Showing results for

- Intel Community
- Software
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library
- Pardiso crashes in MKL 10.3 (worked in 10.2.3)

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

Rawlins__David

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-09-2010
11:03 AM

67 Views

Pardiso crashes in MKL 10.3 (worked in 10.2.3)

I recently upgrated to the MKL version 10.3 from 10.2.3. I don't see any changes in the way Pardiso is called, but I'm getting a segfault error when calling pardiso with phase = 22 where I was not before. Here is the error I an getting:

Program received signal SIGSEGV, Segmentation fault.

0x0000000000598f7d in mkl_pds_lp64_pardiso_c ()

Have you seen this problem before?

Thanks,

Dave

1 Solution

Konstantin_A_Intel

Employee

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-09-2010
08:55 PM

67 Views

It seems like a message catalog from an old version of MKL was used instead of the newest one.

Could you please report your NLSPATH variable value? BTW, am I right that your OS is linux?

Regards,

Konstantin

Link Copied

5 Replies

Rawlins__David

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-09-2010
04:24 PM

67 Views

This is the output from 10.2.3:

=== PARDISO is running in In-Core mode, because iparam(60)=0 ===

================ PARDISO: solving a symmetric indef. system ================

Summary PARDISO: ( reorder to reorder )

================

Times:

======

Time fulladj: 0.000005 s

Time reorder: 0.000070 s

Time symbfct: 0.000028 s

Time malloc : 0.000040 s

Time total : 0.000252 s total - sum: 0.000109 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 1

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 8

#non-zeros in A: 18

non-zeros in A (%): 28.125000

#right-hand sides: 1

< Factors L and U >

#columns for each panel: 128

#independent subgraphs: 0

< Preprocessing with state of the art partitioning metis>

#supernodes: 5

size of largest supernode: 4

number of nonzeros in L 29

number of nonzeros in U 1

number of nonzeros in L+U 30

Reordering completed ...

Number of nonzeros in factors = 30

Number of factorization MFLOPS = 0 Percentage of computed non-zeros for LL^T factorization

0 % 10 % 17 % 31 % 44 % 100 %

================ PARDISO: solving a symmetric indef. system ================

Summary PARDISO: ( factorize to factorize )

================

Times:

======

Time A to LU: 0.000000 s

Time numfct : 0.000121 s

Time malloc : 0.000009 s

Time total : 0.000130 s total - sum: 0.000001 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 1

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 8

#non-zeros in A: 18

non-zeros in A (%): 28.125000

#right-hand sides: 1

< Factors L and U >

#columns for each panel: 128

#independent subgraphs: 0

< Preprocessing with state of the art partitioning metis>

#supernodes: 5

size of largest supernode: 4

number of nonzeros in L 29

number of nonzeros in U 1

number of nonzeros in L+U 30

gflop for the numerical factorization: 0.000000

gflop/s for the numerical factorization: 0.000605

Factorization completed ...

================ PARDISO: solving a symmetric indef. system ================

Summary PARDISO: ( solve to solve )

================

Times:

======

Time solve : 0.000019 s

Time total : 0.000042 s total - sum: 0.000023 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 1

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 8

#non-zeros in A: 18

non-zeros in A (%): 28.125000

#right-hand sides: 1

< Factors L and U >

#columns for each panel: 128

#independent subgraphs: 0

< Preprocessing with state of the art partitioning metis>

#supernodes: 5

size of largest supernode: 4

number of nonzeros in L 29

number of nonzeros in U 1

number of nonzeros in L+U 30

gflop for the numerical factorization: 0.000000

gflop/s for the numerical factorization: 0.000605

Solve completed ...

The solution of the system is:

x [0] = -0.041860

x [1] = -0.003413

x [2] = 0.117250

x [3] = -0.112640

x [4] = 0.024172

x [5] = -0.107633

x [6] = 0.198720

x [7] = 0.190383

==================================

This is the output from 10.3:

Zero Pivot detected

reorder

The local (internal) PARDISO version is : 103000115

1-based array indexing is turned ON

PARDISO double precision computation is turned ON

METIS algorithm at reorder step is turned ON

Single-level factorization algorithm is turned ON

Scaling is turned ON

======Times: Time parlist:Times: Time A to LU:

Time numfct :

Time malloc :

Time cgs :

Time spent in calculations of symmetric matrix portrait(fulladj): #equations: -1139805488

Time spent in reordering of the initial matrix(reorder) : #equations: -1139805488

Time spent in symbolic factorization(symbfct) : #equations: -1139805488

Time spent in in allocation of internal data structures(malloc) : #equations: -1139805488

Time spent in additional calculations : #equations: -1139805488

Total time spent : #equations: -1139805488

#non-zeros in A: 2

Time solve :

#non-zeros in A: 1

#right-hand sides: 2

< Factors L and U >

#columns for each panel: 8

#independent subgraphs: 18

#independent subgraphs: 215

< with multiple minimum degree on the separator nodes >

< no multiple minimum degree on the separator nodes >

#supernodes: 128

size of largest supernode: 0

number of nonzeros in U 2

gflop for the numerical factorization: 0.000000

||A|| 0.000000

L&U for matrix number 29 deleted

Input error is not equal to ZERO, error = 1

Segmentation fault

Konstantin_A_Intel

Employee

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-09-2010
08:55 PM

68 Views

It seems like a message catalog from an old version of MKL was used instead of the newest one.

Could you please report your NLSPATH variable value? BTW, am I right that your OS is linux?

Regards,

Konstantin

Rawlins__David

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-10-2010
03:50 PM

67 Views

Thanks!

Dave

> echo $NLSPATH

/opt/intel/mkl/10.2.3.029/lib/em64t/locale/%l_%t/%N:/usr/sci/linux64/intel/Compiler/11.0/081/lib/intel64/locale/%l_%t/%N:/usr/sci/Linux64/intel/Compiler/11.1/046/idb/intel64/locale/%l_%t/%N:/usr/sci/linux64/intel/Compiler/11.0/083/lib/intel64/locale/%l_%t/%N:/opt/intel/mkl/10.2.3.029/lib/em64t/locale/%l_%t/%N:/usr/sci/linux64/intel/Compiler/11.0/081/lib/intel64/locale/%l_%t/%N:/usr/sci/Linux64/intel/Compiler/11.1/046/idb/intel64/locale/%l_%t/%N:/usr/sci/linux64/intel/Compiler/11.0/083/lib/intel64/locale/%l_%t/%N

Rawlins__David

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-13-2010
02:29 PM

67 Views

Thanks,

Dave

Gennady_F_Intel

Moderator

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-14-2010
01:17 AM

67 Views

yes, libiomp was moved to that folder intentionally

because of this is the common resource using by another components like Compiler, IPP and MKL.

--Gennady

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

For more complete information about compiler optimizations, see our Optimization Notice.