error return from DLATM1 2 => Cannot scale to DMAX (max. sing. value is 0) 3 => Error return from DLAGGE or SLAGSY ===================================================================== 1) Decode and Test the input parameters. Initialize flags & seed. Parameter adjustments */ --iseed; --d; a_dim1 = *lda; a_offset = a_dim1 + 1; a -= a_offset; --work; /* Function Body */ *info = 0; /* Quick return if possible */ if (*m == 0 || *n == 0) { return 0; } /* Decode DIST */ if (strncmp(dist, "U", 1)==0) { idist = 1; } else if (strncmp(dist, "S", 1)==0) { idist = 2; } else if (strncmp(dist, "N", 1)==0) { idist = 3; } else { idist = -1; } /* Decode SYM */ if (strncmp(sym, "N", 1)==0) { isym = 1; irsign = 0; } else if (strncmp(sym, "P", 1)==0) { isym = 2; irsign = 0; } else if (strncmp(sym, "S", 1)==0) { isym = 2; irsign = 1; } else if (strncmp(sym, "H", 1)==0) { isym = 2; irsign = 1; } else { isym = -1; } /* Decode PACK */ isympk = 0; if (strncmp(pack, "N", 1)==0) { ipack = 0; } else if (strncmp(pack, "U", 1)==0) { ipack = 1; isympk = 1; } else if (strncmp(pack, "L", 1)==0) { ipack = 2; isympk = 1; } else if (strncmp(pack, "C", 1)==0) { ipack = 3; isympk = 2; } else if (strncmp(pack, "R", 1)==0) { ipack = 4; isympk = 3; } else if (strncmp(pack, "B", 1)==0) { ipack = 5; isympk = 3; } else if (strncmp(pack, "Q", 1)==0) { ipack = 6; isympk = 2; } else if (strncmp(pack, "Z", 1)==0) { ipack = 7; } else { ipack = -1; } /* Set certain internal parameters */ mnmin = min(*m,*n); /* Computing MIN */ i__1 = *kl, i__2 = *m - 1; llb = min(i__1,i__2); /* Computing MIN */ i__1 = *ku, i__2 = *n - 1; uub = min(i__1,i__2); /* Computing MIN */ i__1 = *m, i__2 = *n + llb; mr = min(i__1,i__2); /* Computing MIN */ i__1 = *n, i__2 = *m + uub; nc = min(i__1,i__2); if (ipack == 5 || ipack == 6) { minlda = uub + 1; } else if (ipack == 7) { minlda = llb + uub + 1; } else { minlda = *m; } /* Use Givens rotation method if bandwidth small enough, or if LDA is too small to store the matrix unpacked. */ givens = FALSE_; if (isym == 1) { /* Computing MAX */ i__1 = 1, i__2 = mr + nc; if ((doublereal) (llb + uub) < (doublereal) max(i__1,i__2) * .3) { givens = TRUE_; } } else { if (llb << 1 < *m) { givens = TRUE_; } } if (*lda < *m && *lda >= minlda) { givens = TRUE_; } /* Set INFO if an error */ if (*m < 0) { *info = -1; } else if (*m != *n && isym != 1) { *info = -1; } else if (*n < 0) { *info = -2; } else if (idist == -1) { *info = -3; } else if (isym == -1) { *info = -5; } else if (abs(*mode) > 6) { *info = -7; } else if (*mode != 0 && abs(*mode) != 6 && *cond < 1.) { *info = -8; } else if (*kl < 0) { *info = -10; } else if (*ku < 0 || isym != 1 && *kl != *ku) { *info = -11; } else if (ipack == -1 || isympk == 1 && isym == 1 || isympk == 2 && isym == 1 && *kl > 0 || isympk == 3 && isym == 1 && *ku > 0 || isympk != 0 && *m != *n) { *info = -12; } else if (*lda < max(1,minlda)) { *info = -14; } if (*info != 0) { i__1 = -(*info); input_error("DLATMS", &i__1); return 0; } /* Initialize random number generator */ for (i = 1; i <= 4; ++i) { iseed[i] = (i__1 = iseed[i], abs(i__1)) % 4096; /* L10: */ } if (iseed[4] % 2 != 1) { ++iseed[4]; } /* 2) Set up D if indicated. Compute D according to COND and MODE */ dlatm1_slu(mode, cond, &irsign, &idist, &iseed[1], &d[1], &mnmin, &iinfo); if (iinfo != 0) { *info = 1; return 0; } /* Choose Top-Down if D is (apparently) increasing, Bottom-Up if D is (apparently) decreasing. */ if (abs(d[1]) <= (d__1 = d[mnmin], abs(d__1))) { topdwn = TRUE_; } else { topdwn = FALSE_; } if (*mode != 0 && abs(*mode) != 6) { /* Scale by DMAX */ temp = abs(d[1]); i__1 = mnmin; for (i = 2; i <= i__1; ++i) { /* Computing MAX */ d__2 = temp, d__3 = (d__1 = d[i], abs(d__1)); temp = max(d__2,d__3); /* L20: */ } if (temp > 0.) { alpha = *dmax__ / temp; } else { *info = 2; return 0; } dscal_(&mnmin, &alpha, &d[1], &c__1); } /* 3) Generate Banded Matrix using Givens rotations. Also the special case of UUB=LLB=0 Compute Addressing constants to cover all storage formats. Whether GE, SY, GB, or SB, upper or lower triangle or both, the (i,j)-th element is in A( i - ISKEW*j + IOFFST, j ) */ if (ipack > 4) { ilda = *lda - 1; iskew = 1; if (ipack > 5) { ioffst = uub + 1; } else { ioffst = 1; } } else { ilda = *lda; iskew = 0; ioffst = 0; } /* IPACKG is the format that the matrix is generated in. If this is different from IPACK, then the matrix must be repacked at the end. It also signals how to compute the norm, for scaling. */ ipackg = 0; dlaset_slu("Full", lda, n, &c_b22, &c_b22, &a[a_offset], lda); /* Diagonal Matrix -- We are done, unless it is to be stored SP/PP/TP (PACK='R' or 'C') */ if (llb == 0 && uub == 0) { i__1 = ilda + 1; dcopy_(&mnmin, &d[1], &c__1, &a[1 - iskew + ioffst + a_dim1], &i__1); if (ipack <= 2 || ipack >= 5) { ipackg = ipack; } } else if (givens) { /* Check whether to use Givens rotations, Householder transformations, or nothing. */ if (isym == 1) { /* Non-symmetric -- A = U D V */ if (ipack > 4) { ipackg = ipack; } else { ipackg = 0; } i__1 = ilda + 1; dcopy_(&mnmin, &d[1], &c__1, &a[1 - iskew + ioffst + a_dim1], & i__1); if (topdwn) { jkl = 0; i__1 = uub; for (jku = 1; jku <= i__1; ++jku) { /* Transform from bandwidth JKL, JKU-1 to JKL, JKU Last row actually rotated is M Last column actually rotated is MIN( M+ JKU, N ) Computing MIN */ i__3 = *m + jku; i__2 = min(i__3,*n) + jkl - 1; for (jr = 1; jr <= i__2; ++jr) { extra = 0.; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = sin(angle); /* Computing MAX */ i__3 = 1, i__4 = jr - jkl; icol = max(i__3,i__4); if (jr < *m) { /* Computing MIN */ i__3 = *n, i__4 = jr + jku; il = min(i__3,i__4) + 1 - icol; L__1 = jr > jkl; dlarot_slu(&c_true, &L__1, &c_false, &il, &c, &s, &a[ jr - iskew * icol + ioffst + icol * a_dim1], &ilda, &extra, &dummy); } /* Chase "EXTRA" back up */ ir = jr; ic = icol; i__3 = -jkl - jku; for (jch = jr - jkl; i__3 < 0 ? jch >= 1 : jch <= 1; jch += i__3) { if (ir < *m) { dlartg_slu(&a[ir + 1 - iskew * (ic + 1) + ioffst + (ic + 1) * a_dim1], &extra, &c, &s, &dummy); } /* Computing MAX */ i__4 = 1, i__5 = jch - jku; irow = max(i__4,i__5); il = ir + 2 - irow; temp = 0.; iltemp = jch > jku; d__1 = -s; dlarot_slu(&c_false, &iltemp, &c_true, &il, &c, & d__1, &a[irow - iskew * ic + ioffst + ic * a_dim1], &ilda, &temp, &extra); if (iltemp) { dlartg_slu(&a[irow + 1 - iskew * (ic + 1) + ioffst + (ic + 1) * a_dim1], &temp, & c, &s, &dummy); /* Computing MAX */ i__4 = 1, i__5 = jch - jku - jkl; icol = max(i__4,i__5); il = ic + 2 - icol; extra = 0.; L__1 = jch > jku + jkl; d__1 = -s; dlarot_slu(&c_true, &L__1, &c_true, &il, &c, & d__1, &a[irow - iskew * icol + ioffst + icol * a_dim1], &ilda, &extra, & temp); ic = icol; ir = irow; } /* L30: */ } /* L40: */ } /* L50: */ } jku = uub; i__1 = llb; for (jkl = 1; jkl <= i__1; ++jkl) { /* Transform from bandwidth JKL-1, JKU to JKL, JKU Computing MIN */ i__3 = *n + jkl; i__2 = min(i__3,*m) + jku - 1; for (jc = 1; jc <= i__2; ++jc) { extra = 0.; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = sin(angle); /* Computing MAX */ i__3 = 1, i__4 = jc - jku; irow = max(i__3,i__4); if (jc < *n) { /* Computing MIN */ i__3 = *m, i__4 = jc + jkl; il = min(i__3,i__4) + 1 - irow; L__1 = jc > jku; dlarot_slu(&c_false, &L__1, &c_false, &il, &c, &s, & a[irow - iskew * jc + ioffst + jc * a_dim1], &ilda, &extra, &dummy); } /* Chase "EXTRA" back up */ ic = jc; ir = irow; i__3 = -jkl - jku; for (jch = jc - jku; i__3 < 0 ? jch >= 1 : jch <= 1; jch += i__3) { if (ic < *n) { dlartg_slu(&a[ir + 1 - iskew * (ic + 1) + ioffst + (ic + 1) * a_dim1], &extra, &c, &s, &dummy); } /* Computing MAX */ i__4 = 1, i__5 = jch - jkl; icol = max(i__4,i__5); il = ic + 2 - icol; temp = 0.; iltemp = jch > jkl; d__1 = -s; dlarot_slu(&c_true, &iltemp, &c_true, &il, &c, &d__1, &a[ir - iskew * icol + ioffst + icol * a_dim1], &ilda, &temp, &extra); if (iltemp) { dlartg_slu(&a[ir + 1 - iskew * (icol + 1) + ioffst + (icol + 1) * a_dim1], &temp, &c, &s, &dummy); /* Computing MAX */ i__4 = 1, i__5 = jch - jkl - jku; irow = max(i__4,i__5); il = ir + 2 - irow; extra = 0.; L__1 = jch > jkl + jku; d__1 = -s; dlarot_slu(&c_false, &L__1, &c_true, &il, &c, & d__1, &a[irow - iskew * icol + ioffst + icol * a_dim1], &ilda, &extra, & temp); ic = icol; ir = irow; } /* L60: */ } /* L70: */ } /* L80: */ } } else { /* Bottom-Up -- Start at the bottom right. */ jkl = 0; i__1 = uub; for (jku = 1; jku <= i__1; ++jku) { /* Transform from bandwidth JKL, JKU-1 to JKL, JKU First row actually rotated is M First column actually rotated is MIN( M +JKU, N ) Computing MIN */ i__2 = *m, i__3 = *n + jkl; iendch = min(i__2,i__3) - 1; /* Computing MIN */ i__2 = *m + jku; i__3 = 1 - jkl; for (jc = min(i__2,*n) - 1; jc >= i__3; --jc) { extra = 0.; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = sin(angle); /* Computing MAX */ i__2 = 1, i__4 = jc - jku + 1; irow = max(i__2,i__4); if (jc > 0) { /* Computing MIN */ i__2 = *m, i__4 = jc + jkl + 1; il = min(i__2,i__4) + 1 - irow; L__1 = jc + jkl < *m; dlarot_slu(&c_false, &c_false, &L__1, &il, &c, &s, & a[irow - iskew * jc + ioffst + jc * a_dim1], &ilda, &dummy, &extra); } /* Chase "EXTRA" back down */ ic = jc; i__2 = iendch; i__4 = jkl + jku; for (jch = jc + jkl; i__4 < 0 ? jch >= i__2 : jch <= i__2; jch += i__4) { ilextr = ic > 0; if (ilextr) { dlartg_slu(&a[jch - iskew * ic + ioffst + ic * a_dim1], &extra, &c, &s, &dummy); } ic = max(1,ic); /* Computing MIN */ i__5 = *n - 1, i__6 = jch + jku; icol = min(i__5,i__6); iltemp = jch + jku < *n; temp = 0.; i__5 = icol + 2 - ic; dlarot_slu(&c_true, &ilextr, &iltemp, &i__5, &c, &s, &a[jch - iskew * ic + ioffst + ic * a_dim1], &ilda, &extra, &temp); if (iltemp) { dlartg_slu(&a[jch - iskew * icol + ioffst + icol * a_dim1], &temp, &c, &s, &dummy); /* Computing MIN */ i__5 = iendch, i__6 = jch + jkl + jku; il = min(i__5,i__6) + 2 - jch; extra = 0.; L__1 = jch + jkl + jku <= iendch; dlarot_slu(&c_false, &c_true, &L__1, &il, &c, &s, &a[jch - iskew * icol + ioffst + icol * a_dim1], &ilda, &temp, &extra); ic = icol; } /* L90: */ } /* L100: */ } /* L110: */ } jku = uub; i__1 = llb; for (jkl = 1; jkl <= i__1; ++jkl) { /* Transform from bandwidth JKL-1, JKU to JKL, JKU First row actually rotated is MIN( N+JK L, M ) First column actually rotated is N Computing MIN */ i__3 = *n, i__4 = *m + jku; iendch = min(i__3,i__4) - 1; /* Computing MIN */ i__3 = *n + jkl; i__4 = 1 - jku; for (jr = min(i__3,*m) - 1; jr >= i__4; --jr) { extra = 0.; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = sin(angle); /* Computing MAX */ i__3 = 1, i__2 = jr - jkl + 1; icol = max(i__3,i__2); if (jr > 0) { /* Computing MIN */ i__3 = *n, i__2 = jr + jku + 1; il = min(i__3,i__2) + 1 - icol; L__1 = jr + jku < *n; dlarot_slu(&c_true, &c_false, &L__1, &il, &c, &s, &a[ jr - iskew * icol + ioffst + icol * a_dim1], &ilda, &dummy, &extra); } /* Chase "EXTRA" back down */ ir = jr; i__3 = iendch; i__2 = jkl + jku; for (jch = jr + jku; i__2 < 0 ? jch >= i__3 : jch <= i__3; jch += i__2) { ilextr = ir > 0; if (ilextr) { dlartg_slu(&a[ir - iskew * jch + ioffst + jch * a_dim1], &extra, &c, &s, &dummy); } ir = max(1,ir); /* Computing MIN */ i__5 = *m - 1, i__6 = jch + jkl; irow = min(i__5,i__6); iltemp = jch + jkl < *m; temp = 0.; i__5 = irow + 2 - ir; dlarot_slu(&c_false, &ilextr, &iltemp, &i__5, &c, &s, &a[ir - iskew * jch + ioffst + jch * a_dim1], &ilda, &extra, &temp); if (iltemp) { dlartg_slu(&a[irow - iskew * jch + ioffst + jch * a_dim1], &temp, &c, &s, &dummy); /* Computing MIN */ i__5 = iendch, i__6 = jch + jkl + jku; il = min(i__5,i__6) + 2 - jch; extra = 0.; L__1 = jch + jkl + jku <= iendch; dlarot_slu(&c_true, &c_true, &L__1, &il, &c, &s, &a[irow - iskew * jch + ioffst + jch * a_dim1], &ilda, &temp, &extra); ir = irow; } /* L120: */ } /* L130: */ } /* L140: */ } } } else { /* Symmetric -- A = U D U' */ ipackg = ipack; ioffg = ioffst; if (topdwn) { /* Top-Down -- Generate Upper triangle only */ if (ipack >= 5) { ipackg = 6; ioffg = uub + 1; } else { ipackg = 1; } i__1 = ilda + 1; dcopy_(&mnmin, &d[1], &c__1, &a[1 - iskew + ioffg + a_dim1], & i__1); i__1 = uub; for (k = 1; k <= i__1; ++k) { i__4 = *n - 1; for (jc = 1; jc <= i__4; ++jc) { /* Computing MAX */ i__2 = 1, i__3 = jc - k; irow = max(i__2,i__3); /* Computing MIN */ i__2 = jc + 1, i__3 = k + 2; il = min(i__2,i__3); extra = 0.; temp = a[jc - iskew * (jc + 1) + ioffg + (jc + 1) * a_dim1]; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = sin(angle); L__1 = jc > k; dlarot_slu(&c_false, &L__1, &c_true, &il, &c, &s, &a[ irow - iskew * jc + ioffg + jc * a_dim1], & ilda, &extra, &temp); /* Computing MIN */ i__3 = k, i__5 = *n - jc; i__2 = min(i__3,i__5) + 1; dlarot_slu(&c_true, &c_true, &c_false, &i__2, &c, &s, &a[ (1 - iskew) * jc + ioffg + jc * a_dim1], & ilda, &temp, &dummy); /* Chase EXTRA back up the matrix */ icol = jc; i__2 = -k; for (jch = jc - k; i__2 < 0 ? jch >= 1 : jch <= 1; jch += i__2) { dlartg_slu(&a[jch + 1 - iskew * (icol + 1) + ioffg + (icol + 1) * a_dim1], &extra, &c, &s, & dummy); temp = a[jch - iskew * (jch + 1) + ioffg + (jch + 1) * a_dim1]; i__3 = k + 2; d__1 = -s; dlarot_slu(&c_true, &c_true, &c_true, &i__3, &c, & d__1, &a[(1 - iskew) * jch + ioffg + jch * a_dim1], &ilda, &temp, &extra); /* Computing MAX */ i__3 = 1, i__5 = jch - k; irow = max(i__3,i__5); /* Computing MIN */ i__3 = jch + 1, i__5 = k + 2; il = min(i__3,i__5); extra = 0.; L__1 = jch > k; d__1 = -s; dlarot_slu(&c_false, &L__1, &c_true, &il, &c, &d__1, &a[irow - iskew * jch + ioffg + jch * a_dim1], &ilda, &extra, &temp); icol = jch; /* L150: */ } /* L160: */ } /* L170: */ } /* If we need lower triangle, copy from upper. No te that the order of copying is chosen to work for 'q' -> 'b' */ if (ipack != ipackg && ipack != 3) { i__1 = *n; for (jc = 1; jc <= i__1; ++jc) { irow = ioffst - iskew * jc; /* Computing MIN */ i__2 = *n, i__3 = jc + uub; i__4 = min(i__2,i__3); for (jr = jc; jr <= i__4; ++jr) { a[jr + irow + jc * a_dim1] = a[jc - iskew * jr + ioffg + jr * a_dim1]; /* L180: */ } /* L190: */ } if (ipack == 5) { i__1 = *n; for (jc = *n - uub + 1; jc <= i__1; ++jc) { i__4 = uub + 1; for (jr = *n + 2 - jc; jr <= i__4; ++jr) { a[jr + jc * a_dim1] = 0.; /* L200: */ } /* L210: */ } } if (ipackg == 6) { ipackg = ipack; } else { ipackg = 0; } } } else { /* Bottom-Up -- Generate Lower triangle only */ if (ipack >= 5) { ipackg = 5; if (ipack == 6) { ioffg = 1; } } else { ipackg = 2; } i__1 = ilda + 1; dcopy_(&mnmin, &d[1], &c__1, &a[1 - iskew + ioffg + a_dim1], & i__1); i__1 = uub; for (k = 1; k <= i__1; ++k) { for (jc = *n - 1; jc >= 1; --jc) { /* Computing MIN */ i__4 = *n + 1 - jc, i__2 = k + 2; il = min(i__4,i__2); extra = 0.; temp = a[(1 - iskew) * jc + 1 + ioffg + jc * a_dim1]; angle = dlarnd_slu(&c__1, &iseed[1]) * 6.2831853071795864769252867663; c = cos(angle); s = -sin(angle); L__1 = *n - jc > k; dlarot_slu(&c_false, &c_true, &L__1, &il, &c, &s, &a[(1 - iskew) * jc + ioffg + jc * a_dim1], &ilda, & temp, &extra); /* Computing MAX */ i__4 = 1, i__2 = jc - k + 1; icol = max(i__4,i__2); i__4 = jc + 2 - icol; dlarot_slu(&c_true, &c_false, &c_true, &i__4, &c, &s, &a[ jc - iskew * icol + ioffg + icol * a_dim1], & ilda, &dummy, &temp); /* Chase EXTRA back down the matrix */ icol = jc; i__4 = *n - 1; i__2 = k; for (jch = jc + k; i__2 < 0 ? jch >= i__4 : jch <= i__4; jch += i__2) { dlartg_slu(&a[jch - iskew * icol + ioffg + icol * a_dim1], &extra, &c, &s, &dummy); temp = a[(1 - iskew) * jch + 1 + ioffg + jch * a_dim1]; i__3 = k + 2; dlarot_slu(&c_true, &c_true, &c_true, &i__3, &c, &s, &a[jch - iskew * icol + ioffg + icol * a_dim1], &ilda, &extra, &temp); /* Computing MIN */ i__3 = *n + 1 - jch, i__5 = k + 2; il = min(i__3,i__5); extra = 0.; L__1 = *n - jch > k; dlarot_slu(&c_false, &c_true, &L__1, &il, &c, &s, &a[ (1 - iskew) * jch + ioffg + jch * a_dim1], &ilda, &temp, &extra); icol = jch; /* L220: */ } /* L230: */ } /* L240: */ } /* If we need upper triangle, copy from lower. No te that the order of copying is chosen to work for 'b' -> 'q' */ if (ipack != ipackg && ipack != 4) { for (jc = *n; jc >= 1; --jc) { irow = ioffst - iskew * jc; /* Computing MAX */ i__2 = 1, i__4 = jc - uub; i__1 = max(i__2,i__4); for (jr = jc; jr >= i__1; --jr) { a[jr + irow + jc * a_dim1] = a[jc - iskew * jr + ioffg + jr * a_dim1]; /* L250: */ } /* L260: */ } if (ipack == 6) { i__1 = uub; for (jc = 1; jc <= i__1; ++jc) { i__2 = uub + 1 - jc; for (jr = 1; jr <= i__2; ++jr) { a[jr + jc * a_dim1] = 0.; /* L270: */ } /* L280: */ } } if (ipackg == 5) { ipackg = ipack; } else { ipackg = 0; } } } } } else { /* 4) Generate Banded Matrix by first Rotating by random Unitary matrices, then reducing the bandwidth using Householder transformations. Note: we should get here only if LDA .ge. N */ if (isym == 1) { /* Non-symmetric -- A = U D V */ dlagge_slu(&mr, &nc, &llb, &uub, &d[1], &a[a_offset], lda, &iseed[1], &work[1], &iinfo); } else { /* Symmetric -- A = U D U' */ dlagsy_slu(m, &llb, &d[1], &a[a_offset], lda, &iseed[1], &work[1], & iinfo); } if (iinfo != 0) { *info = 3; return 0; } } /* 5) Pack the matrix */ if (ipack != ipackg) { if (ipack == 1) { /* 'U' -- Upper triangular, not packed */ i__1 = *m; for (j = 1; j <= i__1; ++j) { i__2 = *m; for (i = j + 1; i <= i__2; ++i) { a[i + j * a_dim1] = 0.; /* L290: */ } /* L300: */ } } else if (ipack == 2) { /* 'L' -- Lower triangular, not packed */ i__1 = *m; for (j = 2; j <= i__1; ++j) { i__2 = j - 1; for (i = 1; i <= i__2; ++i) { a[i + j * a_dim1] = 0.; /* L310: */ } /* L320: */ } } else if (ipack == 3) { /* 'C' -- Upper triangle packed Columnwise. */ icol = 1; irow = 0; i__1 = *m; for (j = 1; j <= i__1; ++j) { i__2 = j; for (i = 1; i <= i__2; ++i) { ++irow; if (irow > *lda) { irow = 1; ++icol; } a[irow + icol * a_dim1] = a[i + j * a_dim1]; /* L330: */ } /* L340: */ } } else if (ipack == 4) { /* 'R' -- Lower triangle packed Columnwise. */ icol = 1; irow = 0; i__1 = *m; for (j = 1; j <= i__1; ++j) { i__2 = *m; for (i = j; i <= i__2; ++i) { ++irow; if (irow > *lda) { irow = 1; ++icol; } a[irow + icol * a_dim1] = a[i + j * a_dim1]; /* L350: */ } /* L360: */ } } else if (ipack >= 5) { /* 'B' -- The lower triangle is packed as a band matrix. 'Q' -- The upper triangle is packed as a band matrix. 'Z' -- The whole matrix is packed as a band matrix. */ if (ipack == 5) { uub = 0; } if (ipack == 6) { llb = 0; } i__1 = uub; for (j = 1; j <= i__1; ++j) { /* Computing MIN */ i__2 = j + llb; for (i = min(i__2,*m); i >= 1; --i) { a[i - j + uub + 1 + j * a_dim1] = a[i + j * a_dim1]; /* L370: */ } /* L380: */ } i__1 = *n; for (j = uub + 2; j <= i__1; ++j) { /* Computing MIN */ i__4 = j + llb; i__2 = min(i__4,*m); for (i = j - uub; i <= i__2; ++i) { a[i - j + uub + 1 + j * a_dim1] = a[i + j * a_dim1]; /* L390: */ } /* L400: */ } } /* If packed, zero out extraneous elements. Symmetric/Triangular Packed -- zero out everything after A(IROW,ICOL) */ if (ipack == 3 || ipack == 4) { i__1 = *m; for (jc = icol; jc <= i__1; ++jc) { i__2 = *lda; for (jr = irow + 1; jr <= i__2; ++jr) { a[jr + jc * a_dim1] = 0.; /* L410: */ } irow = 0; /* L420: */ } } else if (ipack >= 5) { /* Packed Band -- 1st row is now in A( UUB+2-j, j), zero above it m-th row is now in A( M+UUB-j,j), zero below it last non-zero diagonal is now in A( UUB+LLB+1,j ), zero below it, too. */ ir1 = uub + llb + 2; ir2 = uub + *m + 2; i__1 = *n; for (jc = 1; jc <= i__1; ++jc) { i__2 = uub + 1 - jc; for (jr = 1; jr <= i__2; ++jr) { a[jr + jc * a_dim1] = 0.; /* L430: */ } /* Computing MAX Computing MIN */ i__3 = ir1, i__5 = ir2 - jc; i__2 = 1, i__4 = min(i__3,i__5); i__6 = *lda; for (jr = max(i__2,i__4); jr <= i__6; ++jr) { a[jr + jc * a_dim1] = 0.; /* L440: */ } /* L450: */ } } } return 0; /* End of DLATMS */ } /* dlatms_slu */