Julia script running on Linux is slow

alka · October 16, 2019, 10:06pm

I am trying to run julia code in linux machine. But it takes around 4-5 mins to execute. How can I make it faster?

Oscar_Smith · October 16, 2019, 10:12pm

What is the script? Hard to debug something you can’t see. Also, when you say slow on linux, is that compared to Windows/mac or just how you think they should perform?

alka · October 16, 2019, 10:16pm

workspace();

include(“OControllerProx.jl”)
using .OControllerProx;
#include(“Process.jl”)
#using .Process;

using MathProgBase;
using LinearAlgebra;
using MAT;
using DataFrames;
using CSV;
using JuMP;
using ProxSDP;
using Dates;

get the calculated switches status

path = “/Applications/projects/cLEAN/pnnl_control”;

file = matopen(“$(path)/input/one-time-input/switch.mat”);
switches_status = read(file,“switches_status”);

case=CSV.read(“$(path)/input/one-time-input/case.csv”);

#case = parse(Float64,CSV.read(“$(path)/input/one-time-input/case.csv”));

#LoadInformation = CSV.File(“(path)/input/one-time-input/LoadInformation.xls"; normalizenames=true, types=[String, String, String, String, String, String, String, String]); #Loads=CSV.read("(path)/input/one-time-input/LoadInformation.xls”; normalizenames=true)

#using DataFrames
#using StringEncodings
#f=open(“$(path)/input/one-time-input/LoadInformation.xls”,“r”);
#s=StringDecoder(f,“LATIN1”, “UTF-8”);
#LoadInformation=CSV.read(s; normalizenames=true, comment=“#”);
#close(s)
#close(f)

#printf(Loads[2,2])

#(Bbus_processed,Gbus_processed) = Process.GBbus(Ybus, NodeIDs, NumberOfElements);
#Bbus_energizing=switches_status.*Bbus_processed;
#Gbus_energizing=switches_status.*Gbus_processed;

get 5-minute actions

actions = CSV.read(“$(path)/input/5-min-input/actions.csv”; comment=“#”);
optimal_injections=zeros(ComplexF64,8,1);
for i=1:8
optimal_injections[i]=parse(ComplexF64, actions[1,4*i+3]); # convert from string to complex number
end
number_of_energizing_segment=length(actions[:,1]);

pnnl_actions csv file

global df = DataFrame(Dict(“datatime” =>“”, “generator.proposed_diesel_1.1”=> “”, “generator.proposed_diesel_1.2” => “”, “generator.proposed_diesel_1.3” => “”, “generator.ouc_solar.1” => “”,“generator.ouc_solar.2” => “”,“generator.ouc_solar.3” => “”,“generator.gen_none_-60027406210.1” => “”,“generator.gen_none_-60027406210.2” => “”, “generator.gen_none_-60027406210.3” => “”,“generator.proposed_solarpv_1.1” => “”,“generator.proposed_solarpv_1.2” => “”,“generator.proposed_solarpv_1.3” => “”,“generator.gen_none_6002693162.1” => “”,“generator.gen_none_6002693162.2” => “”,“generator.gen_none_6002693162.3” => “”,“storage.proposed_storage_1.1” => “”,“storage.proposed_storage_1.2” => “”,“storage.proposed_storage_1.3” => “”,“storage.proposed_fuelcell_1.1” => “”,“storage.proposed_fuelcell_1.2” => “”,“storage.proposed_fuelcell_1.3” => “”,“storage.proposed_fuelcell_2.1” => “”,“storage.proposed_fuelcell_2.2” => “”,“storage.proposed_fuelcell_2.3” => “”));

save pnnl dynamic actions to the output folder

CSV.write(“$(path)/output/pnnl_actions.csv”, df);

nx=24; # 3 x number of (gen+storage)
mx=16; # number of control actions
px=4; # number of usable measurements
A=-Matrix{Float64}(I, 24,24);
B=[Matrix{Float64}(I, 16,16);zeros(8,16)];
C=zeros(4,24);C[1,3]=1;C[2,6]=1;C[3,9]=1;C[4,12]=1;

finding controller and observer gains

Ax=A;
Bx=B;
Cx=C;

Delta1=1;
Delta2=1;
gamma = (Delta1+Delta2)/5;
OptCtr = OControllerProx.solveControllerObserver(Ax,Bx,Cx,gamma, Delta1, Delta2,nx,mx,px);
P_out=OptCtr[1]; # Lyapunov matrix
K_out=OptCtr[2];
Q_out=OptCtr[3]; # Lyapunov matrix
L_out=OptCtr[4]; # output control gain
##############################

getting control signals to apply to control nodes in every 1 second for 5 minutes

current_observer=zeros(24,1);

global sensor_Mtr_382883
if case==1

for times = 1:number_of_energizing_segment
for i=1:300

# remember to update actions when running for > 5 minute

Step1: Get measurement data from measurement input folder

#sensor_Mtr_382883 =readtable(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”)
#df1 = CSV.File(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”, datarow=10);
#sensor_Mtr_382883 =CSV.read(df1)

current_measurement=zeros(4,1);
if isfile(“(path)/input/measurement-input/sensor_Mtr_-382883.csv")==false current_measurement[1]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_1695832.csv”)==false
current_measurement[2]=0;
elseif isfile(“(path)/input/measurement-input/sensor_Mtr_118122225.csv")==false current_measurement[3]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_119123112.csv”)==false
current_measurement[4]=0;
else

sensor_Mtr_382883 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_-382883.csv"; comment="#"); sensor_Mtr_1695832 =CSV.read("(path)/input/measurement-input/sensor_Mtr_1695832.csv”; comment=“#”);
sensor_Mtr_118122225 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_118122225.csv"; comment="#"); sensor_Mtr_119123112 =CSV.read("(path)/input/measurement-input/sensor_Mtr_119123112.csv”; comment=“#”);

check if data at i-second instant is available; if not then wait until it is available

while (length(sensor_Mtr_382883[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_1695832[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_118122225[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_119123112[:,1])<(i+300*(times-1)))
sleep(0.001);
end

calculate control input based on measurement at i-second instant

if typeof(sensor_Mtr_382883[2,2])==string

convert from string to complex number measurement data

imcurrent_measurement=zeros(ComplexF64,4,1);
imcurrent_measurement[1]= (parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),4]))/3;
imcurrent_measurement[2]= (parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),4]))/3;
imcurrent_measurement[3]= (parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),4]))/3;
imcurrent_measurement[4]= (parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),4]))/3;
for j=1:4
current_measurement[j]= abs(imcurrent_measurement[j]);
end

elseif typeof(sensor_Mtr_382883[2,2])==Int64
current_measurement[1]= sqrt(sensor_Mtr_382883[(i+300*(times-1)),8]^2+sensor_Mtr_382883[(i+300*(times-1)),9]^2);
current_measurement[2]= sqrt(sensor_Mtr_1695832[(i+300*(times-1)),8]^2+sensor_Mtr_1695832[i(i+300*(times-1)),9]^2);
current_measurement[3]= sqrt(sensor_Mtr_118122225[(i+300*(times-1)),8]^2+sensor_Mtr_118122225[(i+300*(times-1)),9]^2);
current_measurement[4]= sqrt(sensor_Mtr_119123112[(i+300*(times-1)),8]^2+sensor_Mtr_119123112[(i+300*(times-1)),9]^2);
end
end

###############################

find the next observer state

global current_observer
next_observer = current_observer + (A+BK_out)current_observer + L_out(current_measurement-Ccurrent_observer);

current_observer =next_observer;

################################

Step 3: return PQ control signals bn

current_control = K_outcurrent_observer;
PQcurrent_control=zeros(ComplexF64,8,1);
for k=1:8
PQcurrent_control[k] = current_control[2k-1] + current_control[2*k]*im;
end
################################

Step 4: calculate the changes to apply to control nodes

dynamic_injections = optimal_injections + PQcurrent_control;

write the file with the stringdata variable information

global df = DataFrame(Dict(“datatime” => DateTime(2000,1,1,0,floor((i+300*(times-1))/60),(i+300*(times-1))-60floor((i+300(times-1))/60)), “generator.proposed_diesel_1.1” => [dynamic_injections[1]], “generator.proposed_diesel_1.2” => [dynamic_injections[1]], “generator.proposed_diesel_1.3” => [dynamic_injections[1]], “generator.ouc_solar.1” => [dynamic_injections[2]],“generator.ouc_solar.2” => [dynamic_injections[2]],“generator.ouc_solar.3” => [dynamic_injections[2]],“generator.gen_none_-60027406210.1” => [dynamic_injections[3]],“generator.gen_none_-60027406210.2” => [dynamic_injections[3]], “generator.gen_none_-60027406210.3” => [dynamic_injections[3]],“generator.proposed_solarpv_1.1” => [dynamic_injections[4]],“generator.proposed_solarpv_1.2” => [dynamic_injections[4]],“generator.proposed_solarpv_1.3” => [dynamic_injections[4]],“generator.gen_none_6002693162.1” => [dynamic_injections[5]],“generator.gen_none_6002693162.2” => [dynamic_injections[5]],“generator.gen_none_6002693162.3” => [dynamic_injections[5]],“storage.proposed_storage_1.1” => [dynamic_injections[6]],“storage.proposed_storage_1.2” => [dynamic_injections[6]],“storage.proposed_storage_1.3” => [dynamic_injections[6]],“storage.proposed_fuelcell_1.1” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.2” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.3” => [dynamic_injections[7]],“storage.proposed_fuelcell_2.1” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.2” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.3” => [dynamic_injections[8]]));

save pnnl dynamic actions to the output folder

CSV.write(“$(path)/output/pnnl_actions.csv”, df,append=true);

end
end

elseif case==2 #incorrect measurement data
for times = 1:number_of_energizing_segment
for i=1:300

  # remember to update actions when running for > 5 minute

Step1: Get measurement data from measurement input folder

#sensor_Mtr_382883 =readtable(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”)
#df1 = CSV.File(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”, datarow=10);
#sensor_Mtr_382883 =CSV.read(df1)
current_measurement=zeros(4,1);
if isfile(“(path)/input/measurement-input/sensor_Mtr_-382883.csv")==false current_measurement[1]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_1695832.csv”)==false
current_measurement[2]=0;
elseif isfile(“(path)/input/measurement-input/sensor_Mtr_118122225.csv")==false current_measurement[3]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_119123112.csv”)==false
current_measurement[4]=0;
else

sensor_Mtr_382883 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_-382883.csv"; comment="#"); sensor_Mtr_1695832 =CSV.read("(path)/input/measurement-input/sensor_Mtr_1695832.csv”; comment=“#”);
sensor_Mtr_118122225 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_118122225.csv"; comment="#"); sensor_Mtr_119123112 =CSV.read("(path)/input/measurement-input/sensor_Mtr_119123112.csv”; comment=“#”);

check if data at i-second instant is available; if not then wait until it is available

while (length(sensor_Mtr_382883[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_1695832[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_118122225[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_119123112[:,1])<(i+300*(times-1)))
sleep(0.001);
end

calculate control input based on measurement at i-second instant

if typeof(sensor_Mtr_382883[2,2])==string

convert from string to complex number measurement data

imcurrent_measurement=zeros(ComplexF64,4,1);
imcurrent_measurement[1]= (parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),4]))/3;
imcurrent_measurement[2]= (parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),4]))/3;
imcurrent_measurement[3]= (parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),4]))/3;
imcurrent_measurement[4]= (parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),4]))/3;
current_measurement[1]=(1.05-0.1*rand(Float64))*abs(imcurrent_measurement[1]);
for j=2:4
current_measurement[j]= abs(imcurrent_measurement[j]);
end

elseif typeof(sensor_Mtr_382883[2,2])==Int64
current_measurement[1]= (1.05-0.1rand(Float64))sqrt(sensor_Mtr_382883[(i+300(times-1)),8]^2+sensor_Mtr_382883[(i+300(times-1)),9]^2);
current_measurement[2]= sqrt(sensor_Mtr_1695832[(i+300*(times-1)),8]^2+sensor_Mtr_1695832[i(i+300*(times-1)),9]^2);
current_measurement[3]= sqrt(sensor_Mtr_118122225[(i+300*(times-1)),8]^2+sensor_Mtr_118122225[(i+300*(times-1)),9]^2);
current_measurement[4]= sqrt(sensor_Mtr_119123112[(i+300*(times-1)),8]^2+sensor_Mtr_119123112[(i+300*(times-1)),9]^2);
end
end

###############################

find the next observer state

global current_observer
next_observer = current_observer + (A+BK_out)current_observer + L_out(current_measurement-Ccurrent_observer);

current_observer =next_observer;

################################

Step 3: return PQ control signals bn

current_control = K_outcurrent_observer;
PQcurrent_control=zeros(ComplexF64,8,1);
for k=1:8
PQcurrent_control[k] = current_control[2k-1] + current_control[2*k]*im;
end
################################

Step 4: calculate the changes to apply to control nodes

dynamic_injections = optimal_injections + PQcurrent_control;

write the file with the stringdata variable information

global df = DataFrame(Dict(“datatime” => DateTime(2000,1,1,0,floor((i+300*(times-1))/60),(i+300*(times-1))-60floor((i+300(times-1))/60)), “generator.proposed_diesel_1.1” => [dynamic_injections[1]], “generator.proposed_diesel_1.2” => [dynamic_injections[1]], “generator.proposed_diesel_1.3” => [dynamic_injections[1]], “generator.ouc_solar.1” => [dynamic_injections[2]],“generator.ouc_solar.2” => [dynamic_injections[2]],“generator.ouc_solar.3” => [dynamic_injections[2]],“generator.gen_none_-60027406210.1” => [dynamic_injections[3]],“generator.gen_none_-60027406210.2” => [dynamic_injections[3]], “generator.gen_none_-60027406210.3” => [dynamic_injections[3]],“generator.proposed_solarpv_1.1” => [dynamic_injections[4]],“generator.proposed_solarpv_1.2” => [dynamic_injections[4]],“generator.proposed_solarpv_1.3” => [dynamic_injections[4]],“generator.gen_none_6002693162.1” => [dynamic_injections[5]],“generator.gen_none_6002693162.2” => [dynamic_injections[5]],“generator.gen_none_6002693162.3” => [dynamic_injections[5]],“storage.proposed_storage_1.1” => [dynamic_injections[6]],“storage.proposed_storage_1.2” => [dynamic_injections[6]],“storage.proposed_storage_1.3” => [dynamic_injections[6]],“storage.proposed_fuelcell_1.1” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.2” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.3” => [dynamic_injections[7]],“storage.proposed_fuelcell_2.1” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.2” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.3” => [dynamic_injections[8]]));

save pnnl dynamic actions to the output folder

CSV.write(“$(path)/output/pnnl_actions.csv”, df,append=true);

end
end

elseif case==3 #failure actuator

for times = 1:number_of_energizing_segment
for i=1:300

  # remember to update actions when running for > 5 minute

Step1: Get measurement data from measurement input folder

#sensor_Mtr_382883 =readtable(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”)
#df1 = CSV.File(“/Applications/projects/cLEAN/pnnl control/measurement input/sensor_Mtr_-382883.csv”, datarow=10);
#sensor_Mtr_382883 =CSV.read(df1)
current_measurement=zeros(4,1);
if isfile(“(path)/input/measurement-input/sensor_Mtr_-382883.csv")==false current_measurement[1]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_1695832.csv”)==false
current_measurement[2]=0;
elseif isfile(“(path)/input/measurement-input/sensor_Mtr_118122225.csv")==false current_measurement[3]=0; elseif isfile("(path)/input/measurement-input/sensor_Mtr_119123112.csv”)==false
current_measurement[4]=0;
else

sensor_Mtr_382883 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_-382883.csv"; comment="#"); sensor_Mtr_1695832 =CSV.read("(path)/input/measurement-input/sensor_Mtr_1695832.csv”; comment=“#”);
sensor_Mtr_118122225 =CSV.read(“(path)/input/measurement-input/sensor_Mtr_118122225.csv"; comment="#"); sensor_Mtr_119123112 =CSV.read("(path)/input/measurement-input/sensor_Mtr_119123112.csv”; comment=“#”);

check if data at i-second instant is available; if not then wait until it is available

while (length(sensor_Mtr_382883[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_1695832[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_118122225[:,1])<(i+300*(times-1))) & (length(sensor_Mtr_119123112[:,1])<(i+300*(times-1)))
sleep(0.001);
end

calculate control input based on measurement at i-second instant

if typeof(sensor_Mtr_382883[2,2])==string

convert from string to complex number measurement data

imcurrent_measurement=zeros(ComplexF64,4,1);
imcurrent_measurement[1]= (parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_382883[(i+300*(times-1)),4]))/3;
imcurrent_measurement[2]= (parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_1695832[(i+300*(times-1)),4]))/3;
imcurrent_measurement[3]= (parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_118122225[(i+300*(times-1)),4]))/3;
imcurrent_measurement[4]= (parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),2])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),3])+parse(ComplexF64, sensor_Mtr_119123112[(i+300*(times-1)),4]))/3;
for j=1:4
current_measurement[j]= abs(imcurrent_measurement[j]);
end

elseif typeof(sensor_Mtr_382883[2,2])==Int64
current_measurement[1]= sqrt(sensor_Mtr_382883[(i+300*(times-1)),8]^2+sensor_Mtr_382883[(i+300*(times-1)),9]^2);
current_measurement[2]= sqrt(sensor_Mtr_1695832[(i+300*(times-1)),8]^2+sensor_Mtr_1695832[i(i+300*(times-1)),9]^2);
current_measurement[3]= sqrt(sensor_Mtr_118122225[(i+300*(times-1)),8]^2+sensor_Mtr_118122225[(i+300*(times-1)),9]^2);
current_measurement[4]= sqrt(sensor_Mtr_119123112[(i+300*(times-1)),8]^2+sensor_Mtr_119123112[(i+300*(times-1)),9]^2);
end
end

###############################

find the next observer state

global current_observer;
next_observer = current_observer + (A+BK_out)current_observer + L_out(current_measurement-Ccurrent_observer);

current_observer =next_observer;

################################

Step 3: return PQ control signals bn

current_control = K_outcurrent_observer;
PQcurrent_control=zeros(ComplexF64,8,1);
for k=1:8
PQcurrent_control[k] = current_control[2k-1] + current_control[2*k]*im;
end

PQcurrent_control[1]=0; #actuators at generator 1 fail

################################

Step 4: calculate the changes to apply to control nodes

dynamic_injections = optimal_injections + PQcurrent_control;

write the file with the stringdata variable information

global df = DataFrame(Dict(“datatime” => DateTime(2000,1,1,0,floor((i+300*(times-1))/60),(i+300*(times-1))-60floor((i+300(times-1))/60)), “generator.proposed_diesel_1.1” => [dynamic_injections[1]], “generator.proposed_diesel_1.2” => [dynamic_injections[1]], “generator.proposed_diesel_1.3” => [dynamic_injections[1]], “generator.ouc_solar.1” => [dynamic_injections[2]],“generator.ouc_solar.2” => [dynamic_injections[2]],“generator.ouc_solar.3” => [dynamic_injections[2]],“generator.gen_none_-60027406210.1” => [dynamic_injections[3]],“generator.gen_none_-60027406210.2” => [dynamic_injections[3]], “generator.gen_none_-60027406210.3” => [dynamic_injections[3]],“generator.proposed_solarpv_1.1” => [dynamic_injections[4]],“generator.proposed_solarpv_1.2” => [dynamic_injections[4]],“generator.proposed_solarpv_1.3” => [dynamic_injections[4]],“generator.gen_none_6002693162.1” => [dynamic_injections[5]],“generator.gen_none_6002693162.2” => [dynamic_injections[5]],“generator.gen_none_6002693162.3” => [dynamic_injections[5]],“storage.proposed_storage_1.1” => [dynamic_injections[6]],“storage.proposed_storage_1.2” => [dynamic_injections[6]],“storage.proposed_storage_1.3” => [dynamic_injections[6]],“storage.proposed_fuelcell_1.1” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.2” => [dynamic_injections[7]],“storage.proposed_fuelcell_1.3” => [dynamic_injections[7]],“storage.proposed_fuelcell_2.1” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.2” => [dynamic_injections[8]],“storage.proposed_fuelcell_2.3” => [dynamic_injections[8]]));

save pnnl dynamic actions to the output folder

CSV.write(“$(path)/output/pnnl_actions.csv”, df,append=true);

end
end

end

alka · October 16, 2019, 10:18pm

When I run the same in Atom it takes few seconds but in linux VM with julia takes 4-5 minutes

aaowens · October 16, 2019, 10:37pm

Is the linux VM reset before you run it? Maybe you’re precompiling all the packages on the VM, which might take a few minutes. In Atom they might already be precompiled.

What happens if you run the script twice in the VM?

alka · October 16, 2019, 10:42pm

I tried running it twice but same speed. 3 minutes both the times. I also get this

sing994@eioc-tdc-02:~/csderms/csderms-docker/scripts$ docker exec csv1 julia StabilizingControl.jl
┌ Warning: Package ProxSDP does not have Random in its dependencies:
│ - If you have ProxSDP checked out for development and have
│ added Random as a dependency but haven’t updated your primary
│ environment’s manifest file, try Pkg.resolve().
│ - Otherwise you may need to report an issue with ProxSDP
└ Loading Random into ProxSDP from project dependency, future warnings for ProxSDP are suppressed.

Is this causing an issue?

jling · October 16, 2019, 11:01pm

in atom, your kernel doesn’t shut down between runs, on linux, if you run it by julia script.jl, every time a new Julia kernel needs to start.

Also, please wrap your code between ``` for readability.

alka · October 16, 2019, 11:09pm

I have to use docker to excecute this Julia script. I use this command -
docker exec csv1 julia StabilizingControl.jl

So, do you have any suggestions to make it faster?

jling · October 16, 2019, 11:13pm

because Julia is not a good candidate for “scripting language” (think of Bash, Python, Perl), you either can just make your “script” a module, so you can pre-compile most of the things away, or you can set up a Jupyter notebook and forward a port out from docker.

Ideally there should be a way to run a Julia kernel in the background, but I’m not sure what’s the easiest way to do this. (maybe Distributed.jl)

johnh · October 17, 2019, 8:51am

Please could you tell us what your exact environment is?
The approximate specs of the base server or laptop and the operating system.
Are you runnign docker on this base system, or are you running a Linxu VM then docker?

As an aside has anyone used docker pause with a Julia docker container?