Hello, I’m looking for multithreading for my Benders cuts: indeed, the subproblem can be decomposed in several independent subproblems that can be executed in parallel (i use CPLEX solver). When I run my program, my notebook tells me that the kernel has crashed. Could you tell me why?
Assuming that you separate the Benders cuts via callbacks let me point you to the relevant lines in the example of the CPLEX.jl readme:
# This is very, very important!!! Only use callbacks in single-threaded mode. MOI.set(model, MOI.NumberOfThreads(), 1)
However, I do not know if this also applies for auto-Benders.
Hi there! Since this is your first post, please read this announcement: PSA: make it easier to help you.
It’s easier to help if you can provide a reproducible example.
This StackOverflow question should be helpful: multithreading - Create JuMP model with multi-threading in Julia - Stack Overflow.
(I need to add something on this to the JuMP documentation: Add documentation on parallel solves of a JuMP model · Issue #2348 · jump-dev/JuMP.jl · GitHub