GitHub - cbouilla/pcg: Prediction algorithms for the PCG pseudo-random generator

cbouilla / pcg Public

Notifications You must be signed in to change notification settings
Fork 1
Star 8

Prediction algorithms for the PCG pseudo-random generator

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 213 Commits
Cknown		Cknown
Cunknown		Cunknown
Python		Python
article		article
challenges		challenges
logs		logs
slides		slides
slides_short		slides_short
.gitignore		.gitignore
G		G
Greduite.sage		Greduite.sage
README		README
good_big_matrix.py		good_big_matrix.py
good_big_matrix.sage		good_big_matrix.sage
lehmer.py		lehmer.py
minL.py		minL.py
minL.sage		minL.sage
remarques_pas_traitees		remarques_pas_traitees

Repository files navigation

Code for the algorithms that recover seeds for the PCG pseudo-random number 
generator.

Written by Julia Sauvage (julia.sauvage@etu.sorbonne-universite.fr) and
Charles Bouillaguet (charles.bouillaguet@sorbonne-universite.fr).

This code is in the public domain.


challenges/ folder: tools to generate challenge inputs and expected output.


Cknown/ folder: case where the default increment is used. 

main.c: Multi-thread program that recovers the seed from 3 consecutive outputs. 
        Takes about 25 CPU-minutes. A simple procedure is repeated 2**37 times.



Cunknown/ folder: case where an unknown increment is used (default in numpy). 

main.c: MPI program that recovers a partial difference between two states from
        10 consecutive outputs. Takes about 12500 CPU-hours. Uses MPI, so is
        parallelizable on a cluster. A simple procedure is repeated 2**51 times.

step2.py: given 64 consecutive outputs and the result from the previous program,
          finds the full difference between any two consecutive states. Very
          fast.

step3.py: given 64 consecutive outputs and the result from the two previous 
          programs, prints the seeds. Very fast.



Other folders: journal article, slides, random junk.