bugGNU Octave - Bugs: bug #38379, SLES 11 with Intel MKL:...

 
 

bug #38379: SLES 11 with Intel MKL: "WARNING: tr1/unordered_map: present but cannot be compiled"

Submitter:  None
Submitted:  Fri 22 Feb 2013 06:37:19 PM UTC
   
 
Category:  Configuration and Build System Severity:  3 - Normal
Priority:  3 - Low Item Group:  Build Failure
Status:  Works For Me Assigned to:  None
Originator Name:  Robyn Landers Originator Email:  -email is unavailable-
Open/Closed:  * Closed Release:  * 3.6.3
Operating System:  * GNU/Linux Fixed Release:  None
Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Wed 23 Oct 2013 12:12:33 AM UTC, comment #5: 

IIRC, there were a few sparse matrix bug fixes in the 3.6 series, so that may be a bug that was identified in 3.6.3 and fixed for the 3.6.4 release.

In any event, it sounds like you have a working build with gcc now, I'll close this bug.

Mike Miller <mtmiller>
Group Member
Tue 22 Oct 2013 05:21:47 PM UTC, comment #4: 

Thanks for your follow-up, Mike.

I did take the recommendation to use gcc, and got Octave 3.6.3
to build on SLES 11 SP1 on IA64.   For Math libraries, I used
FFTW and BLAS from the SLES SDK, but there was no LAPACK
there, so I used SGI's SCSL for that.   This resulted in
successful compilation.

Running self-tests (make check), it failed at this one:

  liboctave/Sparse.cc ....................................panic: Segmentation fault -- stopping myself...
* glibc detected * /usr/source/octave/octave-3.6.3/src/.libs/octave: corrupted double-linked list: 0x6000000003128710 *
make[1]: * [check] Segmentation fault
make[1]: Leaving directory `/usr/source/octave/octave-3.6.3/test'
make: * [check] Error 2

I couldn't figure out how to make it skip that test and
resume with the rest, so I just handed it over as-is
to the users to try, and that's the last I heard.



Robyn

Anonymous
Mon 21 Oct 2013 11:24:02 PM UTC, comment #3: 

Has the OP made any progress building Octave on SLES 11 with either gcc or the Intel compiler? Does Octave in fact compile and run despite the warning messages quoted?

There have been many changes in the build system since the 3.6 releases of Octave, can you try with the current development version from mercurial?

Lowering priority since Octave strongly recommends building with a free compiler such as gcc.

Mike Miller <mtmiller>
Group Member
Fri 22 Feb 2013 07:09:32 PM UTC, comment #2: 

Well, I probably wouldn't even have bothered trying to build
from source if I could have found prebuilt RPMs for SLES 11 SP1 on
IA64 platform.  Maybe I didn't know all the right places to look.

Since it appeared that I had to build it myself,
I chose Intel suite over Gnu suite on the supposition that it
might perform better, but I don't mind switching to gcc etc. if
I run into too much difficulty that seems to stem from Intel/MKL.

Thanks for the suggestion re snatching the unordered_map code
from the config.log.


Robyn

Anonymous
Fri 22 Feb 2013 06:49:59 PM UTC, comment #1: 

Note that we do not recommend building Octave with non-free libraries. If the problem is solely due to using MKL and Intel's non-free compiler, you may be on your own here.

If you want to debug this, try to make a simple C++ program with unordered_map and see if you can compile it, link it, and run it. The program for which configure tried to do this is in config.log, which also has lots more diagnostic information.

Jordi GutiƩrrez Hermoso <jordigh>
Group Member
Fri 22 Feb 2013 06:37:19 PM UTC, original submission:  

I'm trying to build Octave 3.6.3 on an Itanium system
running SLES 11 SP 1, using the Intel 11.1.064 development suite
including MKL numerical libraries.

The configuration stage generates the warning shown in the
subject line and tells me to report it as a bug.

I'm following suggestions posted at
http://software.intel.com/en-us/articles/using-intel-mkl-in-gnu-octave
which are for older versions but seem to give enough of a guide.

To prepare the environment for using Intel compilers etc.,
I run this:


source /opt/intel/Compiler/11.1/064/bin/iccvars.sh ia64


Then I run configure like this:


./configure CC=icc CXX=icc F77=ifort \
--with-blas="-Wl,--start-group -lmkl_intel_lp64 -lmkl_sequential \
-lmkl_core -Wl,--end-group -liomp5 -lpthread" \
--with-lapack="-Wl,--start-group -lmkl_intel_lp64 -lmkl_sequential \
-lmkl_core -Wl,--end-group -liomp5 -lpthread"


Configuration proceeds well until we get to this:


...
checking for unordered_map... no
checking tr1/unordered_map usability... no
checking tr1/unordered_map presence... yes
configure: WARNING: tr1/unordered_map: present but cannot be compiled
configure: WARNING: tr1/unordered_map:     check for missing prerequisite headers?
configure: WARNING: tr1/unordered_map: see the Autoconf documentation
configure: WARNING: tr1/unordered_map:     section "Present But Cannot Be Compiled"
configure: WARNING: tr1/unordered_map: proceeding with the compiler's result
configure: WARNING:     ## ------------------------------------------ ##
configure: WARNING:     ## Report this to http://octave.org/bugs.html ##
configure: WARNING:     ## ------------------------------------------ ##
checking for tr1/unordered_map... no
checking whether unordered_map requires tr1 namespace... no
...



It does continue past that, but since it tells me to
report it as a bug, well, here you go.
(I did search for "unordered_map" in the Bugs category
and found two matches that didn't quite seem like this one.)


Robyn

Anonymous

 

(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 mtmiller (Posted a comment)
  • -email is unavailable- added by jordigh (Posted a comment)
  • -email is unavailable- added by None (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 5 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2013-10-23 mtmiller StatusNeed Info Works For Me
        Open/ClosedOpen Closed
    2013-10-21 mtmiller Priority5 - Normal 3 - Low
        Item GroupNone Build Failure
        StatusNone Need Info

    Back to the top

    Powered by Savane 3.13-cf05.
    Corresponding source code