A general implementation of Shor's algorithm in IBMQ

Following the order of https://arxiv.org/abs/quant-ph/0205095, I construct the Shor's circuit from the basic gates.

Getting Started

For the access of the real quantum devices, you are recommended to register an IBMQ account.

Prerequisites

You should install qiskit (Anaconda environment preferred) to run my code on your machine or using https://quantum-computing.ibm.com/ instead.

conda create --name yourenvname python=3
source activate yourenvname
pip install qiskit
pip install matplotlib #required for plotting
#for CUDA support simulator
pip install qiskit-aer-gpu
(CUDA required)
(Strongly recommended following the tutorial on tensorflow official site)

Structure

• shor.py is the main program.
• gateSet.py implements the gates needed in Shor's algorithm.
• myAlgo.py implements the test of gates and algorithm.
• simulation.py implements the simulation function and argparse in the main program.
• CF.py implements the continuous fraction algorithm and factorize the given N and a.
• factorize.py implements the factorizing main program.

In this repo, I would separate the Shor's algorithm into two parts. First is the circuit experient which would generate results in /normal/result/ or /sequential/result/ depending on the circuit mode. Then run the factorizing part for the factorization.

Run

In [shor.py」」(https://github.com/alfa871212/shor_paper/blob/master/shor.py), the execution flag is listed below

python shor.py [-h] [--simulation local/ibmq | --real]
               (--adder a b n | --phimod n b a N | --cmult n x b a N | --cu n x a N | --nor N a | --seq N a)
               [--output] [--draw] [--log]

Flag explanation: Backend selection:

[-s|-r] select the backend to simulate the circuit, default is local simulator

Test selection:

--adder adder test 
--phimod modular adder test 
--cmult multiplier test 
--cu CU test
--nor Normal Shor's algorithm
--seq Sequential Shor's algorithm

Result selection:

--output plot the result histogram
--draw export a png file of circuit
--log redirect the stdout to file

For example, adder test with a=2 b=3 n=3, we want to do the a+b in n-bit number.

python shor.py -s local --adder 2 2 3 

In factorize.py,

factorize.py [-h] [--type seq/nor] [--individual res len N a | --file N a] [--log]

Result selection:

--indiv just give the result and N,a 
--file visit result csv file

Type selection (--file required):

--type choose the type of circuit, it should directly visit corresponding directory.

Logfile selection:

--log redirect log to file

For example, I want to individually factorize the result 10100000 for 8 bit number with N=15 a=4.

python factorize.py --indiv 10100000 8 15 4

Or, I want to factorize a given csv file sequential/result/15_4.csv.

python factorize.py --file 15 4 --type seq

Some changes in the Qiskit source code

Because the build-in plotting function in Qiskit didn't satisfy my demand, I change the some line in Qiskit source code. If you want the same effect of plot_histogram, you could compare the my_counts_visualization.py with yours (w/o my_) in the qiskit package.

in location of anaconda installation
qiskit/visualization/counts_visualization.py