bugGNU Octave - Bugs: bug #55132, make check fail at sparse/bicg.m

 
 

bug #55132: make check fail at sparse/bicg.m

Submitter:  Leonardo <pocker>
Submitted:  Thu 29 Nov 2018 03:51:15 PM UTC
   
 
Category:  Configuration and Build System Severity:  3 - Normal
Priority:  5 - Normal Item Group:  None
Status:  Fixed Assigned to:  None
Originator Name:  Open/Closed:  * Closed
Release:  * 4.4.1 Operating System:  * GNU/Linux
Fixed Release:  None Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Mon 03 Dec 2018 04:55:48 PM UTC, comment #4: 

I increased the tolerance on the BIST test by a 1 eps here (https://hg.savannah.gnu.org/hgweb/octave/rev/b9debf4436aa).  The fix will be a part of the Octave 5.0 release expected at the end of the year.

Rik <rik5>
Group administrator
Mon 03 Dec 2018 11:55:32 AM UTC, comment #3: 

Following Rik's suggestions, when I changed bicg.m to

assert (norm (A*x - b) / norm (b), 0, relres);

it fails, but PASS when I changed to

assert (norm (b - A * x) / norm (b), 0, relres + eps);

Leonardo <pocker>
Thu 29 Nov 2018 10:13:07 PM UTC, comment #2: 

It's weird, however, that it is failing.  The third output of bicg is supposed to be


RELRES is the ratio of the final residual to its initial
          value, measured in the Euclidean norm.


which in this case should be exactly


norm (b - A * x) / norm (b)


Could the original reporter edit the file bicg.m and change the test to


 assert (norm (A*x - b) / norm (b), 0, relres);


and retest?

If that doesn't work then try this


 assert (norm (b - A * x) / norm (b), 0, relres + eps);



Rik <rik5>
Group administrator
Thu 29 Nov 2018 07:50:00 PM UTC, comment #1: 

Seems to be too low a tolerance (2e-16).
IMO nothing to worry about in practice but it's good you've reported it, thanks.

Philip Nienhuis <philipnienhuis>
Group Member
Thu 29 Nov 2018 03:51:15 PM UTC, original submission:  

I don't know if this is a bug or not, nor where to report it.

I am running linux mint 19 (xfce) and I am trying to build and install octave from source.
After a succeful "make" I did a "make check" and got the results:

PASS                            14924
  FAIL                                1
  XFAIL (reported bug)               35
  SKIP (missing feature)            127
  SKIP (run-time condition)          12


the "FAIL" was at sparse/bicg.m

results of bicg.m (extracted from fntests.log)

*** test
 ## unpreconditioned residual
 A = sparse (toeplitz ([2, 1, 0, 0, 0], [2, -1, 0, 0, 0]));
 b = sum (A, 2);
 M = magic (5);
 [x, flag, relres] = bicg (A, b, [], 2, M);
 assert (norm (b - A * x) / norm (b), 0, relres);
!!!!! test failed
ASSERT errors for:  assert (norm (b - A * x) / norm (b),0,relres)

  Location  |  Observed  |  Expected  |  Reason
     ()        0.85267         0         Abs err 0.85267 exceeds tol 0.85267 by 2e-16


Leonardo <pocker>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by rik5 (Posted a comment)
  • -email is unavailable- added by pocker (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only group members can vote.

     

    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2018-12-03 rik5 StatusNeed Info Fixed
        Open/ClosedOpen Closed
    2018-11-29 rik5 StatusNone Need Info

    Back to the top

    Powered by Savane 3.13-02a9.
    Corresponding source code