#include "blaswrap.h"
#include "f2c.h"

/* Subroutine */ int slarrf_(integer *n, real *d__, real *l, real *ld, 
	integer *clstrt, integer *clend, real *w, real *wgap, real *werr, 
	real *spdiam, real *clgapl, real *clgapr, real *pivmin, real *sigma, 
	real *dplus, real *lplus, real *work, integer *info)
{
/*  -- LAPACK auxiliary routine (version 3.1) --   
       Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..   
       November 2006   
   *   

    Purpose   
    =======   

    Given the initial representation L D L^T and its cluster of close   
    eigenvalues (in a relative measure), W( CLSTRT ), W( CLSTRT+1 ), ...   
    W( CLEND ), SLARRF finds a new relatively robust representation   
    L D L^T - SIGMA I = L(+) D(+) L(+)^T such that at least one of the   
    eigenvalues of L(+) D(+) L(+)^T is relatively isolated.   

    Arguments   
    =========   

    N       (input) INTEGER   
            The order of the matrix (subblock, if the matrix splitted).   

    D       (input) REAL             array, dimension (N)   
            The N diagonal elements of the diagonal matrix D.   

    L       (input) REAL             array, dimension (N-1)   
            The (N-1) subdiagonal elements of the unit bidiagonal   
            matrix L.   

    LD      (input) REAL             array, dimension (N-1)   
            The (N-1) elements L(i)*D(i).   

    CLSTRT  (input) INTEGER   
            The index of the first eigenvalue in the cluster.   

    CLEND   (input) INTEGER   
            The index of the last eigenvalue in the cluster.   

    W       (input) REAL             array, dimension >=  (CLEND-CLSTRT+1)   
            The eigenvalue APPROXIMATIONS of L D L^T in ascending order.   
            W( CLSTRT ) through W( CLEND ) form the cluster of relatively   
            close eigenalues.   

    WGAP    (input/output) REAL             array, dimension >=  (CLEND-CLSTRT+1)   
            The separation from the right neighbor eigenvalue in W.   

    WERR    (input) REAL             array, dimension >=  (CLEND-CLSTRT+1)   
            WERR contain the semiwidth of the uncertainty   
            interval of the corresponding eigenvalue APPROXIMATION in W   

    SPDIAM (input) estimate of the spectral diameter obtained from the   
            Gerschgorin intervals   

    CLGAPL, CLGAPR (input) absolute gap on each end of the cluster.   
            Set by the calling routine to protect against shifts too close   
            to eigenvalues outside the cluster.   

    PIVMIN  (input) DOUBLE PRECISION   
            The minimum pivot allowed in the Sturm sequence.   

    SIGMA   (output) REAL   
            The shift used to form L(+) D(+) L(+)^T.   

    DPLUS   (output) REAL             array, dimension (N)   
            The N diagonal elements of the diagonal matrix D(+).   

    LPLUS   (output) REAL             array, dimension (N-1)   
            The first (N-1) elements of LPLUS contain the subdiagonal   
            elements of the unit bidiagonal matrix L(+).   

    WORK    (workspace) REAL             array, dimension (2*N)   
            Workspace.   

    Further Details   
    ===============   

    Based on contributions by   
       Beresford Parlett, University of California, Berkeley, USA   
       Jim Demmel, University of California, Berkeley, USA   
       Inderjit Dhillon, University of Texas, Austin, USA   
       Osni Marques, LBNL/NERSC, USA   
       Christof Voemel, University of California, Berkeley, USA   

    =====================================================================   


       Parameter adjustments */
    /* Table of constant values */
    static integer c__1 = 1;
    
    /* System generated locals */
    integer i__1;
    real r__1, r__2, r__3;
    /* Builtin functions */
    double sqrt(doublereal);
    /* Local variables */
    static integer i__;
    static real s, bestshift, smlgrowth, eps, tmp, max1, max2, rrr1, rrr2, 
	    znm2, growthbound, fail, fact, oldp;
    static integer indx;
    static real prod;
    static integer ktry;
    static real fail2, avgap, ldmax, rdmax;
    static integer shift;
    extern /* Subroutine */ int scopy_(integer *, real *, integer *, real *, 
	    integer *);
    static logical dorrr1;
    static real ldelta;
    extern doublereal slamch_(char *);
    static logical nofail;
    static real mingap, lsigma, rdelta;
    static logical forcer;
    static real rsigma, clwdth;
    extern logical sisnan_(real *);
    static logical sawnan1, sawnan2, tryrrr1;


    --work;
    --lplus;
    --dplus;
    --werr;
    --wgap;
    --w;
    --ld;
    --l;
    --d__;

    /* Function Body */
    *info = 0;
    fact = 2.f;
    eps = slamch_("Precision");
    shift = 0;
    forcer = FALSE_;
/*     Note that we cannot guarantee that for any of the shifts tried,   
       the factorization has a small or even moderate element growth.   
       There could be Ritz values at both ends of the cluster and despite   
       backing off, there are examples where all factorizations tried   
       (in IEEE mode, allowing zero pivots & infinities) have INFINITE   
       element growth.   
       For this reason, we should use PIVMIN in this subroutine so that at   
       least the L D L^T factorization exists. It can be checked afterwards   
       whether the element growth caused bad residuals/orthogonality.   
       Decide whether the code should accept the best among all   
       representations despite large element growth or signal INFO=1 */
    nofail = TRUE_;

/*     Compute the average gap length of the cluster */
    clwdth = (r__1 = w[*clend] - w[*clstrt], dabs(r__1)) + werr[*clend] + 
	    werr[*clstrt];
    avgap = clwdth / (real) (*clend - *clstrt);
    mingap = dmin(*clgapl,*clgapr);
/*     Initial values for shifts to both ends of cluster   
   Computing MIN */
    r__1 = w[*clstrt], r__2 = w[*clend];
    lsigma = dmin(r__1,r__2) - werr[*clstrt];
/* Computing MAX */
    r__1 = w[*clstrt], r__2 = w[*clend];
    rsigma = dmax(r__1,r__2) + werr[*clend];
/*     Use a small fudge to make sure that we really shift to the outside */
    lsigma -= dabs(lsigma) * 2.f * eps;
    rsigma += dabs(rsigma) * 2.f * eps;
/*     Compute upper bounds for how much to back off the initial shifts */
    ldmax = mingap * .25f + *pivmin * 2.f;
    rdmax = mingap * .25f + *pivmin * 2.f;
/* Computing MAX */
    r__1 = avgap, r__2 = wgap[*clstrt];
    ldelta = dmax(r__1,r__2) / fact;
/* Computing MAX */
    r__1 = avgap, r__2 = wgap[*clend - 1];
    rdelta = dmax(r__1,r__2) / fact;

/*     Initialize the record of the best representation found */

    s = slamch_("S");
    smlgrowth = 1.f / s;
    fail = (real) (*n - 1) * mingap / (*spdiam * eps);
    fail2 = (real) (*n - 1) * mingap / (*spdiam * sqrt(eps));
    bestshift = lsigma;

/*     while (KTRY <= KTRYMAX) */
    ktry = 0;
    growthbound = *spdiam * 8.f;
L5:
    sawnan1 = FALSE_;
    sawnan2 = FALSE_;
/*     Ensure that we do not back off too much of the initial shifts */
    ldelta = dmin(ldmax,ldelta);
    rdelta = dmin(rdmax,rdelta);
/*     Compute the element growth when shifting to both ends of the cluster   
       accept the shift if there is no element growth at one of the two ends   
       Left end */
    s = -lsigma;
    dplus[1] = d__[1] + s;
    if (dabs(dplus[1]) < *pivmin) {
	dplus[1] = -(*pivmin);
/*        Need to set SAWNAN1 because refined RRR test should not be used   
          in this case */
	sawnan1 = TRUE_;
    }
    max1 = dabs(dplus[1]);
    i__1 = *n - 1;
    for (i__ = 1; i__ <= i__1; ++i__) {
	lplus[i__] = ld[i__] / dplus[i__];
	s = s * lplus[i__] * l[i__] - lsigma;
	dplus[i__ + 1] = d__[i__ + 1] + s;
	if ((r__1 = dplus[i__ + 1], dabs(r__1)) < *pivmin) {
	    dplus[i__ + 1] = -(*pivmin);
/*           Need to set SAWNAN1 because refined RRR test should not be used   
             in this case */
	    sawnan1 = TRUE_;
	}
/* Computing MAX */
	r__2 = max1, r__3 = (r__1 = dplus[i__ + 1], dabs(r__1));
	max1 = dmax(r__2,r__3);
/* L6: */
    }
    sawnan1 = sawnan1 || sisnan_(&max1);
    if (forcer || max1 <= growthbound && ! sawnan1) {
	*sigma = lsigma;
	shift = 1;
	goto L100;
    }
/*     Right end */
    s = -rsigma;
    work[1] = d__[1] + s;
    if (dabs(work[1]) < *pivmin) {
	work[1] = -(*pivmin);
/*        Need to set SAWNAN2 because refined RRR test should not be used   
          in this case */
	sawnan2 = TRUE_;
    }
    max2 = dabs(work[1]);
    i__1 = *n - 1;
    for (i__ = 1; i__ <= i__1; ++i__) {
	work[*n + i__] = ld[i__] / work[i__];
	s = s * work[*n + i__] * l[i__] - rsigma;
	work[i__ + 1] = d__[i__ + 1] + s;
	if ((r__1 = work[i__ + 1], dabs(r__1)) < *pivmin) {
	    work[i__ + 1] = -(*pivmin);
/*           Need to set SAWNAN2 because refined RRR test should not be used   
             in this case */
	    sawnan2 = TRUE_;
	}
/* Computing MAX */
	r__2 = max2, r__3 = (r__1 = work[i__ + 1], dabs(r__1));
	max2 = dmax(r__2,r__3);
/* L7: */
    }
    sawnan2 = sawnan2 || sisnan_(&max2);
    if (forcer || max2 <= growthbound && ! sawnan2) {
	*sigma = rsigma;
	shift = 2;
	goto L100;
    }
/*     If we are at this point, both shifts led to too much element growth   
       Record the better of the two shifts (provided it didn't lead to NaN) */
    if (sawnan1 && sawnan2) {
/*        both MAX1 and MAX2 are NaN */
	goto L50;
    } else {
	if (! sawnan1) {
	    indx = 1;
	    if (max1 <= smlgrowth) {
		smlgrowth = max1;
		bestshift = lsigma;
	    }
	}
	if (! sawnan2) {
	    if (sawnan1 || max2 <= max1) {
		indx = 2;
	    }
	    if (max2 <= smlgrowth) {
		smlgrowth = max2;
		bestshift = rsigma;
	    }
	}
    }
/*     If we are here, both the left and the right shift led to   
       element growth. If the element growth is moderate, then   
       we may still accept the representation, if it passes a   
       refined test for RRR. This test supposes that no NaN occurred.   
       Moreover, we use the refined RRR test only for isolated clusters. */
    if (clwdth < mingap / 128.f && dmin(max1,max2) < fail2 && ! sawnan1 && ! 
	    sawnan2) {
	dorrr1 = TRUE_;
    } else {
	dorrr1 = FALSE_;
    }
    tryrrr1 = TRUE_;
    if (tryrrr1 && dorrr1) {
	if (indx == 1) {
	    tmp = (r__1 = dplus[*n], dabs(r__1));
	    znm2 = 1.f;
	    prod = 1.f;
	    oldp = 1.f;
	    for (i__ = *n - 1; i__ >= 1; --i__) {
		if (prod <= eps) {
		    prod = dplus[i__ + 1] * work[*n + i__ + 1] / (dplus[i__] *
			     work[*n + i__]) * oldp;
		} else {
		    prod *= (r__1 = work[*n + i__], dabs(r__1));
		}
		oldp = prod;
/* Computing 2nd power */
		r__1 = prod;
		znm2 += r__1 * r__1;
/* Computing MAX */
		r__2 = tmp, r__3 = (r__1 = dplus[i__] * prod, dabs(r__1));
		tmp = dmax(r__2,r__3);
/* L15: */
	    }
	    rrr1 = tmp / (*spdiam * sqrt(znm2));
	    if (rrr1 <= 8.f) {
		*sigma = lsigma;
		shift = 1;
		goto L100;
	    }
	} else if (indx == 2) {
	    tmp = (r__1 = work[*n], dabs(r__1));
	    znm2 = 1.f;
	    prod = 1.f;
	    oldp = 1.f;
	    for (i__ = *n - 1; i__ >= 1; --i__) {
		if (prod <= eps) {
		    prod = work[i__ + 1] * lplus[i__ + 1] / (work[i__] * 
			    lplus[i__]) * oldp;
		} else {
		    prod *= (r__1 = lplus[i__], dabs(r__1));
		}
		oldp = prod;
/* Computing 2nd power */
		r__1 = prod;
		znm2 += r__1 * r__1;
/* Computing MAX */
		r__2 = tmp, r__3 = (r__1 = work[i__] * prod, dabs(r__1));
		tmp = dmax(r__2,r__3);
/* L16: */
	    }
	    rrr2 = tmp / (*spdiam * sqrt(znm2));
	    if (rrr2 <= 8.f) {
		*sigma = rsigma;
		shift = 2;
		goto L100;
	    }
	}
    }
L50:
    if (ktry < 1) {
/*        If we are here, both shifts failed also the RRR test.   
          Back off to the outside   
   Computing MAX */
	r__1 = lsigma - ldelta, r__2 = lsigma - ldmax;
	lsigma = dmax(r__1,r__2);
/* Computing MIN */
	r__1 = rsigma + rdelta, r__2 = rsigma + rdmax;
	rsigma = dmin(r__1,r__2);
	ldelta *= 2.f;
	rdelta *= 2.f;
	++ktry;
	goto L5;
    } else {
/*        None of the representations investigated satisfied our   
          criteria. Take the best one we found. */
	if (smlgrowth < fail || nofail) {
	    lsigma = bestshift;
	    rsigma = bestshift;
	    forcer = TRUE_;
	    goto L5;
	} else {
	    *info = 1;
	    return 0;
	}
    }
L100:
    if (shift == 1) {
    } else if (shift == 2) {
/*        store new L and D back into DPLUS, LPLUS */
	scopy_(n, &work[1], &c__1, &dplus[1], &c__1);
	i__1 = *n - 1;
	scopy_(&i__1, &work[*n + 1], &c__1, &lplus[1], &c__1);
    }
    return 0;

/*     End of SLARRF */

} /* slarrf_ */