gpu computing and parallel computing (vectorize code)

Hi, again I am coming back as having simple example.

At first, I want to speed up using parallel tool box like CPU multicore and GPU computing(Quadro M6000)

But I can't improve my code to speed up using them

I think the reasion is while loop (iteration) .

Please give me genious idea

N Particles simultaneously are computed in the while loop

If you have nice idea about speeding up using GPU (Euler method computing through while loop),

overall, you can modify this code; remove while loop or use good function to solve Euler with gpu.

tic;
gd = gpuDevice();
S = gpuArray(linspace(0, 100, 1E3));
% S = linspace(0, 100, 1E3);
b = arrayfun(@(x) plus_(x),S);

toc;

function b = plus_(a)
dt = 1E-1;
j = 0;

while a < 1E4
    j = j + 1;
    a = a + j * dt;
end

b = a;
end

Answer 1

[Matt J]

b=max(0, floor( (1e4-S)/dt)   )*dt + S;