You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The solvers accept an InitialPtFile to provide a starting point. This could be hooked up to MOI to allow the user to specify an initial search point (aka a warmstart). This probably wouldn't be too hard because we can reuse some of our printing code.
The text was updated successfully, but these errors were encountered:
It would be cool also to have some way to load the problem using e.g. BigFloats, then write + solve it at 64 bit precision with plain SDPA, use that solution as an initial point, and then solve it again at high precision with SDPA-GMP. I wonder if that could end up a lot quicker than just using SDPA-GMP the whole way.
We could probably do it in a such a way that the problem data doesn't have to get loaded into the Optimizer object twice, and of course only do the presolve once as well.
The solvers accept an
InitialPtFile
to provide a starting point. This could be hooked up to MOI to allow the user to specify an initial search point (aka a warmstart). This probably wouldn't be too hard because we can reuse some of our printing code.The text was updated successfully, but these errors were encountered: