Error while vectorizing my code

x = 0:0.01:5;  %% range of x values for graph
y = zeta_func(x);
plot(x,y),xlabel('x'), ylabel('zeta(x)'), title('Zeta(x) Graph'),

%% function definition 
function zeta_val = zeta_func(x)    
    %% zeta(x) = summation of 1/x^n
    %% here summing from n = 1 to 50
    n = 1:1:50;
    terms = 1./power(n,x);        %% performing 1/x^i for all n   
    zeta_val = sum(terms);        %% summing the values
end

I get the following error

Error using .^

Matrix dimensions must agree.

Error in zeta_gpu>zeta_func (line 11)

terms = 1./power(n,x); %% performing 1/x^i for all n

Error in zeta_gpu (line 2)

y = zeta_func(x);

Over here the function is being called on the whole array x. But I want to do element wise.

How do I acheive this over here?

Answer 1

[Cris LaPierre]

The error is because your vectors x and n are not the same length. x has 501 elements while n has 50. Also, power is doing n.^x. From how you worded things, perhaps you want x.^n?

Assuming you want every n to be raised to every x, I would look into using meshgrid. You can set it up so the rows represent unique values of x and the columns unique values of n (or vice versa). Then just sum by each x so you can create your plot.

x = 0:0.01:5;  %% range of x values for graph
y = zeta_func(x);
plot(x,y),xlabel('x'), ylabel('zeta(x)'), title('Zeta(x) Graph'),

%% function definition 
function zeta_val = zeta_func(x)    
    %% zeta(x) = summation of 1/x^n
    %% here summing from n = 1 to 50
    n = 1:1:50;
    % create meshgrid, with unique x in each column, unique n in each row (50x501)
    [X,N]=meshgrid(x,n);
    terms = 1./power(N,X);        %% performing 1/x^i for all n   
    zeta_val = sum(terms);        %% summing the values
end

[Jan]

Or alternatively:

terms = 1 ./ power(n.', x);

[madhan ravi]

Ah didn't see your comment Jan :), have to increase my font size ;)

[Mohsin Shaikh]

Thank you so much!

Answer 2

[madhan ravi]

terms = 1 ./ (n( : ) .^ x) % < 2016b 1 ./ bsxfun(@power, n( : ), x) 

[Mohsin Shaikh]

Thanks a lot!

Answer 3

[Luigi Emanuel di Grazia]

Shouldn't it be power(x,n)?