My submission for the final M3C project 2016
This is a study of network traversal problems under the Erdos Renyi model. I used Python for data analysis, and the F2Py module to compute the backend processes in Fortran. I also used OpenMP to parallelise the processes and studied performance differences.
Unfortunately I don't have the questions stored as they were on a temporary website, but hopefully the files make sense.
I have kept all of my working files in 'OtherFilesUsed'
#OtherFilesUsed build.sh - automate f2py wrapping network_dev.f90 - dev version of network.f90 omptest.py - testing omp, python/fortran code part1_dev.f90 - dev version of part1.f90 test.py - more tests part1_dev.py - dev version of part1_dev.py
#Rest network.f90, part1.py, part1.f90, part1.py - as described by question
p11.png - plot of greatest number of walkers at each step p12.png - plot of variation of F(Ntime,X0)
#OtherFilesUsed part2_dev.f90 - dev version of part2.f90 network.f90 - original network.f90 file part2_dev.py - dev version of part2.py
#Rest all files as descibed by question
added these inputs to solveFluNet in part2.py verFlag - tells you which version of RHS to use ('python','fortran','omp') N0,L,Nt - as described by question. Don't think these are needed anymore, I was originally initializing inside the function
#OtherFilesUsed part31_dev.f90 - dev version of part31_dev.f90 part3_dev.py - dev version of part3_dev.py
#Rest part3.py,part31.f90 as described by question
Omar Haque email: omar.haque14@imperial.ac.uk CID: 00944961