Suppressing the output information resulted from PARDISO solver.

[Ngu_Soon_Hui]

I am testing the PARDISO solver now. The program that I am currently testing can be found here.

One thing I would like to do, I would like to supress all the logging information of the PARDISO solver ( see above), I don't want to see all the statistics regarding the solver running. Is there anyway I can suppress them?

From the documentation, I understand that by setting the iparm(18) and iparm(19) to 0, those diagnostic information should be suppressed, but when I modify the above program accordingly all the output still gets presented!.

Any idea how to fix this? Or is there anything I miss?

[Gennady_F_Intel · Accepted solution]

please set

msglvl    = 0    instead ofmsglvl == 1.

-- Gennady

[Ngu_Soon_Hui]

Thanks, but i think there is still this

0

1

2

3

4

5

appearing on the screen. How to suppress them?

[Alexander_K_Intel2]

Hi,

The testcase you provided call PARDISO 4.1.0 from pardiso-project. Intel doesn't support this project so we could provide you some ideas like set msglvl=1.

With best regards,

Alexander Kalinkin

[Ngu_Soon_Hui]

Alexander,

Actually the same sample program is also available in Intel MKL reference Manual, under Section C, Pardiso Code Examples.

[Alexander_K_Intel2]

Hi,

Yes, similar example exists in Intel MKL and we support it. The Gennady's advise about msglvl will resolve your issue if you will use MKL functionality.

With best regards,

Alexander Kalinkin

[Ngu_Soon_Hui]

Alex,

gennady advise will solve the problem only partially, but there is still this

0

1

2

3

4

5

numerical value appearing on the screen, I afraid.

[Alexander_K_Intel2]

Hi again,

MKL example is similar but not the same with example from pardiso project. On which example your problem appeared? If this problem exists in MKL example could you provide the exact name of this example or attach it here?

With best regards,

Alexander Kalinkin

[Ngu_Soon_Hui]

The example I am talking about is Example C-6, pardiso_sym.f

[Alexander_K_Intel2]

Hi,

Could you also provide the version of MKL that you currently use to help me reproduce your issue?

With best regards,

Alexander Kalinkin

[Gennady_F_Intel]

Just an example of output which I received with pardiso_unsym.c example:

version on MKL 10.2 update7:

msglvl disabled

Reordering completed ...

Number of nonzeros in factors = 21

Number of factorization MFLOPS = 0

Factorization completed ...

Solve completed ...

The solution of the system is:

x [0] = -0.522321

x [1] = -0.008929

x [2] = 1.220982

x [3] = -0.504464

x [4] = -0.214286

iparm[63] = 102000115

and when msglvl == enabled

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

================ PARDISO: solving a real nonsymmetric system ================

The local (internal) PARDISO version is : 102000115

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

Matching is turned ON

Summary PARDISO: ( reorder to reorder )

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

Times:

======

Time spent in calculations of symmetric matrix portrait(fulladj): 0.000007 s

Time spent in reordering of the initial matrix(reorder) : 0.000158 s

Time spent in symbolic factorization(symbfct) : 0.001695 s

Time spent in in allocation of internal data structures(malloc) : 0.000246 s

Time spent in additional calculations : 0.003461 s

Total time spent : 0.005567 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 2

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 5

#non-zeros in A: 13

non-zeros in A (): 52.000000

#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: 2

size of largest supernode: 4

number of nonzeros in L 19

number of nonzeros in U 2

number of nonzeros in L+U 21

Reordering completed ...

Number of nonzeros in factors = 21

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

10 %

94 %

100 %

================ PARDISO: solving a real nonsymmetric system ================

Summary PARDISO: ( factorize to factorize )

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

Times:

======

Time spent in copying matrix to internal data structure(A to LU): 0.000000 s

Time spent in factorization step(numfct) : 0.000139 s

Time spent in in allocation of internal data structures(malloc) : 0.000044 s

Time spent in additional calculations : 0.000001 s

Total time spent : 0.000184 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 2

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 5

#non-zeros in A: 13

non-zeros in A (): 52.000000

#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: 2

size of largest supernode: 4

number of nonzeros in L 19

number of nonzeros in U 2

number of nonzeros in L+U 21

gflop for the numerical factorization: 0.000000

gflop/s for the numerical factorization: 0.000352

Factorization completed ...

================ PARDISO: solving a real nonsymmetric system ================

Summary PARDISO: ( solve to solve )

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

Times:

======

Time spent in direct solver at solve step (solve) : 0.000033 s

Time spent in additional calculations : 0.000052 s

Total time spent : 0.000085 s

Statistics:

===========

< Parallel Direct Factorization with #processors: > 2

< Numerical Factorization with Level-3 BLAS performance >

< Linear system Ax = b>

#equations: 5

#non-zeros in A: 13

non-zeros in A (): 52.000000

#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: 2

size of largest supernode: 4

number of nonzeros in L 19

number of nonzeros in U 2

number of nonzeros in L+U 21

gflop for the numerical factorization: 0.000000

gflop/s for the numerical factorization: 0.000352

Solve completed ...

The solution of the system is:

x [0] = -0.522321

x [1] = -0.008929

x [2] = 1.220982

x [3] = -0.504464

x [4] = -0.214286

iparm[63] = 102000115

[Ngu_Soon_Hui]

Alex,

After further investigation the 1,2,3,4,5 is coming from my part of the program. The pardiso doesn't output those, sorry for the confusion!

[Alexander_K_Intel2]

Hi,

That's fine! Feel free to ask more about Pardiso and other part of MKL!

With best regards,

Alexander Kalinkin