This is the pseudocode for the up-sweep. (Notice the parfor over the j variable, which means that this block of code can be parallelized over threads indexed by j):

input: x0, ..., xn-1
initialize:
    for i = 0 to n - 1:
        yi := xi
begin:
for k=0 to log2(n) - 1:
    parfor j=0 to n - 1: 
        if j is divisible by 2k+1:
            yj+2k+1-1 = yj+2k-1  yj +2k+1 -1
            else:
            continue
end
output: y0, ..., yn-1