Re: Lion make check woes

[Doug Stewart]

On Sat, Jan 14, 2012 at 5:33 PM, Robinson, Melvin D <address@hidden> wrote:

I'm still having difficulty with the make check and would appreciate just a bit of guidance on how I can make some progress. 

 

With ATLAS installed I have the following problems with make check:

  src/DLD-FUNCTIONS/betainc.cc ........................... ***MESSAGE FROM ROUTINE INITS IN LIBRARY SLATEC.
 ***POTENTIALLY RECOVERABLE ERROR, PROG ABORTED, TRACEBACK REQUESTED
 *  Chebyshev series too short for specified accuracy
 *  ERROR NUMBER = 1
 *
 ***END OF MESSAGE

 ***JOB ABORT DUE TO UNRECOVERED ERROR.
0          ERROR MESSAGE SUMMARY
 LIBRARY    SUBROUTINE MESSAGE START             NERR     LEVEL     COUNT
 SLATEC     INITS      Chebyshev series too         1         1         1

 ***MESSAGE FROM ROUTINE INITS IN LIBRARY SLATEC.
 ***POTENTIALLY RECOVERABLE ERROR, PROG ABORTED, TRACEBACK REQUESTED
 *  Chebyshev series too short for specified accuracy
 *  ERROR NUMBER = 1
 *
 ***END OF MESSAGE

 ***JOB ABORT DUE TO UNRECOVERED ERROR.
0          ERROR MESSAGE SUMMARY
 LIBRARY    SUBROUTINE MESSAGE START             NERR     LEVEL     COUNT
 SLATEC     INITS      Chebyshev series too         1         1         2

 PASS    4/6    FAIL 2
  src/DLD-FUNCTIONS/bsxfun.cc ............................ PASS   73/73
  src/DLD-FUNCTIONS/cellfun.cc ........................... PASS  121/121
  src/DLD-FUNCTIONS/chol.cc .............................. PASS   29/29
  src/DLD-FUNCTIONS/conv2.cc ............................. PASS    4/4
  src/DLD-FUNCTIONS/convhulln.cc ......................... PASS    2/3    FAIL 1
  src/DLD-FUNCTIONS/dassl.cc ............................. PASS    4/4
  src/DLD-FUNCTIONS/det.cc ............................... PASS    5/5
  src/DLD-FUNCTIONS/dlmread.cc ........................... PASS   20/20
  src/DLD-FUNCTIONS/dmperm.cc ............................ PASS    5/5
  src/DLD-FUNCTIONS/dot.cc ...............................

The convhulln test is one that I will have to work on, but I'm not understanding why it would lock up in dot.cc.  This is simply the source code to comptes the dot product of vectors, right?  I'm trying to narrow this down

 

octave:1> test dot.cc verbose
>>>>> /Users/melrobin/octave/src/DLD-FUNCTIONS/dot.cc
  ***** test
 x = [2, 1; 2, 1];
 y = [-0.5, 2; 0.5, -2];
 assert(dot (x, y), [0 0]);
  ***** test
 x = [ 1+i, 3-i; 1-i, 3-i];
 assert(dot (x, x), [4, 20]);

 

At this point it is locks up.  I verified this behavior by using ./run-octave:

octave:1>  x = [ 1+i, 3-i; 1-i, 3-i];
octave:2> dot(x,x)

and it locks up dotting x with itself.  This smells like it is a bug with ATLAS because I do not have the same issue when using -framework veclib (the default BLAS/LAPACK for Mac OS X).  My problem using -framework veclib is that it segfaults on eigs.cc, but one problem at a time.

 

I would like to recreate this problem for this test case by calling ATLAS directly.  Can you tell me what call dot(x,x) makes for complex values?  I'm going to try it with zdot, but wanted to make sure that I know what I'm talking about. 

 

Thanks for your help.

With Ubuntu and octave Version 3.6.0-rc0

I get

octave:5> x = [ 1+i, 3-i; 1-i, 3-i];

octave:6> dot(x,x)

ans =

    4   20

octave:7> 

--

DAS

https://linuxcounter.net/user/206392.html