> Just to confirm my understanding, are you saying if I make the > changes you describe below, that as long as each call to the solver > referencing the GLPK data structure instances for a single LP problem > are made from the one thread, i.e. sequentially, I'm fine?
Yes. Please see an implementation example: http://lists.gnu.org/archive/html/help-glpk/2009-12/msg00008.html (note that the routine names were changed in v4.43; you need to change the glpenv02.c module). > I only run into thread safety issues if I were to create multiple > threads to make GLPK calls on the same LP problem. > In summary, one thread per LP problem is fine? As this is what I > want, I only need to be able to run X threads solving X LP problems, > one each concurrently to be at the MT level I require. > I do not require MT safe access to a single LP problem. _______________________________________________ Help-glpk mailing list Help-glpk@gnu.org http://lists.gnu.org/mailman/listinfo/help-glpk