Global variables are inefficient and. make errors difficult to diagnose

I'm writing code that builds the Vortex Lattice method, but I'm having trouble calling the variables in my Main code. I'm getting the error in the title on every code and I don't understand why.

Answer 1

[Cris LaPierre]

Why not pass the variables as inputs and outputs to your functions? I did this by putting all the code into one script, then used the Refactor Code option to break each piece out to its own local function. They could just as easily be function files as well.

%--------------------------------------------------------------------------
% Parametreler
%--------------------------------------------------------------------------
c = 1;  % Veter uzunluğu
b = 1;  % Kanat açıklığı
NJ = 10;  % Panel sayısı Y
NI = 1;  % Panel sayısı X
NJ1 = NJ+1;
NI1 = NI + 1;
Gamma = 1;

P = Geometry(NI1, NJ1, b, NJ, c, NI);
[G, C] = Panels(NJ1, P, NJ);
[PC, P1, P2] = DGP(NI, NJ, C, G);
Vor3D(PC, P1, P2, Gamma);

function P = Geometry(NI1, NJ1, b, NJ, c, NI)
P = zeros(NI1, NJ1, 3);
dy = b/NJ;
dx = c/NI;
for j = 1:NJ1
    for i = 1:NI1
        P(i,j,:) = [(i-1)*dx, (j-1)*dy, 0];
    end
end
end

function [G, C] = Panels(NJ1, P, NJ)
%--------------------------------------------------------------------------
%   Girdap uç noktalari 
%--------------------------------------------------------------------------
B_alpha_deg = 10;  % Derece olarak
L_uzunluk   = 10;  % L'nin uzunluÄŸu
L_vektor    = [L_uzunluk * cosd(B_alpha_deg), 0, L_uzunluk * sind(B_alpha_deg)]';

G = zeros(3, NJ1, 3);
for j = 1:NJ1
    G(1,j,:) = P(1,j,:) + 0.25*(P(2,j,:) - P(1,j,:));
    G(2,j,:) = P(2,j,:);
    G(3,j,:) = squeeze(G(2,j,:)) + L_vektor;
end
% disp (G);

%--------------------------------------------------------------------------

%--------------------------------------------------------------------------
C = zeros(1, NJ, 3);  

for j = 1:NJ
    CL = P(1,j  ,:) + 0.75*(P(2,j  ,:) - P(1,j  ,:));
    CR = P(1,j+1,:) + 0.75*(P(2,j+1,:) - P(1,j+1,:));
    C(1,j,:) = (CL + CR) / 2;
    % C(1,j,:) = (P(1,j,:) + 0.75*(P(2,j,:) - P(1,j,:)) + P(1,j+1,:) + 0.75*(P(2,j+1,:) - P(1,j+1,:))) / 2;
end
% disp (C);

%--------------------------------------------------------------------------
% Yuzey birim normal vektorlerini hesaplama
%--------------------------------------------------------------------------
zeta = zeros(1, NJ, 3); 

for j = 1:NJ
    R1 = squeeze(P(1, j, :) - P(2, j+1, :));
    R2 = squeeze(P(2, j , :) - P(1, j+1, :))';

    R12 = cross(R1, R2);
    normal_buyukluk = norm(R12);

    zeta(1, j, :) = R12 / normal_buyukluk;
end
end

function [PC, P1, P2] = DGP(NI, NJ, C, G)
for i = 1:NI
    for j = 1:NJ

        PC = squeeze(C(i,j,:));
        P1 = squeeze(G(i,j  ,:));
        P2 = squeeze(G(i,j+1,:));

    end
end
end

function Vor3D(PC, P1, P2, Gamma)
r1 = PC - P1;
r2 = PC - P2;
r0 = P2 - P1;
r12 = cross(r1, r2);
r1r2       = r12 / (norm(r12))^2;
Term2      = (r1 / norm(r1)) - (r2 / norm(r2));
r0Term2    = dot(r0, Term2);
DeltaV     = Gamma / (4*pi) * r1r2 * r0Term2;
end

[Volkan]

I work with a mentor and he wanted me to call the function command by giving the name of the file, and he wanted me to make it universal for the entire code by calling the inputs and outputs with the global command.

That solution is working too thank u so much for support but , probably my mentor will be want to convert my codes to global variation.

Answer 2

[Walter Roberson]

The message is a warning, not an error message.

You are getting the warning message because using global variables is the most expensive form of accessing variables. Literally every other possibility is checked before checking whether a variable is a global. That makes accessing global variables relatively slow.

You are also getting that warning message because historically people have found using global variables to be error-prone, and that it can be quite difficult to track down where global variables are changed. They might not be bad when they are set in a single block of code and are read-only in all other blocks of code, but much more typical is that they end up being modified in several places.

Because the use of global variables is slow and error-prone, the editor gives a warning about them. It is an opportunity to reflect upon whether you really need global variables, and upon whether you have taken precautions to ensure that the global variables are being used correctly.

You can turn off the warning. Right-click on the word "global" and select "Suppress Message" and then either "on this line" or "in this file". But turning off the warning... just turns off the warning, without fixing any of the underlying problems.

[Volkan]

Thanks for the information, I got the same advice wherever I go. I think my code will have two different paths and I will learn to work with the global command and face this code in the future.