Slashdot Mirror
Physicists Made An Unprecedented 53 Qubit Quantum Simulator (vice.com)
Two teams of researchers have published papers [1, 2] in the journal Nature detailing how they were able to create unprecedented quantum simulators consisting of over 50 qubits. The University of Maryland team and National Institute of Standards and Technology team -- the two teams behind one of the two new papers -- were able to create a quantum simulator with 53 qubits. Motherboard reports: Quantum simulators are a special type of quantum computer that uses qubits to simulate complex interactions between particles. Qubits are the informational medium of quantum computers, analogous to a bit in an ordinary computer. Yet rather than existing as a 1 or 0, as is the case in a conventional bit, a qubit can exist in some superposition of both of these states at the same time. For the Maryland experiment, each of the qubits was a laser cooled ytterbium ion. Each ion had the same electrical charge, so they repelled one another when placed in close proximity. The system created by Monroe and his colleagues used an electric field to force the repelled ions into neat rows. At this point, lasers are used to manipulate all the ytterbium qubits into the same initial state. Then another set of lasers is used to manipulate the qubits so that they act like atomic magnets, where each ion has a north and south pole. The qubits either orient themselves with their neighboring ions to form a ferromagnet, where their magnetic fields are aligned, or at random. By changing the strength of the laser beams that are manipulating the qubits, the researchers are able to program them to a desired state (in terms of magnetic alignment).
According to Zhexuan Gong, a physicist at the University of Maryland, the 53 qubits can be used to simulate over a quadrillion different magnetic configurations of the qubits, a number that doubles with each additional qubit added to the array. As these types of quantum simulators keep adding more qubits into the mix, they will be able to simulate ever more complex atomic interactions that are far beyond the capabilities of conventional supercomputers and usher in a new era of physics research. Another team from Harvard and Maryland also released a paper today in which it demonstrated a quantum simulator using 51 qubits.
According to Zhexuan Gong, a physicist at the University of Maryland, the 53 qubits can be used to simulate over a quadrillion different magnetic configurations of the qubits, a number that doubles with each additional qubit added to the array. As these types of quantum simulators keep adding more qubits into the mix, they will be able to simulate ever more complex atomic interactions that are far beyond the capabilities of conventional supercomputers and usher in a new era of physics research. Another team from Harvard and Maryland also released a paper today in which it demonstrated a quantum simulator using 51 qubits.
Scott Aaronson, a prominent quantum computing expert made comments about some very similar work that is relevant https://www.scottaaronson.com/blog/?p=3512. The short summary is that we should expect people to continue to push up how many qubits can be practically simulatable. But that sort of improvement through clever tricks and the like doesn't really do much to address the more interesting issue of quantum supremacy https://en.wikipedia.org/wiki/Quantum_supremacy, whether there are problems that a quantum computer can solve that a classical computer practically cannot. Note that the "practically" in the previous sentence is really important. Everything a quantum computer can do a classical computer can do with exponential slow down; standard conjectures essentially amount to saying that a classical computer cannot do any better than that.
The sources that claim 4k or 10k for a 2048k key are misinformed themselves or don't know how quantum computing (theory) works.
You're more than welcome to point out the specific thing they get wrong, as there are plenty of detailed accounts of how many qubits are needed for Shor's algorithm, e.g. here's a paper discussing the trade-off of space vs. speed scaling. Much like classical digital circuitry, you need some scratch space to store intermediate values, and more space lets you use fewer operations for quicker results, or less space requires more operations.
Even the naive implementation of Shor's algorithm for n bit number requires 2n qubits, and takes special efforts to make that work with a minimum n+1 qubits by reusing a qubit n times (greatly increasing the time the whole system needs to hold together).
How is it circular? This is no different than saying, "A hydraulic computer is not modeling water flow without water, but is an analog computer using water to model other things."
GGP said this uses no actual quantum states. GP says that is flat out wrong: This computer uses qubits, but is not a general purpose quantum computer. It is the same difference as between a general purpose cpu and a fixed logic circuit that does one thing, but still uses bits.
If that seems circular, then it is because your reading comprehension has been impaired by your preoccupied with where people stick their genitalia.