IBM Research announced a 50 qubit quantum computer system in November, 2017. In a presentation in Sept, 2017, IBM indicated that they are doubling qubits every 8 months.
If IBM maintains an 8 month qubit doubling rate then they will announce a
100 qubit quantum computer in June, 2018
a 200 qubit system in Feb, 2019 and
400 qubits in October 2019.
IBM projects DWave could surpass the best classical optimizers in 7 years
Cracking Crypto could be 15 years away at 1.4 yr qubit doubling or 7 years with 8 month doubling
Shor’s algorithm promises polynomial-time factorization
–Extremely valuable, if rather destructive
* Requires general-purpose qubits (IBM-Q, not D-Wave current version)
–Thousands of them
* Assuming 1.4 years per doubling, we have about 15 years until QC cracks 1000-bit RSA
–Also assumes that Shor’s algorithm actually works on real hardware
–On the other hand, IBM-Q may be adding qubits faster than 1.4 years per doubling, doubling every 8 months from May 2016 to May 2017
–So it might not be too early to start work on QC-resistant cyphers
The projected time is delayed years if the error rates are not reduced or the error correction problem remains difficult.
IBM Research introduced the concept of quantum volume. If we want to use quantum computers to solve real problems, the number of qubits is important, but so is the error rate. In practical devices, the effective error rate depends on the accuracy of each operation, but also on how many operations it takes to solve a particular problem as well as how the processor performs the operations.
The quantum volume measures the useful amount of quantum computing done by a device in space and time.
As we build larger quantum computing devices capable of performing more complicated algorithms, it is important to quantify their power. The origin of a quantum computer’s power is already subtle, and a quantum computer’s performance depends on many factors that can make assessing its power challenging. These factors include:
1. The number of physical qubits;
2. The number of gates that can be applied before errors make the device behave essentially classically;
3. The connectivity of the device;
4. The number of operations that can be run in parallel.
The quantum volume, to summarize performance against these factors.