We are a Canadian open source development company based in Toronto. In order to achieve quantum speedup across our distributed graph database, we have been undertaking research into efficient quantum simulation on classical hardware.
In essence, we are building a software kernel that will enable general quantum computation (arbitrary algorithm) on a typical CPU or GPU such as those built by Intel, Nvidia, AMD and ARM. Thus we are coding the ideal quantum computer, albeit at the computational cost of the simulator's complexity.
Although efficient, general quantum computation on Turing machines is considered a hard, open problem; the majority of the QI community feels that such a thing is either "impossible" or, at least "unlikely". Regardless, there are [http://arxiv.org/find/quant-ph/1/ti:+AND+simulator+AND+classical+quantum many recent, limited cases] where efficient classical simulation is proven. It is known that classical computers can do all quantum computations deterministically - therefore ''optimization'' is the sole barrier between where we are today, and efficiency.