>>>>>>>>>>>>> FILELIST <<<<<<<<<<<<< >>>>>>>>>>>>> README <<<<<<<<<<<<< RCS file: RCS/README,v Working file: README description: README file for Meschach v. 1.2a and onwards ---------------------------- revision 1.1 date: 1994/01/13 04:14:16; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> arnoldi.c <<<<<<<<<<<<< RCS file: RCS/arnoldi.c,v Working file: arnoldi.c description: Arnoldi method for finding eigenvalues of large matrices via a Krylov subspace technique. ---------------------------- revision 1.3 date: 1994/01/13 05:45:40; author: des; state: Exp; lines: +132 -9 Version 1.2 update -- changed names etc not to be maintained in future ---------------------------- revision 1.2 date: 1991/10/25 02:16:36; author: des; state: Exp; lines: +19 -11 Modified to conform to changes in norm.c: n1() => norm1() etc ---------------------------- revision 1.1 date: 1991/09/15 23:06:22; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> bdfactor.c <<<<<<<<<<<<< RCS file: RCS/bdfactor.c,v Working file: bdfactor.c description: Band matrix routines -- utility and factorise/solve routines ---------------------------- revision 1.7 date: 1994/09/02 18:37:39; author: des; state: Exp; lines: +3 -2 Bug fix by Zbigniew Leyk. ---------------------------- revision 1.6 date: 1994/08/25 15:21:00; author: des; state: Exp; lines: +56 -2 Added bd_mv_mlt() to multiply a banded matrix by a vector. Addition contributed by Ruediger Franke of the Technical University of Ilmenau, Germany. ---------------------------- revision 1.5 date: 1994/05/17 23:18:32; author: des; state: Exp; lines: +3 -2 MEM_COPY() call in bd_transp() has "*sizeof(Real)" added. Bug noticed by Hugh Secker-Walker ---------------------------- revision 1.4 date: 1994/02/14 00:38:27; author: des; state: Exp; lines: +6 -5 Used Zbigniew's update to bd_copy() --- uses bd_get() if B == MNULL ---------------------------- revision 1.3 date: 1994/02/14 00:33:47; author: des; state: Exp; lines: +1 -1 RCS Id string now setup properly ---------------------------- revision 1.2 date: 1994/02/13 21:47:53; author: des; state: Exp; lines: +1 -1 Fixed bug in bd_copy(A,B) for B == NULL Bug noted by Scott Bigham (dsb@cs.duke.edu) ---------------------------- revision 1.1 date: 1994/01/13 05:36:04; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> bkpfacto.c <<<<<<<<<<<<< RCS file: RCS/bkpfacto.c,v Working file: bkpfacto.c description: Bunch-Kaufman-Parlett symmetric indefinite factorise and solve ---------------------------- revision 1.7 date: 1994/01/13 05:45:50; author: des; state: Exp; lines: +112 -69 Version 1.2 update -- changed names etc ---------------------------- revision 1.6 date: 1992/10/29 06:12:23; author: des; state: Exp; lines: +7 -2 BKPsolve() now raises E_SING error if a 1 x 1 block is singular Before it was only for the singular 2 x 2 blocks ---------------------------- revision 1.5 date: 1992/10/22 01:19:16; author: des; state: Exp; lines: +4 -4 BKPsolve() modified to use both upper and lower triangular parts of the factored BKP form to get stride 1. ---------------------------- revision 1.4 date: 1992/10/22 01:15:03; author: des; state: Exp; lines: +27 -22 Now does the transposed operations in BKPfactor() to get mainly stride 1 ops however, both lower & upper triangular parts are modified ---------------------------- revision 1.3 date: 1992/03/22 23:13:48; author: des; state: Exp; lines: +18 -17 Fixed bugs in BKPfactor() and BKPsolve() -- don't know where they came from These were true bugs in the algorithms! ---------------------------- revision 1.2 date: 1992/02/17 05:13:20; author: des; state: Exp; lines: +10 -10 Removed printf statements ---------------------------- revision 1.1 date: 1991/08/14 00:21:03; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> chfactor.c <<<<<<<<<<<<< RCS file: RCS/chfactor.c,v Working file: chfactor.c description: Cholseky and LDL^T factorise and solve routines ---------------------------- revision 1.2 date: 1994/01/13 05:36:36; author: des; state: Exp; lines: +36 -16 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> configgnu <<<<<<<<<<<<< RCS file: RCS/configgnu,v Working file: configgnu description: Configuration script for using gcc ---------------------------- revision 1.1 date: 1994/01/14 01:33:40; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> configure <<<<<<<<<<<<< RCS file: RCS/configure,v Working file: configure description: Configuration script, generated using autoconf and configure.in ---------------------------- revision 1.1 date: 1994/01/13 05:53:55; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> configure.in <<<<<<<<<<<<< RCS file: RCS/configure.in,v Working file: configure.in description: Input file for constructing configure (and configgnu) using autoconf ---------------------------- revision 1.3 date: 1994/03/08 05:41:32; author: des; state: Exp; lines: +14 -14 Replaced the obsolete AC_MEMORY_H with AC_HAVE_HEADERS(memory.h) ---------------------------- revision 1.2 date: 1994/02/22 00:16:30; author: des; state: Exp; lines: +14 -1 Added Brook Milligan's check for prototypes in structures Creates macro HAVE_PROTOTYPES_IN_STRUCT for machine.h[.in] ---------------------------- revision 1.1 date: 1994/01/13 05:47:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> conjgrad.c <<<<<<<<<<<<< RCS file: RCS/conjgrad.c,v Working file: conjgrad.c description: Conjugate gradient type routines Includes pre-conditioned CG, CGS and LSQR ---------------------------- revision 1.4 date: 1994/01/13 05:36:45; author: des; state: Exp; lines: +74 -44 Version 1.2 update -- uses new names for routines not to be maintained in future ---------------------------- revision 1.3 date: 1992/03/19 02:08:51; author: des; state: Exp; lines: +10 -10 Changed use of norm2() to v_norm2() to conform to norm.c ---------------------------- revision 1.2 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +10 -10 Modified to reflect change in norm.c: n1() => norm1(), n2() => norm2() etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> copy.c <<<<<<<<<<<<< RCS file: RCS/copy.c,v Working file: copy.c description: Copy routines for basic data structures ---------------------------- revision 1.2 date: 1994/01/13 05:37:14; author: des; state: Exp; lines: +163 -34 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> copyright <<<<<<<<<<<<< RCS file: RCS/copyright,v Working file: copyright description: Copyright notice for Meschach ---------------------------- revision 1.1 date: 1994/01/13 05:54:22; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> dmacheps.c <<<<<<<<<<<<< RCS file: RCS/dmacheps.c,v Working file: dmacheps.c description: Computes double precision machine epsilon ---------------------------- revision 1.1 date: 1994/01/13 05:37:18; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> err.c <<<<<<<<<<<<< RCS file: RCS/err.c,v Working file: err.c description: Error handling/reporting routines ---------------------------- revision 1.6 date: 1995/01/30 14:49:14; author: des; state: Exp; lines: +28 -26 Added E_BREAKDOWN error for the iterative routines. Modified by Zbigniew Leyk ---------------------------- revision 1.5 date: 1994/01/13 05:37:35; author: des; state: Exp; lines: +273 -108 Version 1.2 update -- uses new names for routines and new functionality -- has extendable error handling now ---------------------------- revision 1.4 date: 1993/07/19 01:45:11; author: des; state: Exp; lines: +8 -3 Added "#define ERR_C" so that err.c contains the "master" line[] array. See matdef.h revision 1.4 log. ---------------------------- revision 1.3 date: 1991/11/11 06:35:55; author: des; state: Exp; lines: +14 -3 Added count_errs() function to set/unset error counting inside ev_err(). ---------------------------- revision 1.2 date: 1991/10/16 06:47:18; author: des; state: Exp; lines: +3 -2 Added "internal inconsistency (data structure)" error (E_INTERN) ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> err.h <<<<<<<<<<<<< RCS file: RCS/err.h,v Working file: err.h description: Version 1.2 update -- uses new names for routines See also err.c ---------------------------- revision 1.2 date: 1995/01/30 14:48:05; author: des; state: Exp; lines: +4 -2 Added E_BREAKDOWN error for the iterative routines. ---------------------------- revision 1.1 date: 1994/01/13 05:38:03; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> extras.c <<<<<<<<<<<<< RCS file: RCS/extras.c,v Working file: extras.c description: Extra (support) routines -- mem_copy() and mem_zero() -- BLAS-style routines ---------------------------- revision 1.3 date: 1994/01/13 05:45:36; author: des; state: Exp; lines: +110 -47 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1992/04/29 06:58:09; author: des; state: Exp; lines: +1 -1 Just getting the header field right ---------------------------- revision 1.1 date: 1992/04/29 06:56:15; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> fft.c <<<<<<<<<<<<< RCS file: RCS/fft.c,v Working file: fft.c description: Fast Fourier Transform routine radix 2 only! ---------------------------- revision 1.3 date: 1994/01/13 05:45:33; author: des; state: Exp; lines: +35 -7 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1993/05/22 03:49:37; author: des; state: Exp; lines: +11 -1 Added ifft() to compute the inverse FFT (incl. scaling) ---------------------------- revision 1.1 date: 1993/05/10 02:33:00; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> fmacheps.c <<<<<<<<<<<<< RCS file: RCS/fmacheps.c,v Working file: fmacheps.c description: Computes single precision floating point machine epsilon ---------------------------- revision 1.1 date: 1994/01/13 05:39:24; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> givens.c <<<<<<<<<<<<< RCS file: RCS/givens.c,v Working file: givens.c description: Givens' rotations' operations ---------------------------- revision 1.2 date: 1994/01/13 05:39:42; author: des; state: Exp; lines: +57 -20 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> hessen.c <<<<<<<<<<<<< RCS file: RCS/hessen.c,v Working file: hessen.c description: Hessenberg factorisation This is preliminary to Schur decomposition: see also schur.c ---------------------------- revision 1.2 date: 1994/01/13 05:36:24; author: des; state: Exp; lines: +54 -19 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> hsehldr.c <<<<<<<<<<<<< RCS file: RCS/hsehldr.c,v Working file: hsehldr.c description: Householder operations: see also qrfactor.c ---------------------------- revision 1.2 date: 1994/01/13 05:36:29; author: des; state: Exp; lines: +38 -11 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> init.c <<<<<<<<<<<<< RCS file: RCS/init.c,v Working file: init.c description: Initialisation and resetting routines Initialises basic data structures to zeros, ones, id. matrix etc ---------------------------- revision 1.6 date: 1994/01/13 05:36:58; author: des; state: Exp; lines: +123 -85 Version 1.2 update -- uses new names for routines -- also added new initialisation routines ---------------------------- revision 1.5 date: 1992/10/15 05:07:18; author: des; state: Exp; lines: +34 -7 Added mrandlist(), and removed "%" operations which slow it down on RS6000's ---------------------------- revision 1.4 date: 1992/10/15 04:05:33; author: des; state: Exp; lines: +93 -3 Added own pseudo-random number generator routines: mrand() returning double's in range [0,1) smrand(int) for initialising or resetting the mrand() seed(s). ---------------------------- revision 1.3 date: 1992/10/13 01:16:31; author: des; state: Exp; lines: +3 -3 Modified rand_vec() and rand_mat() so that MAX_RAND is converted to double before doing division. ---------------------------- revision 1.2 date: 1992/02/28 03:12:23; author: des; state: Exp; lines: +1 -2 Removed the doubling up of #include "matdef.h" and #include "machine.h" ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> iotort.c <<<<<<<<<<<<< RCS file: RCS/iotort.c,v Working file: iotort.c des: 1.1 description: Input/output torture test ---------------------------- revision 1.1 locked by: des; date: 1994/01/13 05:38:15; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> iter.h <<<<<<<<<<<<< RCS file: RCS/iter.h,v Working file: iter.h description: Header file for iterative routines; see also iter0.c ---------------------------- revision 1.2 date: 1994/03/08 05:48:27; author: des; state: Exp; lines: +7 -2 Added #ifdef PROTOTYPES_IN_STRUCT wrappers for the ITER data structure. ---------------------------- revision 1.1 date: 1994/01/13 05:38:34; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> iter0.c <<<<<<<<<<<<< RCS file: RCS/iter0.c,v Working file: iter0.c description: Basic iterative utility routines ---------------------------- revision 1.3 date: 1995/01/30 14:50:56; author: des; state: Exp; lines: +18 -26 Update of iterative routines by Zbigniew Leyk ---------------------------- revision 1.2 date: 1994/06/20 05:22:10; author: des; state: Exp; lines: +19 -11 Fixed bug in iter-resize() --- Bug noticed and fixed by Zbigniew Leyk ---------------------------- revision 1.1 date: 1994/01/13 05:38:23; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> iternsym.c <<<<<<<<<<<<< RCS file: RCS/iternsym.c,v Working file: iternsym.c description: Non-symmetric iterative routines ---------------------------- revision 1.6 date: 1995/01/30 14:53:01; author: des; state: Exp; lines: +96 -67 Added E_BREAKDOWN error, and corrected bug in iter_gmres() due to exactly finding an invariant Krylov subspace. Related bugs fixed in iter_mgcr(). Modified by Zbigniew Leyk ---------------------------- revision 1.5 date: 1995/01/09 22:39:00; author: des; state: Exp; lines: +6 -9 Minor patch to iter_gmres() to prevent premature exit for small matrices ---------------------------- revision 1.4 date: 1994/10/12 23:23:04; author: des; state: Exp; lines: +1 -14 Deleted mail header ---------------------------- revision 1.3 date: 1994/09/20 17:05:56; author: des; state: Exp; lines: +46 -28 Changed use of Gram-Schmidt to modified GRam Schmidt in arnoldi() ---------------------------- revision 1.2 date: 1994/02/02 06:02:20; author: des; state: Exp; lines: +11 -3 Added test foir h_rem <= 0.0 to avoid divide by zero using 1.0/h_rem ---------------------------- revision 1.1 date: 1994/01/13 05:39:10; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> itersym.c <<<<<<<<<<<<< RCS file: RCS/itersym.c,v Working file: itersym.c description: Symmetric matrix iterative routines ---------------------------- revision 1.2 date: 1995/01/30 14:55:54; author: des; state: Exp; lines: +13 -8 Added E_BREAKDOWN error. Also fixed iter_cg() to test for finding an exactly invariant Krylov subspace. Modified by Zbigniew Leyk. ---------------------------- revision 1.1 date: 1994/01/13 05:38:59; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> itertort.c <<<<<<<<<<<<< RCS file: RCS/itertort.c,v Working file: itertort.c des: 1.1 description: Torture test for iterative routines ---------------------------- revision 1.1 locked by: des; date: 1994/01/13 05:38:50; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> ivecop.c <<<<<<<<<<<<< RCS file: RCS/ivecop.c,v Working file: ivecop.c description: Integer vector data structure and operations ---------------------------- revision 1.5 date: 1994/01/13 05:45:30; author: des; state: Exp; lines: +350 -283 Version 1.2 update -- changed names etc ---------------------------- revision 1.4 date: 1993/08/27 06:44:22; author: des; state: Exp; lines: +3 -1 Added static line[MAXLINE] declaration to avoid name collisions. ---------------------------- revision 1.3 date: 1993/05/10 02:10:35; author: des; state: Exp; lines: +108 -5 Modified fout_ivec() to print "IntVector: dim: " instead of just "IntVector: " ---------------------------- revision 1.2 date: 1991/11/01 05:35:05; author: des; state: Exp; lines: +17 -1 Added cp_ivec() routine; does not have any i0 parameters unlike cp_vec(). ---------------------------- revision 1.1 date: 1991/10/26 03:09:41; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> lanczos.c <<<<<<<<<<<<< RCS file: RCS/lanczos.c,v Working file: lanczos.c description: Basic Lanczos and Cullum & Willoughby type Lanczos routines ---------------------------- revision 1.4 date: 1994/01/13 05:28:24; author: des; state: Exp; lines: +53 -21 Version 1.2 update -- changed names etc not to be maintained ---------------------------- revision 1.3 date: 1992/02/16 22:00:27; author: des; state: Exp; lines: +1 -1 Replaced HUGE with HUGE_VAL ---------------------------- revision 1.2 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +2 -2 Modified to reflect change in norm.c: n1() => norm1(), n2() => norm2() etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> ls.dat <<<<<<<<<<<<< RCS file: RCS/ls.dat,v Working file: ls.dat description: Data file for least squares tutorial problem; see tutorial.c ---------------------------- revision 1.1 date: 1994/01/13 05:47:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> lufactor.c <<<<<<<<<<<<< RCS file: RCS/lufactor.c,v Working file: lufactor.c description: LU factorise and solve routines ---------------------------- revision 1.7 date: 1994/03/13 23:08:25; author: des; state: Exp; lines: +2 -2 LUcondest() now returns HUGE_VAL instead of HUGE for an exactly singular matrix. ---------------------------- revision 1.6 date: 1994/01/13 05:28:42; author: des; state: Exp; lines: +101 -82 Version 1.2 update -- changed names etc ---------------------------- revision 1.5 date: 1992/02/16 22:00:09; author: des; state: Exp; lines: +2 -2 Replaced HUGE with HUGE_VAL ---------------------------- revision 1.4 date: 1992/01/09 06:05:54; author: des; state: Exp; lines: +15 -6 Improved LUcondest() ---------------------------- revision 1.3 date: 1992/01/09 05:49:58; author: des; state: Exp; lines: +15 -3 Fixed minor problem in LUcondest() ---------------------------- revision 1.2 date: 1992/01/09 05:06:29; author: des; state: Exp; lines: +39 -1 Added LUcondest() for estimating condition numbers ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.c <<<<<<<<<<<<< RCS file: RCS/machine.c,v Working file: machine.c description: Machine dependent code. This code is the basis of most of the basic dense matrix/vector operations. Change this file for "tuning" to a particular machine/compiler. ---------------------------- revision 1.4 date: 1994/01/13 05:28:56; author: des; state: Exp; lines: +65 -132 Version 1.2 update -- added the "#ifdef VUNROLL" parts for selective unrolling of loops ---------------------------- revision 1.3 date: 1991/08/29 06:42:11; author: des; state: Exp; lines: +13 -2 Used loop unrolling on __mltadd__() ---------------------------- revision 1.2 date: 1991/08/29 06:15:24; author: des; state: Exp; lines: +157 -28 Modified __ip__() and added __ip4__(), __lc4__() and __ma4__() to take help take advantage of vector & super-scalar architectures. For full advantage would need to re-write matop.c, lufactor.c etc etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h.in <<<<<<<<<<<<< RCS file: RCS/machine.h.in,v Working file: machine.h.in description: Input file for constructing machine.h ---------------------------- revision 1.2 date: 1994/03/13 23:07:30; author: des; state: Exp; lines: +31 -1 Now defines both HUGE and HUGE_VAL if either is defined. ---------------------------- revision 1.1 date: 1994/01/13 05:47:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile.in <<<<<<<<<<<<< RCS file: RCS/makefile.in,v Working file: makefile.in description: Input file for constructing makefile using configure ---------------------------- revision 1.4 date: 1994/03/14 01:24:06; author: des; state: Exp; lines: +2 -2 Now produces distribution names "mesch12b" instead of "mesch12a" ---------------------------- revision 1.3 date: 1994/03/08 05:42:29; author: des; state: Exp; lines: +13 -4 Replaced remaining ranlib's with $(RANLIB), and added realclean target and dependencies amongst the *part* targets. Changes suggested by Blair Zajac. ---------------------------- revision 1.2 date: 1994/01/14 05:49:40; author: des; state: Exp; lines: +19 -6 Now "make list" produces a file FILELIST which lists all Meschach files ---------------------------- revision 1.1 date: 1994/01/13 05:47:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matlab.c <<<<<<<<<<<<< RCS file: RCS/matlab.c,v Working file: matlab.c description: Matlab save/load facility routines ---------------------------- revision 1.7 date: 1994/01/13 05:30:09; author: des; state: Exp; lines: +40 -13 Version 1.2 update -- changed names etc ---------------------------- revision 1.6 date: 1993/05/31 05:27:04; author: des; state: Exp; lines: +30 -1 Added d_save() which stores a single double precision number a la Matlab. ---------------------------- revision 1.5 date: 1993/05/31 00:29:11; author: des; state: Exp; lines: +3 -3 Vectors are now saved as **column**, not row, vectors ---------------------------- revision 1.4 date: 1993/03/30 23:03:46; author: des; state: Exp; lines: +6 -36 No change ---------------------------- revision 1.3 date: 1992/01/20 01:37:26; author: des; state: Exp; lines: +3 -1 Wrapped the #define MACH_ID so that it can be #define'd in machine.h. ---------------------------- revision 1.2 date: 1991/10/03 23:43:37; author: des; state: Exp; lines: +49 -53 Fixed problem with malloc() declarations etc ---------------------------- revision 1.1 date: 1991/08/14 00:20:38; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matlab.h <<<<<<<<<<<<< RCS file: RCS/matlab.h,v Working file: matlab.h description: Header file for MATLAB machine types etc ---------------------------- revision 1.2 date: 1994/01/13 05:44:47; author: des; state: Exp; lines: +75 -5 Version 1.2 update -- changed names etc Uses new scheme for dealing with single vs double precision and machine characteristics as determined by configure ---------------------------- revision 1.1 date: 1993/05/10 02:24:09; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matop.c <<<<<<<<<<<<< RCS file: RCS/matop.c,v Working file: matop.c description: Basic matrix operations: see also vecop.c ---------------------------- revision 1.3 date: 1994/01/13 05:30:28; author: des; state: Exp; lines: +53 -40 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1992/01/23 22:57:49; author: des; state: Exp; lines: +14 -9 Modified mv_mltadd() and vm_mltadd() to take a scale factor as an argument ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matrix.h <<<<<<<<<<<<< RCS file: RCS/matrix.h,v Working file: matrix.h description: Include file for the basic routines ---------------------------- revision 1.18 date: 1994/04/16 00:33:37; author: des; state: Exp; lines: +5 -5 Fixed v_add_val(), v_sub_val(), m_add_val() and m_sub_val() macros so that the function in the error() call is correct. Previously the calls were to error(..., ".._set_val"). Bug noticed by Tony Rossini, Harvard Uni. ---------------------------- revision 1.17 date: 1994/03/08 05:49:16; author: des; state: Exp; lines: +1 -1 Added #ifdef PROTOTYPES_IN_STRUCT wrappers for v_map() and _v_map(). ---------------------------- revision 1.16 date: 1994/03/08 05:47:14; author: des; state: Exp; lines: +12 -1 Now declares m_version() -- identified by lint output from Blair Zajac. ---------------------------- revision 1.15 date: 1994/01/13 05:30:42; author: des; state: Exp; lines: +449 -208 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.14 date: 1993/05/31 05:28:36; author: des; state: Exp; lines: +65 -15 Added d_save() which stores a single double precision number a la Matlab. ---------------------------- revision 1.13 date: 1992/03/19 06:49:53; author: des; state: Exp; lines: +5 -1 Added m_load(), m_save() and v_save() declatatrations ---------------------------- revision 1.12 date: 1992/03/19 06:04:21; author: des; state: Exp; lines: +2 -2 Added m_norm1() declaration for ANSI_C ---------------------------- revision 1.11 date: 1992/01/31 02:52:39; author: des; state: Exp; lines: +2 -2 Replaced __STDC__ in previous modification with ANSI_C ---------------------------- revision 1.10 date: 1992/01/31 02:42:28; author: des; state: Exp; lines: +11 -1 Now uses (size_t) casts for memory allocation macros if __STDC__ is defined. ---------------------------- revision 1.9 date: 1992/01/23 23:42:16; author: des; state: Exp; lines: +3 -3 Updated to match mv_mltadd() and vm_mltadd() changes in matop.c ---------------------------- revision 1.8 date: 1992/01/20 01:27:39; author: des; state: Exp; lines: +3 -3 Changed READ() and FREAD() macros to input() and finput() macros. ---------------------------- revision 1.7 date: 1991/11/11 06:36:31; author: des; state: Exp; lines: +3 -1 Added count_errs() function to set/unset error counting inside ev_err(). ---------------------------- revision 1.6 date: 1991/11/01 05:38:55; author: des; state: Exp; lines: +9 -1 Added IVEC add & subtract operations ---------------------------- revision 1.5 date: 1991/11/01 05:35:34; author: des; state: Exp; lines: +14 -4 Added IVEC routines ---------------------------- revision 1.4 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +9 -7 Modified to reflect change in norm.c: n1() => norm1(), n2() => norm2() etc ---------------------------- revision 1.3 date: 1991/10/16 06:48:57; author: des; state: Exp; lines: +3 -2 Added Added "internal inconsistency (data structure)" error (E_INTERN) ---------------------------- revision 1.2 date: 1991/10/03 23:40:50; author: des; state: Exp; lines: +13 -6 Modified to use MALLOCDECL macro (see: machine.h) ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matrix2.h <<<<<<<<<<<<< RCS file: RCS/matrix2.h,v Working file: matrix2.h description: Include file for the matrix factorisation routines ---------------------------- revision 1.7 date: 1994/01/13 05:30:58; author: des; state: Exp; lines: +182 -43 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.6 date: 1993/06/09 23:04:24; author: des; state: Exp; lines: +10 -8 Add m_inverse() to declarations ---------------------------- revision 1.5 date: 1992/12/09 02:30:18; author: des; state: Exp; lines: +6 -5 Modified QRsolve(), makeQ() routines not to use a beta vector Also added QRTsolve(). ---------------------------- revision 1.4 date: 1992/01/21 03:07:20; author: des; state: Exp; lines: +11 -7 Added BKPfactor(), BKPsolve(), schur_vecs(). ---------------------------- revision 1.3 date: 1992/01/09 05:17:10; author: des; state: Exp; lines: +2 -0 Added LUcondest() header info ---------------------------- revision 1.2 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +0 -7 Modified to reflect change in norm.c: n1() => norm1(), n2() => norm2() etc ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> matrixio.c <<<<<<<<<<<<< RCS file: RCS/matrixio.c,v Working file: matrixio.c description: Input/output for basic data structures ---------------------------- revision 1.4 date: 1994/01/13 05:31:10; author: des; state: Exp; lines: +91 -57 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.3 date: 1993/08/27 06:42:51; author: des; state: Exp; lines: +3 -1 Added static line[MAXLINE] declaration to avoid name collisions. ---------------------------- revision 1.2 date: 1993/06/14 22:30:43; author: des; state: Exp; lines: +131 -124 Functions bfin_mat() and bfin_vec() now will use the input file's dimensions. This is at least consistent with the documentation. ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> maxint.c <<<<<<<<<<<<< RCS file: RCS/maxint.c,v Working file: maxint.c description: Computes the maximum representable int ---------------------------- revision 1.1 date: 1994/01/13 05:31:17; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> meminfo.c <<<<<<<<<<<<< RCS file: RCS/meminfo.c,v Working file: meminfo.c description: Memory tracking routines -- keeps information on number and amount of objects/memory used by the various types ---------------------------- revision 1.1 date: 1994/01/13 05:31:39; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> meminfo.h <<<<<<<<<<<<< RCS file: RCS/meminfo.h,v Working file: meminfo.h description: Header file for meminfo.c ---------------------------- revision 1.1 date: 1994/01/13 05:32:24; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> memory.c <<<<<<<<<<<<< RCS file: RCS/memory.c,v Working file: memory.c description: Memory allocation and de-allocation for basic data structures ---------------------------- revision 1.13 date: 1994/04/05 02:10:37; author: des; state: Exp; lines: +4 -5 Bug fixes as suggested by Jean-Pierre Haeberly. -- free'd matrix before matrix->base bug fixes (now free'd in reverse order) -- bugs in code for SEGMENTED architectures resulting in memory leaks using m_resize() and m_free(). ---------------------------- revision 1.12 date: 1994/03/08 05:44:13; author: des; state: Exp; lines: +1 -4 Removed declarations of m_zero(), px_ident() and v_zero() inside *_resize() functions. ---------------------------- revision 1.11 date: 1994/01/13 05:32:32; author: des; state: Exp; lines: +924 -255 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.10 date: 1993/08/26 07:19:05; author: des; state: Exp; lines: +33 -10 Modified the get_...() routines so that they don't call malloc() or calloc() requesting 0 bytes of memory. ---------------------------- revision 1.9 date: 1993/07/19 22:53:07; author: des; state: Exp; lines: +3 -3 Used to not deallocate in v_free() or px_free() if the dim/size > MAXDIM. It now only checks if (int)(x->dim) < 0 or (int)(px->size) < 0. ---------------------------- revision 1.8 date: 1993/06/14 22:32:17; author: des; state: Exp; lines: +2 -2 See previous entry ---------------------------- revision 1.7 date: 1993/06/14 22:20:29; author: des; state: Exp; lines: +4 -3 Changed declaration of i from u_int to int to avoid problems with m_resize(A,0,0). Bug found by Hidetoshi Shidomaira ---------------------------- revision 1.6 date: 1993/05/10 02:12:21; author: des; state: Exp; lines: +2 -2 Fixed bug in m_resize(): replaced "new_size = new_m*new_n" with "new_size = new_max_m*new_max_n". ---------------------------- revision 1.5 date: 1993/03/30 23:03:01; author: des; state: Exp; lines: +11 -2 Fixed problem with resizing SEGMENTED matrices ---------------------------- revision 1.4 date: 1992/02/28 03:12:23; author: des; state: Exp; lines: +2 -2 Removed the doubling up of #include "matdef.h" and #include "machine.h" ---------------------------- revision 1.3 date: 1992/01/27 22:17:06; author: des; state: Exp; lines: +3 -1 Fixed bug in v_resize(): newly allocated area was not zero'd -- it now is ---------------------------- revision 1.2 date: 1991/10/04 04:04:26; author: des; state: Exp; lines: +65 -7 Added code to deal with SEGMENTED memory machines in get_mat(), m_free() and m_resize(). ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> memstat.c <<<<<<<<<<<<< RCS file: RCS/memstat.c,v Working file: memstat.c description: Static workspace routines -- used to allow freeing of workspace on demand ---------------------------- revision 1.1 date: 1994/01/13 05:32:44; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> memtort.c <<<<<<<<<<<<< RCS file: RCS/memtort.c,v Working file: memtort.c des: 1.1 description: Memory allocation/deallocation/resize torture test ---------------------------- revision 1.1 locked by: des; date: 1994/01/13 05:33:17; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> mfunc.c <<<<<<<<<<<<< RCS file: RCS/mfunc.c,v Working file: mfunc.c description: Matrix functions -- exponentials, polynomials and powers ---------------------------- revision 1.2 date: 1994/11/01 05:57:56; author: des; state: Exp; lines: +11 -4 Added code so that m_pow() can handle negative exponents ---------------------------- revision 1.1 date: 1994/01/13 05:33:58; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> mfuntort.c <<<<<<<<<<<<< RCS file: RCS/mfuntort.c,v Working file: mfuntort.c description: Torture test for mfunc.c ---------------------------- revision 1.2 date: 1994/01/14 01:08:06; author: des; state: Exp; lines: +2 -1 Matrix functions (powers, polynomials and exponentials) torture test ---------------------------- revision 1.1 date: 1994/01/13 05:34:16; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> norm.c <<<<<<<<<<<<< RCS file: RCS/norm.c,v Working file: norm.c description: Matrix and vector norms: only the easy ones! ---------------------------- revision 1.6 date: 1994/01/13 05:34:35; author: des; state: Exp; lines: +38 -7 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.5 date: 1993/06/01 01:12:33; author: des; state: Exp; lines: +2 -2 Replaced string "m_norm_inf" in m_norm1() with "m_nrom1" for error() call ---------------------------- revision 1.4 date: 1993/06/01 00:03:45; author: des; state: Exp; lines: +3 -3 Changed error() calls in _v_norm_inf() to pass the string "_v_norm_inf" instead of "_norm_inf". ---------------------------- revision 1.3 date: 1992/03/19 06:03:49; author: des; state: Exp; lines: +5 -5 Fixed incorrect limits for loops in m_norm1(). ---------------------------- revision 1.2 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +22 -22 Changed names of norm functions: n1() => norm1(), n_inf() => norm_inf() etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> oldnames.h <<<<<<<<<<<<< RCS file: RCS/oldnames.h,v Working file: oldnames.h description: Backwards compatibility header file ---------------------------- revision 1.1 date: 1994/01/13 05:34:40; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> otherio.c <<<<<<<<<<<<< RCS file: RCS/otherio.c,v Working file: otherio.c description: Other input/output -- general purpose routines ---------------------------- revision 1.2 date: 1994/01/13 05:34:52; author: des; state: Exp; lines: +44 -10 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> pxop.c <<<<<<<<<<<<< RCS file: RCS/pxop.c,v Working file: pxop.c description: Basic permutation operations: see also vecop.c matop.c ---------------------------- revision 1.5 date: 1994/03/23 23:58:50; author: des; state: Exp; lines: +42 -40 Added guard to px_inv() to prevent crashes if px is not a permutation. Now raises E_BOUNDS error. ---------------------------- revision 1.4 date: 1994/03/08 05:45:16; author: des; state: Exp; lines: +9 -1 Replaced old-style declaration of m_get() with proper prototype based declarations. ---------------------------- revision 1.3 date: 1994/01/13 05:35:00; author: des; state: Exp; lines: +149 -74 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.2 date: 1992/01/17 04:13:23; author: des; state: Exp; lines: +2 -2 Made myqsort() static to avoid name clashes ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> qrfactor.c <<<<<<<<<<<<< RCS file: RCS/qrfactor.c,v Working file: qrfactor.c description: QR factorise and solve routines Also includes QR with column pivoting ---------------------------- revision 1.5 date: 1994/01/13 05:35:07; author: des; state: Exp; lines: +71 -40 Version 1.2 update -- uses new names for routines (by this stage, have removed "beta" parameters) ---------------------------- revision 1.4 date: 1992/12/09 02:29:28; author: des; state: Exp; lines: +114 -42 Added (modified & corrected) QRTsolve() Also fixed bugs in calculating beta values without using beta vector ---------------------------- revision 1.3 date: 1992/05/04 04:27:54; author: des; state: Exp; lines: +2 -2 QRCPsolve() must return a VEC *, not a MAT * ! ---------------------------- revision 1.2 date: 1992/05/04 04:24:33; author: des; state: Exp; lines: +339 -199 Added QRCPsolve() and QRcondest() functions ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> rk4.dat <<<<<<<<<<<<< RCS file: RCS/rk4.dat,v Working file: rk4.dat description: Data file for Runge-Kutta tutorial problem; see tutorial.c ---------------------------- revision 1.1 date: 1994/01/13 05:47:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> schur.c <<<<<<<<<<<<< RCS file: RCS/schur.c,v Working file: schur.c description: Schur decomposition For computing eigenvalues of general matrices Computes real, rather than comples, Schur form ---------------------------- revision 1.7 date: 1994/03/17 05:36:53; author: des; state: Exp; lines: +34 -10 Now puts 2 x 2 blocks with complex eigenvalues into the form [a b; c a] with b*c < 0 (at least to within roundoff errors). ---------------------------- revision 1.6 date: 1994/03/16 05:12:12; author: des; state: Exp; lines: +23 -13 Second bug for 2 x 2 real e-val block fixed Replaced ".. : tmp +sqrt(discrim)" with ".. : - tmp + sqrt(discrim)" Also added code to force splitting of blocks with degenerate 2 x 2 lower-right blocks after 5 iterations of no splitting. This forced splitting uses relative tolerances of sqrt(MACHEPS) rather than O(MACHEPS). This is a problem with QR algorithms for eigenvalues/vectors in general. ---------------------------- revision 1.5 date: 1994/02/21 01:11:44; author: des; state: Exp; lines: +36 -1 Bug fix!!! schur() went into an infinite loop if it found a 2x2 diagonal block with real eigenvalues. Modification explicitly deals with this case. Bug found by Graeme Boyle Tektronix Inc., Oregon ---------------------------- revision 1.4 date: 1994/01/13 05:35:52; author: des; state: Exp; lines: +299 -144 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.3 date: 1991/11/27 02:52:13; author: des; state: Exp; lines: +2 -2 Modified to use new names for the norm functions ---------------------------- revision 1.2 date: 1991/10/25 02:07:24; author: des; state: Exp; lines: +2 -2 Modified to reflect change in norm.c: n1() => norm1(), n2() => norm2() etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> solve.c <<<<<<<<<<<<< RCS file: RCS/solve.c,v Working file: solve.c description: General matrix solve routines for triangular & diagonal matrices ---------------------------- revision 1.3 date: 1994/01/13 05:29:57; author: des; state: Exp; lines: +39 -16 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1993/05/27 07:26:39; author: des; state: Exp; lines: +102 -91 Added sizeof(double) factor to the mem_copy calls in UTsolve() and LTsolve() ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sparse.c <<<<<<<<<<<<< RCS file: RCS/sparse.c,v Working file: sparse.c description: Basic sparse matrix operations ---------------------------- revision 1.10 date: 1994/03/08 05:46:07; author: des; state: Exp; lines: +1 -2 Removed unnecessary declaration of sprow_foutput() in sp_copy2(). ---------------------------- revision 1.9 date: 1994/02/28 22:38:44; author: des; state: Exp; lines: +25 -18 Bug in sp_resize(): fixed failure for (new) n > A->n. Now resizes A->start_row[] and A->start_idx[] if n > A->max_n and updates A->n in **all** cases where a != A->n. ---------------------------- revision 1.8 date: 1994/01/13 05:32:38; author: des; state: Exp; lines: +895 -764 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.7 date: 1993/04/29 03:30:41; author: des; state: Exp; lines: +3 -2 Added #include as fabs() used ---------------------------- revision 1.6 date: 1992/03/24 01:54:19; author: des; state: Exp; lines: +2 -2 Removed printf in sp_compact() ---------------------------- revision 1.5 date: 1992/03/24 01:12:27; author: des; state: Exp; lines: +64 -6 Fixed some bugs in sp_row_copy(), _row_mltadd() and sp_row_merge() [set r_out->len before calling row_xpd()] ---------------------------- revision 1.4 date: 1992/02/24 06:02:15; author: des; state: Exp; lines: +9 -2 Now sp_resize() reallocates the start_row and start_idx entries! ---------------------------- revision 1.3 date: 1992/01/21 22:20:33; author: des; state: Exp; lines: +2 -2 Fixed error reporting by: sp_vm_mlt should have been sp_mv_mlt ---------------------------- revision 1.2 date: 1991/11/27 00:50:34; author: des; state: Exp; lines: +4 -6 Fixed bugs in sp_free_mat(). Now cleans up after itself properly! ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sparse.h <<<<<<<<<<<<< RCS file: RCS/sparse.h,v Working file: sparse.h description: Include file for basic sparse matrix routines ---------------------------- revision 1.2 date: 1994/01/13 05:33:36; author: des; state: Exp; lines: +186 -46 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sparse2.h <<<<<<<<<<<<< RCS file: RCS/sparse2.h,v Working file: sparse2.h description: Include file for sparse matrix factorisations and solve and e-val routines ---------------------------- revision 1.4 date: 1994/01/13 05:33:46; author: des; state: Exp; lines: +60 -24 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.3 date: 1992/12/07 00:08:58; author: des; state: Exp; lines: +7 -1 spBKPfactor() and spBKPsolve() added to functions ---------------------------- revision 1.2 date: 1992/10/13 03:45:25; author: des; state: Exp; lines: +3 -1 Added spILUfactor() entry. See also splufctr.c. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sparseio.c <<<<<<<<<<<<< RCS file: RCS/sparseio.c,v Working file: sparseio.c description: Sparse matrix input/output ---------------------------- revision 1.4 date: 1994/01/13 05:34:25; author: des; state: Exp; lines: +80 -46 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.3 date: 1993/08/27 06:42:10; author: des; state: Exp; lines: +2 -2 Made line[] array static to avoid name collisions. ---------------------------- revision 1.2 date: 1993/06/11 05:06:42; author: des; state: Exp; lines: +63 -53 Sparse matrix output routine now puts a '#\n' after printing the matrix This should prevent a following number being read as a new entry ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> spbkp.c <<<<<<<<<<<<< RCS file: RCS/spbkp.c,v Working file: spbkp.c description: Sparse Bunch-Kaufman-Parlett symmetric indefinite factorisation Uses Markowitz-style pivot selection a la Duff, Gould, Reid, Scott & Turner ---------------------------- revision 1.5 date: 1994/01/13 05:44:35; author: des; state: Exp; lines: +141 -135 Version 1.2 update -- changed names etc ---------------------------- revision 1.4 date: 1993/06/07 22:38:33; author: des; state: Exp; lines: +4 -5 Modified "#ifndef _H_MALLOC" to "#ifdef MALLOCDECL". Replaced "min_val = iv->ive" with "min_val = iv->ive[0]", and chase_col() now returns a row_elt *. ---------------------------- revision 1.3 date: 1993/05/25 02:32:49; author: des; state: Exp; lines: +30 -7 Bug in iv_min() corrected in line 413: "min_val = iv->ive;" should have been "min_val = iv->ive[0];" ---------------------------- revision 1.2 date: 1992/12/07 00:02:53; author: des; state: Exp; lines: +600 -697 Radically modified to use Karen George's idea of unordered sparse rows ---------------------------- revision 1.1 date: 1992/11/04 03:06:27; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> spchfctr.c <<<<<<<<<<<<< RCS file: RCS/spchfctr.c,v Working file: spchfctr.c description: Sparse Cholseky factorisation routines Includes symbolic and incomplete factorisations ---------------------------- revision 1.4 date: 1994/01/13 05:31:32; author: des; state: Exp; lines: +86 -58 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.3 date: 1992/05/25 23:35:38; author: des; state: Exp; lines: +3 -1 Added extra ANSI_C wrapper around calloc() and reallco() (re)definitions. Bug spotted by Alex Sergejew, U of Q Dept Psychiatry(!) ---------------------------- revision 1.2 date: 1992/02/16 21:58:40; author: des; state: Exp; lines: +2 -2 Replaced the #ifndef RS6000 with #ifndef MALLOCDECL for portabilbity ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> splufctr.c <<<<<<<<<<<<< RCS file: RCS/splufctr.c,v Working file: splufctr.c description: Sparse LU factorisation ---------------------------- revision 1.4 date: 1994/03/23 04:38:05; author: des; state: Exp; lines: +43 -9 This is a major revision of spILUfactor() with both efficiency improvements and bug fixes. On the efficiency side the new version uses the nxt_row and nxt_idx fields to jump to the new rows in a given column. Bugs became apparent with alpha == 0.0. ---------------------------- revision 1.3 date: 1994/01/13 05:35:56; author: des; state: Exp; lines: +74 -44 Version 1.2 update -- uses new names for routines ---------------------------- revision 1.2 date: 1992/01/16 02:56:31; author: des; state: Exp; lines: +95 -6 Now initialises merge variable in spLUfactor() ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sprow.c <<<<<<<<<<<<< RCS file: RCS/sprow.c,v Working file: sprow.c description: Sparse row utility routines -- were part of sparse.c ---------------------------- revision 1.1 date: 1994/01/13 05:35:36; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> spswap.c <<<<<<<<<<<<< RCS file: RCS/spswap.c,v Working file: spswap.c description: Collection of routines based around swapping sparse rows/cols and chasing column entries ---------------------------- revision 1.3 date: 1994/01/13 05:44:43; author: des; state: Exp; lines: +46 -17 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1992/12/07 00:05:29; author: des; state: Exp; lines: +39 -31 Modified along with spbkp.c (v.1.2) ---------------------------- revision 1.1 date: 1992/11/04 03:07:02; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> sptort.c <<<<<<<<<<<<< RCS file: RCS/sptort.c,v Working file: sptort.c des: 1.1 description: Sparse "torture" test ---------------------------- revision 1.1 locked by: des; date: 1994/01/13 04:26:50; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> submat.c <<<<<<<<<<<<< RCS file: RCS/submat.c,v Working file: submat.c description: Sub-matrix and sub-vector routines ---------------------------- revision 1.2 date: 1994/01/13 05:28:12; author: des; state: Exp; lines: +128 -96 Version 1.2 update -- changed names etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> svd.c <<<<<<<<<<<<< RCS file: RCS/svd.c,v Working file: svd.c description: Singular value decomposition ---------------------------- revision 1.6 date: 1994/01/13 05:44:16; author: des; state: Exp; lines: +47 -12 Version 1.2 update -- changed names etc ---------------------------- revision 1.5 date: 1992/05/03 23:38:37; author: des; state: Exp; lines: +2 -2 The error messages in svd() now pass "svd" as the function name They did say "bifactor"! ---------------------------- revision 1.4 date: 1992/02/17 05:16:30; author: des; state: Exp; lines: +0 -1 Removed funny (non-printing) character at end ---------------------------- revision 1.3 date: 1992/01/17 04:05:27; author: des; state: Exp; lines: +91 -0 Now svd() and bisvd() returns nonnegative singular values sorted in decreasing order (as the theory is usually represented). ---------------------------- revision 1.2 date: 1991/11/27 02:53:10; author: des; state: Exp; lines: +1 -1 Modified to use the new names for the norm functions ---------------------------- revision 1.1 date: 1991/10/25 02:07:24; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> symmeig.c <<<<<<<<<<<<< RCS file: RCS/symmeig.c,v Working file: symmeig.c description: Symmetric eigenvalue routines Has routines for tridiagonal and general symmetric matrices ---------------------------- revision 1.5 date: 1994/02/16 03:23:39; author: des; state: Exp; lines: +4 -4 Changed "i_min < i_max-2" to "i_min < i_max-1" in trieig(). This now works correctly, so that it **does** use b_ve[i_max-1] ---------------------------- revision 1.4 date: 1994/02/16 01:25:35; author: des; state: Exp; lines: +10 -7 Bug fixes similar to previous as found by Zbigniew on lines 109, 110, 112 and 138 --- bug would result in failure for n == 1. ---------------------------- revision 1.3 date: 1994/02/16 00:50:52; author: des; state: Exp; lines: +2 -2 Changed l.137 of trieig() to avoid accessing b_ve[i] for i >= i_max. Bug notified by Andy Vick, Durham, UK on a DEC alpha with OSF1.2a ---------------------------- revision 1.2 date: 1994/01/13 05:27:47; author: des; state: Exp; lines: +46 -11 Version 1.2 update -- changed names etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> torture.c <<<<<<<<<<<<< RCS file: RCS/torture.c,v Working file: torture.c description: Basic torture test for Meschach ---------------------------- revision 1.6 date: 1994/08/25 15:22:11; author: des; state: Exp; lines: +16 -18 Added call to bd_mv_mlt(). See bdfactor.c ---------------------------- revision 1.5 date: 1994/05/18 01:53:58; author: des; state: Exp; lines: +3 -2 Modified torture.c to test non-in situ band transposition to catch by found in bd_transp(). ---------------------------- revision 1.4 date: 1994/02/21 01:14:37; author: des; state: Exp; lines: +14 -2 Added test for infinite loop in schur() for 2x2 block with real eigenvalues. See schur.c; based on Graeme Boyle (Tektronix Inc., Oregon) code. ---------------------------- revision 1.3 date: 1994/01/13 05:42:55; author: des; state: Exp; lines: +350 -88 Version 1.2 update -- changed names etc ---------------------------- revision 1.2 date: 1993/06/08 23:51:43; author: des; state: Exp; lines: +7 -4 Added call to m_version(). ---------------------------- revision 1.1 date: 1993/05/10 02:25:33; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> tutadv.c <<<<<<<<<<<<< RCS file: RCS/tutadv.c,v Working file: tutadv.c description: Tutorial in using extended type facilities for, e.g., rank 3 tensors ---------------------------- revision 1.2 date: 1994/01/14 03:43:32; author: des; state: Exp; lines: +1 -0 Minor modifications ---------------------------- revision 1.1 date: 1994/01/13 05:26:55; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> tutorial.c <<<<<<<<<<<<< RCS file: RCS/tutorial.c,v Working file: tutorial.c description: Tutorial code revised for version 1.2 ---------------------------- revision 1.3 date: 1994/01/16 22:53:09; author: des; state: Exp; lines: +1 -1 Minor change to $Id: $ line ---------------------------- revision 1.2 date: 1994/01/16 22:52:01; author: des; state: Exp; lines: +3 -2 Minor corrections (see ZL) ---------------------------- revision 1.1 date: 1994/01/13 05:26:26; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> update.c <<<<<<<<<<<<< RCS file: RCS/update.c,v Working file: update.c description: Update routines for LDL^T and QR factorisations ---------------------------- revision 1.2 date: 1994/01/13 05:26:06; author: des; state: Exp; lines: +32 -3 V.1.2a update ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> vecop.c <<<<<<<<<<<<< RCS file: RCS/vecop.c,v Working file: vecop.c description: Basic vector operations: see also matop.c ---------------------------- revision 1.4 date: 1994/03/08 05:50:39; author: des; state: Exp; lines: +10 -2 Added #ifdef PROTOTYPES_IN_STRUCT wrappers for v_map() and _v_map(). ---------------------------- revision 1.3 date: 1994/01/13 04:29:06; author: des; state: Exp; lines: +178 -77 V.1.2a update: added componentwise operations incl. convolutions ---------------------------- revision 1.2 date: 1992/10/15 23:38:05; author: des; state: Exp; lines: +208 -1 Added componentwise operations v_map(), v_max(), v_min(), v_star() (Hadamard product), v_slash() (Hadamard inverse), v_sort() and v_sum(). ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> version.c <<<<<<<<<<<<< RCS file: RCS/version.c,v Working file: version.c description: For official netlib release ---------------------------- revision 1.9 date: 1994/03/24 00:04:05; author: des; state: Exp; lines: +6 -1 Added notes on changes to spILUfactor() and px_inv(). ---------------------------- revision 1.8 date: 1994/02/21 04:32:25; author: des; state: Exp; lines: +10 -4 Set version to 1.2b with bug fixes in schur() and installation. ---------------------------- revision 1.7 date: 1994/01/13 05:43:57; author: des; state: Exp; lines: +36 -21 Version 1.2 update ---------------------------- revision 1.6 date: 1993/06/09 00:08:07; author: des; state: Exp; lines: +4 -1 Replaced $Header$ wit $Id$ for RCS ---------------------------- revision 1.5 date: 1993/05/10 02:18:33; author: des; state: Exp; lines: +9 -2 Update to version 1.1b from 1.1 ---------------------------- revision 1.4 date: 1993/03/23 01:45:45; author: des; state: Exp; lines: +5 -2 Updated version number for 2nd netlib release ---------------------------- revision 1.3 date: 1992/03/24 23:46:10; author: des; state: Exp; lines: +10 -2 Added comment describing use of m_version() ---------------------------- revision 1.2 date: 1992/03/24 23:42:41; author: des; state: Exp; lines: +3 -2 Updated version number for netlib release ---------------------------- revision 1.1 date: 1992/03/24 23:37:40; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zcopy.c <<<<<<<<<<<<< RCS file: RCS/zcopy.c,v Working file: zcopy.c description: Copying complex vectors & matrices ---------------------------- revision 1.1 date: 1994/01/13 04:28:42; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zfunc.c <<<<<<<<<<<<< RCS file: RCS/zfunc.c,v Working file: zfunc.c description: Operations on complex numbers ---------------------------- revision 1.1 date: 1994/01/13 04:28:29; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zgivens.c <<<<<<<<<<<<< RCS file: RCS/zgivens.c,v Working file: zgivens.c description: Complex Givens' rotations ---------------------------- revision 1.1 date: 1994/01/13 04:27:11; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zhessen.c <<<<<<<<<<<<< RCS file: RCS/zhessen.c,v Working file: zhessen.c description: Complex Hessenberg form computations ---------------------------- revision 1.1 date: 1994/01/13 04:27:00; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zhsehldr.c <<<<<<<<<<<<< RCS file: RCS/zhsehldr.c,v Working file: zhsehldr.c description: Complex Householder operations ---------------------------- revision 1.2 date: 1994/04/07 01:43:47; author: des; state: Exp; lines: +12 -4 Avoids problems with zero values of out->ve[i0] now. Bug noticed by Gary D. Brushe of DSTO, Australia. ---------------------------- revision 1.1 date: 1994/01/13 04:26:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zlufctr.c <<<<<<<<<<<<< RCS file: RCS/zlufctr.c,v Working file: zlufctr.c description: Complex LU factorisation (Gaussian Elimination) ---------------------------- revision 1.1 date: 1994/01/13 04:26:20; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmachine.c <<<<<<<<<<<<< RCS file: RCS/zmachine.c,v Working file: zmachine.c description: Basic operations on (naked!) complex arrays ---------------------------- revision 1.1 date: 1994/01/13 04:25:41; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmatio.c <<<<<<<<<<<<< RCS file: RCS/zmatio.c,v Working file: zmatio.c description: Human readable complex vector/matrix input/output ---------------------------- revision 1.1 date: 1994/01/13 04:25:18; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmatlab.c <<<<<<<<<<<<< RCS file: RCS/zmatlab.c,v Working file: zmatlab.c description: Complex matrix/vector/number save/load in MATLAB (binary) format ---------------------------- revision 1.1 date: 1994/01/13 04:24:57; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmatop.c <<<<<<<<<<<<< RCS file: RCS/zmatop.c,v Working file: zmatop.c description: Complex matrix operations ---------------------------- revision 1.1 date: 1994/01/13 04:24:46; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmatrix.h <<<<<<<<<<<<< RCS file: RCS/zmatrix.h,v Working file: zmatrix.h description: Header file for basic complex operations ---------------------------- revision 1.2 date: 1994/03/08 05:50:09; author: des; state: Exp; lines: +5 -0 Added #ifdef PROTOTYPES_IN_STRUCT wrappers for zv_map() and _zv_map(). ---------------------------- revision 1.1 date: 1994/01/13 04:24:34; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmatrix2.h <<<<<<<<<<<<< RCS file: RCS/zmatrix2.h,v Working file: zmatrix2.h description: Header file for complex factorisation routines ---------------------------- revision 1.1 date: 1994/01/13 04:22:14; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zmemory.c <<<<<<<<<<<<< RCS file: RCS/zmemory.c,v Working file: zmemory.c description: Memory allocation/deallocation and resizing for complex matrices and vectors ---------------------------- revision 1.2 date: 1994/04/05 02:13:14; author: des; state: Exp; lines: +6 -7 Bug fixes as suggested by Jean-Pierre Haeberly. -- free'd matrix before matrix->base bug fixes (now free'd in reverse order) -- bugs in code for SEGMENTED architectures resulting in memory leaks using m_resize() and m_free(). ---------------------------- revision 1.1 date: 1994/01/13 04:21:55; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> znorm.c <<<<<<<<<<<<< RCS file: RCS/znorm.c,v Working file: znorm.c description: Computes norms of complex vectors and matrices ---------------------------- revision 1.1 date: 1994/01/13 04:21:31; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zqrfctr.c <<<<<<<<<<<<< RCS file: RCS/zqrfctr.c,v Working file: zqrfctr.c description: Complex QR factorisation ---------------------------- revision 1.1 date: 1994/01/13 04:21:22; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zschur.c <<<<<<<<<<<<< RCS file: RCS/zschur.c,v Working file: zschur.c description: Complex Schur form (produces triangular matrices, rather than just quasi-triangular matrices as in schur.c) ---------------------------- revision 1.1 date: 1994/01/13 04:20:48; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zsolve.c <<<<<<<<<<<<< RCS file: RCS/zsolve.c,v Working file: zsolve.c description: Solve for complex triangular systems of equations ---------------------------- revision 1.1 date: 1994/01/13 04:20:33; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> ztorture.c <<<<<<<<<<<<< RCS file: RCS/ztorture.c,v Working file: ztorture.c des: 1.1 description: "Torture" test for complex vector/matrix operations ---------------------------- revision 1.1 locked by: des; date: 1994/01/13 04:20:15; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> zvecop.c <<<<<<<<<<<<< RCS file: RCS/zvecop.c,v Working file: zvecop.c description: Complex vector operations ---------------------------- revision 1.2 date: 1994/03/08 05:51:07; author: des; state: Exp; lines: +10 -2 Added #ifdef PROTOTYPES_IN_STRUCT wrappers for zv_map() and _zv_map(). ---------------------------- revision 1.1 date: 1994/01/13 04:26:36; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> fnindex.txt <<<<<<<<<<<<< >>>>>>>>>>>>> tutorial.txt <<<<<<<<<<<<< >>>>>>>>>>>>> Cray <<<<<<<<<<<<< >>>>>>>>>>>>> GCC <<<<<<<<<<<<< >>>>>>>>>>>>> Linux <<<<<<<<<<<<< >>>>>>>>>>>>> RS6000 <<<<<<<<<<<<< >>>>>>>>>>>>> SGI <<<<<<<<<<<<< >>>>>>>>>>>>> SPARC <<<<<<<<<<<<< >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> patch.1 <<<<<<<<<<<<< >>>>>>>>>>>>> patch.2 <<<<<<<<<<<<< >>>>>>>>>>>>> patch.3 <<<<<<<<<<<<< >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.c <<<<<<<<<<<<< RCS file: RCS/machine.c,v Working file: machine.c description: Machine dependent code. This code is the basis of most of the basic dense matrix/vector operations. Change this file for "tuning" to a particular machine/compiler. ---------------------------- revision 1.4 date: 1994/01/13 05:28:56; author: des; state: Exp; lines: +65 -132 Version 1.2 update -- added the "#ifdef VUNROLL" parts for selective unrolling of loops ---------------------------- revision 1.3 date: 1991/08/29 06:42:11; author: des; state: Exp; lines: +13 -2 Used loop unrolling on __mltadd__() ---------------------------- revision 1.2 date: 1991/08/29 06:15:24; author: des; state: Exp; lines: +157 -28 Modified __ip__() and added __ip4__(), __lc4__() and __ma4__() to take help take advantage of vector & super-scalar architectures. For full advantage would need to re-write matop.c, lufactor.c etc etc ---------------------------- revision 1.1 date: 1991/08/14 00:10:43; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> machine.h <<<<<<<<<<<<< RCS file: RCS/machine.h,v Working file: machine.h description: Contains machine dependencies Change this file before anything else! ---------------------------- revision 1.4 date: 1994/01/13 05:29:37; author: des; state: Exp; lines: +162 -36 Version 1.2 update -- works with configure script ---------------------------- revision 1.3 date: 1992/02/16 21:59:19; author: des; state: Exp; lines: +8 -1 Added code to ensure that HUGE_VAL is defined, in terms of HUGE if not already. ---------------------------- revision 1.2 date: 1991/10/03 23:39:57; author: des; state: Exp; lines: +13 -0 Added MALLOCDECL macro for testing whther r or not malloc() etc have been declared. ---------------------------- revision 1.1 date: 1991/08/14 00:06:35; author: des; state: Exp; Initial revision ============================================================================= >>>>>>>>>>>>> makefile <<<<<<<<<<<<< RCS file: RCS/makefile,v Working file: makefile description: "Vanilla" makefile for Meschach ---------------------------- revision 1.2 date: 1994/02/21 04:35:43; author: des; state: Exp; lines: +192 -28 Updated to mesch12b.tar and mesch12b.zip ---------------------------- revision 1.1 date: 1993/05/10 02:28:37; author: des; state: Exp; Initial revision ============================================================================= .