Intel® Fortran Compiler
Build applications that can scale for the future with optimized code designed for Intel® Xeon® and compatible processors.

What's this error mean?

truelies
Beginner
624 Views
Fatal: ERROR 344 from d_error_post. d_/lin_sol_self using Aasen's
decomposition, the tridiagonal's upper triangular factor has a
small diagonal at number 10
after use this function:
call d_lin_sol_self(rtemp1, rtemp2 , rcdeltaq2)
What's this error mean?
Thank you so much!
0 Kudos
4 Replies
TimP
Honored Contributor III
624 Views
That looks like an application specific message, best answered by the package documentation. On the face of it, it appears to mean that the solver detects a near singularity in the data, within the numerical accuracy.
0 Kudos
Steven_L_Intel1
Employee
624 Views
I'm guessing that this is an IMSL routine, but Tim's explanation is useful.
0 Kudos
truelies
Beginner
624 Views
Thank you Tim! There is some near singular item in the matrix, do you have some methods to solve this without the error344? Thank you so much!
0 Kudos
TimP
Honored Contributor III
624 Views
That question goes well beyond the scope of this forum, the information you have provided, and my competence. I see that Visual Numerics even posted on this topic http://www.vni.com/books/whitepapers/Aasen.html
An incomplete list of possibilities:
1. Use double precision if required
2. Bug in your program or data input
3. Formulation inappropriate to your problem (maybe even overkill?)

Not having gone so far as to attempt teaching in this area, I won't attempt to recommend textbooks.
0 Kudos
Reply