topic Re: Runtime error with LAPAACK solver DGELS in Intel® oneAPI Math Kernel Library

Runtime error with LAPAACK solver DGELS

rvwilson — Tue, 13 Sep 2005 22:00:07 GMT

I am getting a runtime error (segmentation fault) when I call the LAPACK solver 'DGELS' from my Fortran program. I used a small example program that I found on the net to set up the problem and make the call to DGELS.

Did I use the correct fortran command line options? Did I link with the MKL libraries properly? Are there any compiler bugs that would cause this problem? I also set the stacksize to unlimited (e.g., limit stacksize unlimited).

Any help resolving this issue would be greatly appreciated.

Thanks in advance,
Bob

==================
Problem specifics
==================

Compiler command:
-----------------
ifort myprog.f -L$MKLPATH -lmkl_lapack -lmkl_ia32 -lguide -lpthread

where MKLPATH=/opt/intel/mkl721/lib/32

Runtime execution:
------------------
[bwilson@spalding ~/lapack]$ ./a.out
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
libc.so.6 55772D79 Unknown Unknown Unknown

Stack trace terminated abnormally.

Operating system:
-----------------
[bwilson@spalding ~]$ uname -a
Linux spalding 2.6.8-24.10-smp #1 SMP Wed Dec 22 11:54:27 UTC 2004 x86_64 x86_64 x86_64 GNU/Linux

Compiler version:
------------------
[bwilson@spalding ~]$ ifort -V
Intel Fortran Compiler for 32-bit applications, Version 9.0 Build 20050430Z Package ID: l_fc_p_9.0.021

Re: Runtime error with LAPAACK solver DGELS

TimP — Wed, 14 Sep 2005 00:13:47 GMT

Since you say your problem is small, it may not matter whether you build your test application as a 32-bit application, but it's still curious that you did that on a 64-bit system.
ifort 9.0.023 or 024 should be more reliable than 021, but you haven't given any information which would tell if that is relevant.
You should get better information on the location of the crash by using -g option, with and without optimization.

Re: Runtime error with LAPAACK solver DGELS

rvwilson — Wed, 14 Sep 2005 01:08:53 GMT

Tim,

Thanks for the reply. I recompiled with the -g option and used the debugger with the following the results included at the bottom. It gave a warning about an unrecognized programming langauge. Sorry, but I do not why this occured. Does this help to shed any light on the problem?

Thanks,
Bob

[bwilson@spalding ~/lapack]$ ifort -g myprog.f $MKLPATH/libmkl_lapack.a $MKLPATH/libmkl_ia32.a $MKLPATH/libguide.a -lpthread

[bwilson@spalding ~/lapack]$ idb ./a.out
Linux Application Debugger for 32-bit applications, Version 7.2.2, Build 20030624
------------------
object file name: ./a.out
Reading symbolic information ...WARNING: Unrecognized programming language encountered.
Attempting to use C Language rules
done
(idb) run
A:
1.0000 6.0000 2.0000
1.0000 -2.0000 -8.0000
1.0000 -2.0000 4.0000
1.0000 6.0000 14.0000

b:
96.0000
192.0000
192.0000
-96.0000
Thread received signal SEGV
stopped at [ __GI_strcat(...) 0x55746d79]

Information: An type was presented during execution of the previous command. For complete type information on this symbol, recompilation of the program will be necessary. Consult the compiler man pages for details on producing full symbol table information using the '-g' (and '-gall' for cxx) flags.

Message Edited by rvwilson on 09-13-2005 11:24 AM

Re: Runtime error with LAPAACK solver DGELS

TimP — Wed, 14 Sep 2005 03:29:57 GMT

It does look like it would be worth a problem report on your premier.intel.com account. Please fill out all the information including both ifort and MKL versions.

Re: Runtime error with LAPAACK solver DGELS

rvwilson — Wed, 14 Sep 2005 03:50:36 GMT

Thanks. I will submit an inquiry.

Bob