I have a linear problem with a sparse, psd and strictly diagonally dominant matrix. Can you please point me to some known best solvers (in terms of runtime, or easy to be practically optimized for multicore and SIMD system)?

Currently my direction was to try conjugate gradient method, and it seems to solve the problem in about 8-10 iterations (for specific use case), which is pretty good. However I wonder if there are other solvers that take the advantages of those characteristics even better.

Much thanks!

  • $\begingroup$ When you say linear problem, do you mean linear objective function and linear constraints? $\endgroup$ Aug 6 '14 at 14:03
  • $\begingroup$ And in general, do consider commercial or free solver? $\endgroup$ Aug 6 '14 at 14:03
  • $\begingroup$ @AC_MOSEK - yes, exactly. linear objective function and a set of linear constraints. Commercial solvers can be applied. Thanks! $\endgroup$
    – rursw1
    Aug 10 '14 at 13:04
  • $\begingroup$ I would say that there are no advantages you can take from the problem structure, except for sparsity, afaik. Sparsity is usually exploited by every decent LP solver. Please do not use gradient base method in this case! $\endgroup$ Aug 11 '14 at 5:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.