1 files changed, 71 insertions, 0 deletions
diff --git a/Utilib/src/ALDDLF.f b/Utilib/src/ALDDLF.f
new file mode 100644
index 0000000..f690da7
--- /dev/null
+++ b/Utilib/src/ALDDLF.f
@@ -0,0 +1,71 @@
+*DECK ALDDLF
+      SUBROUTINE ALDDLF (L4,ASS,MU1)
+*
+*-----------------------------------------------------------------------
+*
+*Purpose:
+* in-place L-D-L(T) factorization of a symmetric positive definite
+* matrix in compressed diagonal storage mode.
+* Double precision version.
+*
+*Copyright:
+* Copyright (C) 1989 Ecole Polytechnique de Montreal
+* This library is free software; you can redistribute it and/or
+* modify it under the terms of the GNU Lesser General Public
+* License as published by the Free Software Foundation; either
+* version 2.1 of the License, or (at your option) any later version
+*
+*Author(s): A. Hebert
+*
+*Parameters: input
+* L4      order of the coefficient matrix.
+* ASS     coefficient matrix in compressed diagonal storage mode.
+*         A(I,J)=ASS(MU1(I)-I+J) if J.le.I and J.gt.I+MU1(I-1)-MU1(I)
+*               =A(J,I)          if I.lt.J
+*               =0.0             else
+*         DIMENSION ASS(MU1(L4)-MU1(1)+1)
+* MU1     position of each diagonal element in vector ASS.
+*
+*Parameters: output
+* ASS     LDL(T) factors.
+*
+*-----------------------------------------------------------------------
+*
+*----
+*  SUBROUTINE ARGUMENTS
+*----
+      INTEGER L4,MU1(L4)
+      DOUBLE PRECISION ASS(*)
+*----
+*  LOCAL VARIABLES
+*----
+      DOUBLE PRECISION S,R
+*
+      ASS(MU1(1))=1.0D0/ASS(MU1(1))
+      IF (L4.EQ.1) RETURN
+      DO 10 K=2,L4
+      K1=MU1(K)-K
+      KM=MU1(K-1)+1-K1
+      IF(KM+1.GT.K) GO TO 7
+      DO 2 I=KM+1,K-1
+      R=ASS(K1+I)
+      ASS(K1+I)=0.0D0
+      S=0.0D0
+      I1=MU1(I)-I
+      IM=MU1(I-1)+1-I1
+      DO 5 J=MAX0(IM,KM),I
+      S=S+ASS(K1+J)*ASS(I1+J)
+ 5    CONTINUE
+      ASS(K1+I)=R-S
+ 2    CONTINUE
+      S=0.0D0
+      DO 6 I=KM,K-1
+      R=ASS(K1+I)
+      ASS(K1+I)=R*ASS(MU1(I))
+      S=S+R*ASS(K1+I)
+ 6    CONTINUE
+      ASS(MU1(K))=ASS(MU1(K))-S
+ 7    ASS(MU1(K))=1.0D0/ASS(MU1(K))
+ 10   CONTINUE
+      RETURN
+      END