Compile

mpicc test_module.c HyperCubeLib.c  -o run

Run

mpirun -n 64 run

More

Why I did this?

• Inspired by CS525 at Purdue, here is the related slide.

Complexity

• see it at the top of each function here

Different implementations

• see the MACROs here, and you could compare the complexities among different implementations.

(New feature) Weighted Scan

• definition: $S[i] = \sum_{k=0}^i sum[k] (\Pi_{l=k+1}^i mul[l])$
• sequential handling:

for(int i = 1; i < n; i++)
    S[i] = sum[i] + mul[i] * S[i - 1];

• Which is built for Mamba, that Scan