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
It would be great to be able to configure solver parameters like a timeout (in terms of cpu time, clock time, etc) as mentioned in #97. Other configuration options could be useful as well. Solvers that call library functions from JuMP, etc, should pass parameters like timeout along.
The text was updated successfully, but these errors were encountered:
Hi
I spent a lot of time trying to make a timeout using Julia, but I failed.
Finally, I used python's subprocess library, but it's not ideal. My code is:
`
try:
max_time = 150 # 150 Sec
output = subprocess.check_output(['Julia', 'juliafile.jl'], timeout=max_time).splitlines()
# do something
except subprocess.TimeoutExpired:
print("the process was too long")
Thanks!
I think what we will ultimately do is something along the lines of adding a SolverConfig struct with the relevant fields (TBD). For solvers that call JuMP, we can pass time limits along to JuMP.set_time_limit_sec. For other solvers we may have to keep track of time ourselves.
Although I've never looked into it, I think also calling yield() in the solvers can allow them to be run asynchronously and terminated externally after a set amount of time. This could also be a direction to look at.
It would be great to be able to configure solver parameters like a timeout (in terms of cpu time, clock time, etc) as mentioned in #97. Other configuration options could be useful as well. Solvers that call library functions from JuMP, etc, should pass parameters like timeout along.
The text was updated successfully, but these errors were encountered: