[Bug-gsl] gsl_eigen_hermv

[Peter Christopher]

Hi there,
        
        I believe that there is a bug in gsl_eigen_hermv. I attach a
program which demonstrates this problem. The output of this program
(included at the bottom of this letter) shows that the hermetian solver
fails on a real symmetric matrix that the symmetric solver handles. Some
time ago there was a bug reported for the function gsl_eigen_hermv.  
But I'm using gsl 1.4, and I checked that the bug fix was included in
gsl 1.4 and it looked like it was. I don't know what the story is.  
Although this bug is so glaring that I feel like it is very likely that
I may have messed up somehow. Any help would be great.

**********************STATS***********************
compiling with gcc 2.96
complied gsl 1.4

******************MAKEFILE (for my LOCAL gsl install)***************

GSLHOME=/theory/u0/pchristo/pack/gsl-1.4

bug-report:
        gcc -I $(GSLHOME)/gsl\
            -I $(GSLHOME)/\
            -lm\
             bug-report.c \
             $(GSLHOME)/.libs/libgsl.a\
             $(GSLHOME)/cblas/.libs/libgslcblas.a

********************** PROGRAM OUTPUT ************************

COMPLEX VERSION FAILS!!
=======================

( -4.0000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)(  
0.6000,  0.0000)
(  0.6000,  0.0000)( -3.5000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)(  
0.6000,  0.0000)
(  0.6000,  0.0000)(  0.6000,  0.0000)( -3.3333,  0.0000)(  0.6000,  0.0000)(  
0.6000,  0.0000)
(  0.6000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)( -3.2500,  0.0000)(  
0.6000,  0.0000)
(  0.6000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)(  0.6000,  0.0000)( 
-3.2000,  0.0000)


              0  -4.000000e+00  -4.485842e+00
              1  -3.500000e+00  -3.970429e+00
              2  -3.333333e+00  -3.968993e+00
              3  -3.250000e+00  -3.823090e+00
              4  -3.200000e+00  -1.034980e+00


(9.2678e-01,0.0000e+00)(0.0000e+00,0.0000e+00)(-1.8687e-02,0.0000e+00)(0.0000e+00,0.0000e+00)(3.7515e-01,0.0000e+00)
(     nan,     nan)(     nan,     nan)(     nan,     nan)(     nan,     nan)(   
  nan,     nan)
(     nan,     nan)(     nan,     nan)(     nan,     nan)(     nan,     nan)(   
  nan,     nan)
(     nan,     nan)(     nan,     nan)(     nan,     nan)(     nan,     nan)(   
  nan,     nan)
(     nan,     nan)(     nan,     nan)(     nan,     nan)(     nan,     nan)(   
  nan,     nan)


REAL VERSION WORKS!!
========================

     -4.0000      0.6000      0.6000      0.6000      0.6000
      0.6000     -3.5000      0.6000      0.6000      0.6000
      0.6000      0.6000     -3.3333      0.6000      0.6000
      0.6000      0.6000      0.6000     -3.2500      0.6000
      0.6000      0.6000      0.6000      0.6000     -3.2000


              0  -4.000000e+00  -4.489462e+00
              1  -3.500000e+00  -4.044666e+00
              2  -3.333333e+00  -3.896537e+00
              3  -3.250000e+00  -3.821292e+00
              4  -3.200000e+00  -1.031375e+00


  9.2228e-01  8.4548e-02 -3.8924e-02  2.1665e-02  3.7453e-01
 -2.6176e-01  8.4853e-01 -1.3458e-01  6.0533e-02  4.3555e-01
 -1.8331e-01 -4.2173e-01 -7.4414e-01  1.5058e-01  4.6057e-01
 -1.5943e-01 -2.4119e-01  5.8837e-01  5.8769e-01  4.7419e-01
 -1.4786e-01 -1.9190e-01  2.8363e-01 -7.9235e-01  4.8275e-01

bug-report.c
Description: buggy program