Viv Kendon Durham University viv.kendon@durham.ac.uk HPC & - - PowerPoint PPT Presentation

Hybrid Quantum & Classical Computing

Viv Kendon

JQC & QLM Physics Dept Durham University viv.kendon@durham.ac.uk

HPC & Quantum Summit 2019 (Westminster Hall) Tuesday 5th February 2019 Key collaborators: Susan Stepney

(York Cross-disciplinary Centre for Systems Analysis)

Nick Chancellor (UKRI Innovation Fellow, Durham) Funding: EPSRC Fellowship in Quantum T echnologies

February 3, 2019

Hybrid Quantum Computing

GOAL: increase computing power . . .

⋆ current computers already very powerful – two barriers to more computing power:

• 1. silicon chip technology reaching limits
• 2. energy consumption far from optimal:

– resource limits; global warming

note these are related: can’t cool Si chips any faster

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

2/13

February 3, 2019

Hybrid Quantum Computing

beyond silicon . . .

quantum: IBM 5 qubit BZ reaction chemical reservoir computer rat neuron on silicon encoding for DNA computer

⋆ plenty of examples ⋆

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

3/13

February 3, 2019

Hybrid Quantum Computing

hybrid computers . . .

practice: co-processors: unconventional: control + substrate: conventional:

• graphics cards
• ASIC application-specific integrated circuit
• FPGA field-programmable gate array
• quantum
• NMR
• reservoir
• slime mould

⋆ hybrid computational systems are the norm ⋆ theory: single paradigm:

• classical – T

uring Machine

• analog – Shannon’s GPAC
• quantum – gate model, QTM . . .
• linear optics (Bosons) [Aaronson/Arkhipov STOC 2011 ECCC TRI-10 170]

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

4/13

February 3, 2019

Hybrid Quantum Computing

quantum information processing

Quantum Information is built on the idea that: Quantum Logic allows greater efficiency than Classical Logic

classical quantum bits, 0 or 1 qubits, α |0 + β |1 yes or no, binary decisions yes and no, superpositions HEADS or TAILS, random numbers random measurement outcomes ⇒ quantum gives different computation from classical: how different?

• computability – what can be computed?
• complexity – how efficiently can it be computed?

⇒ quantum computability is the same as classical complexity differs: some problems can be computed more EFFICIENTL Y

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

5/13

February 3, 2019

Hybrid Quantum Computing

measurement-based quantum computing

[Raussendorf/Briegel PRL 86, 518 (2001)]

classical controls integral part of the architecture: control layer base layer

Richard Jozsa [arχiv:quant-ph/0508124v2] – first to highlight the role of classical processing in quantum computing

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

6/13

February 3, 2019

Hybrid Quantum Computing

measurement-based quantum computing

Anders/Browne PRL 102 050502 (2009):

correlated resource control computer

⊕L (parity-L)

track parity of each qubit

P (universal classical – Clifford group) combination gives BQP (universal quantum computing)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

7/13

February 3, 2019

Hybrid Quantum Computing

heterotic Computing

• from the greek: heterosis ≡ hybrid vigour

compose different types of computational devices −→ more powerful hybrid computers characterise in terms of the computational power of the parts – if the whole is more than the sum of the parts: heterotic −→ not specifically quantum: Theo Murphy Meeting at Chicheley Hall 7–8 Nov 2013

• Phil. Trans. Royal Soc. A 2015 373 20150091;

DOI: 10.1098/rsta.2015.0091. Heterotic computing: exploiting hybrid computational devices

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

8/13

February 3, 2019

Hybrid Quantum Computing

reservoir computing

engineering stage use stage

A Substrate-Independent Framework to Characterise Reservoir Computers, Matthew Dale, Julian F . Miller, Susan Stepney, Martin A. Trefzer arχiv:1810.07135

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

9/13

February 3, 2019

Hybrid Quantum Computing

hybrid algorithms . . .

⋆ hardware is first step towards maximising computing power . . . algorithms need to exploit the full capabilities of hardware example hybrid quantum-classical algorithms:

X

E state

algorithms match hardware: quantum annealers with limited precision

Nick Chancellor: New Journal of Physics 19, 2, 023024 (2017) (local searches) Nick Chancellor: arXiv:1609.05875 (genetic algorithms)

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

10/13

February 3, 2019

Hybrid Quantum Computing

quantum computing diversity

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

11/13

February 3, 2019

Hybrid Quantum Computing

continuous-time quantum computing

family of computational models:

• discrete – qubits for

efficient encoding

• continuous time evolution

using system Hamiltonians

• coupling to low temp bath –
• pen system effects

cooling

QA QW AQC

unitary

• pen

em

noise (high

−→ makes sense because qubits do superpositions; classical bits don’t

Quantum search with hybrid adiabatic-quantum walk algorithms and realistic noise Morley/Chancellor/Bose/VK, to appear in PRA, arχiv:1709.00371

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

12/13

February 3, 2019

Hybrid Quantum Computing

future of computing:

huge investment in silicon: continuing to develop:

• squeeze more performance out by making specialised chips
• team up with quantum computer developers

+ quantum co-processors

• cloud services and networked devices – chips with everything

( – is your toaster spying on you??)

networked embodied hybrid smarter multicore co-processors optimized

future −→ diversified

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

13/13

February 3, 2019

Hybrid Quantum Computing

• pen questions:
• heterotic in theory – complexity classes (one example so far!)
• many ways to gain an advantage:

– faster (especially for real time applications) – cheaper (for mass markets) – more robust (for portability or extreme environments) – easier to program – more reliable – . . . which of these are heterotic?

• heterotic in practice – all hybrid is for an advantage of some sort

⋆ when is the combination more than the sum of its parts? ⋆

• + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

14/13