Hi,
after figuring out how to use ModelingToolkit to create functions/derivates etc. I would like to run code generated by ModelingToolkit in parallel. E.g. solving the Lorenz system for multiple initial states, see the following example:
using Distributed
using ModelingToolkit
using DifferentialEquations
# create the Lorenz system
eqs_mt = []
push!(eqs_mt, dt(x) ~ lorenz_mt(x,y,z)[1])
push!(eqs_mt, dt(y) ~ lorenz_mt(x,y,z)[2])
push!(eqs_mt, dt(z) ~ lorenz_mt(x,y,z)[3])
de_mt = ODESystem(eqs_mt)
ode_func_mt = ODEFunction(de_mt, [x,y,z], [sigma,rho,beta])
u0 = [19.,20.,50.]
params = [16.,45.92,4]
ode_prob_mt = ODEProblem(ode_func_mt, u0, (0., 10.),params)
# add processes to workspace
addprocs(2)
@everywhere begin
using DifferentialEquations
using ModelingToolkit
function solve_lorenz(ode_problem)
print(solve(ode_problem))
end
end
This code errors with
On worker 2:
UndefVarError: ###375 not defined
deserialize_datatype at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:1115
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:771
deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
deserialize_datatype at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:1139
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:771
deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
deserialize_datatype at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:1139
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:771
deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
deserialize_datatype at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:1139
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:771
deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:778
deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
#3 at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:869
ntuple at ./tuple.jl:136
deserialize_tuple at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:869
handle_deserialize at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:761
deserialize_msg at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Serialization/src/Serialization.jl:731
#invokelatest#1 at ./essentials.jl:742 [inlined]
invokelatest at ./essentials.jl:741 [inlined]
message_handler_loop at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Distributed/src/process_messages.jl:160
process_tcp_streams at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.1/Distributed/src/process_messages.jl:117
#105 at ./task.jl:259
I assume that the generated function, with name ###375, is not available on the worker process as it was created on the main process. Is there a possibility to make this function available to worker processes without calling ODEFunction? I would like to avoid costly generation of Jacobians, etc. on each worker process, when they are already precalculated on the main process.