Home

Welcome to the sqaod wiki!

Beta1 has been released (May 4th, 2018).

The current contents are for Alaph2, and are going to be updated to Beta1 soon.

• important fixes and updates:
  • Fixed: In alpha1, CPU/CUDA annealers with FP32 did not work due to incorrect numerical type conversion in python c-extension. In alpha2, this issue has been fixed and tested.
  • API change : set_x() is used to set initial bit vectors to annealers in alpha1. In alpha2, set_q() is used intead. The set_q() accepts one spin vector or an array of spin vectors.
  • Several fixes in python code and c-extensions.
  • Current versions are 1.0.1 for debs and 1.0.1a2 for whl in pypi.

Sqaod is a collection of solvers for simulated quantum annealing.

Algorithm

Currently 2 algorithms have been implemented.

• Brute-force searcher
  Carrying out brute-force search for every combination of bits to get a strict solution for a given QUBO.
• Simulated quantum annealing.
  Carrying out simulated quantum annealing to get a solution of a given QUBO.

Graph structure

Solvers are provided for 2 graph types. Graph-specific optimizations are applied to each solver.

• Dense graph
  
• Bipartite graph
  

Python

All solvers are callable from python 2.7 and python 3.3 or later.

Implementation

Implementations are available for language/devices shown in the table below:

language	python package	description
python	sqaod.py	Provided to show agorithm.
CPU	sqaod.cpu	CPU-based solvers, parallelized and accelerated by OpenMP.
CUDA	sqaod.cuda	GPU solvers, parallelized and accelerated by CUDA.

Examples

Please visit example directory. The 4 examples are provided.

File name	graph	algorithm
dense_graph_annealer.py	dense graph	simulated quantum annealing
bipartite_graph_annealer.py	bipartite graph	simulated quantum annealing
dense_graph_bf_searcher.py	dense graph	brute-force searcher
bipartite_graph_bf_searcher.py	bipartite graph	brute-force searcher