Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
Beginner
13 Views

Pardiso Error Output Meaning

I am trying to use Pardiso for my computation, and I am getting the following error messages why I try to run the Problem

*** error PARDISO ( sequence_ido,parameters) error_num= 10
*** Input check: iparam(6) -66969200 (out of bounds)
*** input parameters inconsistent error= 10 max_fac_store_in: 1
matrix_number_in: 1 matrix_type_in: 11
ido_in : 33 neqns_in : 540
ia(neqns_in+1)-1 : 25578 nb_in : 1
Can Somebody shed some light to the meaning of these. Also when I did the Symbolic Factorisation and the LU factorisation, I am getting an error value of 1 instead of 0. What is the meaning of this, Such kind of error values are not defined in the mkl manual file.

Please shed some light in to these issues.
0 Kudos
4 Replies
Highlighted
13 Views

Hi,

Firstly could you check your input matrix by matrix checker (set iparm(27) = 1)? Also what version of library you use to call PARDISO: lp64 or ilp64? And could you provide to us value of array iparm before call to PARDISO?
With best regards,
Alexander Kalinkin
0 Kudos
Highlighted
Beginner
13 Views

Thank you sir,
I am using lp64. The values of iparam are

iparam(1) = 1
iparam(2) = 2
iparam(5) = 2
iparam(6) = 1
iparam(10) = 13
iparam(11) = 1
iparam(13) = 1
iparam(18) = -1
iparam(21) = 1
iparam(27) = 1

I have kept the iparam(27) to 1. Now where should I look to check the input matrix
0 Kudos
Highlighted
13 Views

Hi,
PARDISO automatically check input matrix if iparam(27) = 1 and return error mesagge if there are some errors in matrix structure. Nevertheless it's seems to be some error in input parameters, could you provide to us some testcase that we could compile and run to reproduce problem? It could help us to find rootcase of this issue.
With best regards,
Alexander Kalinkin
0 Kudos
Highlighted
Black Belt
13 Views

There is something amiss here.

Your input value of iparam(6) = 1, but Pardiso says it is -66969200.

Make sure that argument arrays are dimensioned and initialized properly before calling Pardiso. Complete code demonstrating the problem would be useful.
0 Kudos