Errors in converting to C code

Question

Maurilio Matracia 2022-9-18

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1807355-errors-in-converting-to-c-code

回答： Yash 2023-8-30

Screenshot 2022-09-12 104857.jpg

在 MATLAB Online 中打开

Hello everyone,

I have the following code that I want to speed up by converting to C (I never did that before):

function [MD_a,mismatch_SminusA] = metaDistr_fun
tic
zero = 1e-6  ;     Pi = 3.14159 ;
alpha = 3.5 ;
m = 1 ;
iterations = 3e1 ;
instants = 9e0 ; 
sigma2 = 1e-11 ;    
lambda0 = 1e-6 ;
Rd  = 0.5e3 ;  
Ra = 840  ;        
Rb = 3286 ;   
infty = 2e4 ;
xi = 20 ;   
Chi = 0 ;
tau = 0.2 ;
 X = linspace(0,1, 1) ;  nX = length(X) ;
 pTag = zeros(1,2) ; coord = {[0 0];[0 0]} ; ED = {0;0} ; EDmin = ED ;  Hc = ED ; pr =0 ;
 Ru=Rd
MD_s = zeros(1,nX) ;   reliab = zeros(1,iterations) ;  SINR = zeros(iterations,instants) ;
for i = 1:iterations
nD = poissrnd (Chi * lambda0 * pi * (Rd^2-0^2)) ;   
thetaD = rand(nD,1) * 2*pi ;    
rD = sqrt(rand(nD,1)*(Rd^2-0^2) + 0^2) ;          
coord{1} = [ rD.*cos(thetaD) ,  rD.*sin(thetaD) ] ;     
    nTa = poissrnd(lambda0 * pi * (Ra^2-Rd^2)); 
    nTb = poissrnd(lambda0 * pi * (infty^2-Rb^2)) ;
    thetaT = rand(nTa+nTb,1) * 2*pi ;
    rT = [sqrt(rand(nTa,1)*(Ra^2-Rd^2) + Rd^2); sqrt(rand(nTb,1)*(infty^2-Rb^2) + Rb^2)] ;  
    coord{2} = [ rT.*cos(thetaT) ,  rT.*sin(thetaT) ] ;   
        for B = 1:2                   
        ED{B} = sqrt( (Ru-coord{B}(:, 1)).^2 + coord{B}(:, 2).^2 ) ;                
               if isempty(ED{B})
               ED{B} = infty ;   EDmin{B} = infty  ;   
               else
               EDmin{B} = min(ED{B})  ;   
               end       
        pTag(B) = xi * EDmin{B}^-alpha;  
        Hc{B} = gamrnd( m,1/m,  length(ED{B}),instants )   ;
        end
tag = find( pTag == max(pTag) ) ;      
    for t = 1:instants
    pr = Hc{tag}( find(ED{tag}==EDmin{tag}),t ) *xi*EDmin{tag}^-alpha ;
    PtotD = xi * sum(Hc{1}(:,t).*ED{1}.^-alpha)  ;
    Ptot = PtotD + xi * sum(Hc{2}(:,t).*ED{2}.^-alpha) ;
    SINR(i,t) = pr/(Ptot-pr+sigma2) ;
    end
reliab(i) = mean(SINR(i,:)>tau) ;
end 
    for x = 1 : nX
    MD_s(x) = mean(reliab >= X(x)) ;
    end
end

However, I am getting the following errors (screenshot attached), could anyone suggest me how to fix them?

Thanks in advance!

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Yash 2023-8-30

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1807355-errors-in-converting-to-c-code#answer_1297071

Hi Maurilio,

I was able to reproduce the error using the code you provided. After analysing your code, I noticed that the cell array 'ED' changes its shape in each iteration (line 32).

You can confirm this by removing the semicolon at the end of line 32.

It's important to note that not all MATLAB codes can be converted into C code using the MATLAB Coder. Since C does not support variable-size arrays, the variable size of 'ED' cannot be converted from MATLAB to C.

One possible workaround is to use padding. You can add extra zeros at the end of the arrays to fix their size. Preallocate the size for the arrays by creating an array of zeros or ones and avoid changing their size within the loop.