d-wave next generation plans and activities
Post on 19-Nov-2021
1 Views
Preview:
TRANSCRIPT
D-Wave Next Generation Plans and Activities
SOS23 / Multi-ExaOp Platforms (beyond 2021)
Asheville, NC
Dr. Joel M. Gottlieb
March 29, 2019
Copyright © D-Wave Systems Inc. 2
Overview
• D-Wave Introduction and Background
• DOE Activity
• Future Systems
• Conclusion
Copyright © D-Wave Systems Inc. 3
Quantum Annealing Outlined by TokyoTech
PHYSICAL REVIEW E VOLUME 58, NUMBER 5 NOVEMBER 1998
Quantum annealing in the transverse Ising modelTadashi Kadowaki and Hidetoshi Nishimori
Department of Physics, Tokyo Institute of Technology, Oh-okayama, Meguro-ku, Tokyo 152-8551, Japan
(Received 30 April 1998)
We introduce quantum fluctuations into the simulated annealing process of optimization problems, aiming at faster convergence to the optimal state. Quantum fluctuations cause transitions between states and thus play the same role as thermal fluctuations in the conventional approach. The idea is tested by the transverse Ising model, in which the transverse field is a function of time similar to the temperature in the conventional method. The goal is to find the ground state of the diagonal part of the Hamiltonian with high accuracy as quickly as possible. We have solved the time-dependent Schrödinger equation numerically for small size systems with various exchange interactions. Comparison with the results of the corresponding classical (thermal) method reveals that the quantum annealing leads to the ground state with much larger probability in almost all cases if we use the same annealing schedule.[S1063-651X~98!02910-9]
1960 1970 1980 1990 2000 2010 2020
Copyright © D-Wave Systems Inc. 3
Copyright © D-Wave Systems Inc. 4
D-Wave’s Goal
• Create adiabatic quantum optimization (AQO) quantum processing unit (QPU) as programmable chip
• Build programmable chip as a grid of flux qubits
• Chip must be fabricated as integrated circuit, and robust against fabrication variations
• Must work well at millikelvin temperatures, without significant interference and thermal effects
• Must be able to represent and solve Ising model problems using voltages and magnetic fields to control chip circuitry
• Must be able to read flux qubits at end of adiabatic evolution when quantum mechanical tunneling is done
• Each problem must be independent of the one before – fully reprogrammable
Copyright © D-Wave Systems Inc. 5
D-Wave Company Background
• Founded in 1999
• World’s first quantum computing company
• Public customers:
– Lockheed Martin/USC
– Google/NASA Ames/USRA
– Los Alamos National Laboratory
– Cybersecurity - 1
– Oak Ridge National Laboratory
• Other customer projects done via cloudaccess
• ~150 U.S. patentsCopyright © D-Wave Systems Inc. 5
Copyright © D-Wave Systems Inc. 6
Landscape metaphor
• Space of solutions defines an energy landscape & best solution is lowest valley
• Classical algorithms can only walk over this landscape
• Quantum annealing uses quantum effects to go through the hills
Copyright © D-Wave Systems Inc. 8
Processor Environment
• Cooled to 0.015 Kelvin, 175x colder than interstellar space
• Shielded to 50,000× less than Earth’s magnetic field
• In a high vacuum: pressure is 10 billion times lower than atmospheric pressure
• On low vibration floor
• <25 kW total power consumption – for the next few generations
15mK15mk
Copyright © D-Wave Systems Inc. 9
• Quantum Processor mounted on a “cold finger”
• Connected to the real-world via specialized wiring and RF read-out chain
• Operates in an extreme environment – to get there we use a special kind of “Fridge” (dilution refrigerator)
• Lives in a uniquely isolated environment – electrical, magnetic, and thermal
Shielded Enclosure
Quantum Processing Unit
© 2016 D-Wave Systems Inc. Proprietary and Confidential Information | 10
Copyright © D-Wave Systems Inc. 11
The system samples from the 𝑞𝑖 that minimize the objective
QUBIT 𝒒𝒊Quantum bit which participates in annealing cycle and settles into one of two possible final states: 0,1
COUPLER 𝒒𝒊𝒒jPhysical device that allows one qubit to influence another qubit
WEIGHT 𝒂𝒊
Real-valued constant associated with each qubit, which influences the qubit’s tendency to collapse into its two possible final states; controlled by the programmer
STRENGTH 𝒃𝒊𝒋
Real-valued constant associated with each coupler, which controls the influence exerted by one qubit on another; controlled by the programmer
OBJECTIVE 𝑂𝑏𝑗Real-valued function which is minimized during the annealing cycle
𝑶𝒃𝒋(𝒂𝒊, 𝒃𝒊𝒋; 𝒒𝒊) =
𝒊
𝒂𝒊 𝒒𝒊 +
𝒊𝒋
𝒃𝒊𝒋𝒒𝒊 𝒒𝒋
Copyright © D-Wave Systems Inc. 12
Leap Quantum Application Environment (QAE)
https://cloud.dwavesys.com/leap
Copyright © D-Wave Systems Inc. 13
How do I program with QUBOs?
• Package of Python tools for Ising/QUBO problems.
• Majority is open-source code available on GitHub.
• Extensions and features from community welcome!
Copyright © D-Wave Systems Inc. 15
Gate Model Machines - 2019
• ~50 – 100 qubit models running
• No large scale error correction
• Noisy Intermediate Scale QC’s (NISQ)*
• Know if some problems will run without
error correction
• Quantum Material Science?
• Quantum “Supremacy” perhaps for
synthetic benchmark
• Importance of error correction and potential apps becomes clear* “Quantum Computing in the NISQ era and beyond”, John Preskill, Cal Tech, arXiv:1801.00862
Copyright © D-Wave Systems Inc. 15
Copyright © D-Wave Systems Inc. 16
Overview
• D-Wave Introduction and Background
• DOE Activity
• Future Systems
• Conclusion
Customer Application Areas
Copyright © D-Wave Systems Inc.
17
• Lockheed/USC ISI
– Software Verification and Validation
– Optimization – Aeronautics
– Performance Characterization & Physics
• Google/NASA Ames/USRA
– Machine Learning
– Optimization
– Performance Characterization & Physics
– Research
• Los Alamos National Laboratory
– Optimization
– Machine Learning, Sampling
– Software Stack
– Simulating Quantum Systems
– Other (good) Ideas
• CS - 1
– Cybersecurity
• Oak Ridge National Laboratory
– Similar to Los Alamos
– Material Science & Chemistry
Copyright © D-Wave Systems Inc. 18Los Alamos National Laboratory
D-Wave “Rapid Response” Projects (Stephan Eidenbenz,
ISTI)
Round 1 (June 2016)
1. Accelerating Deep Learning with
QuantumAnnealing
2. Constrained Shortest Path Estimation
3. D-Wave Quantum Computer as an
Efficient Classical Sampler
4. Efficient Combinatorial Optimization
using Quantum Computing
5. Functional Topological Particle Padding
6. gms2q—Translation of B-QCQP to
D-Wave
7. Graph Partitioning using the D-Wave for
Electronic Structure Problems
8. Ising Simulations on the D-Wave QPU
9. Inferring Sparse Representations for
Object Classification using the
Quantum D-Wave 2X machine
10. Quantum Uncertainty Quantification for
Physical Models using ToQ.jl
11. Phylogenetics calculations
Round 2 (December 2016)1. Preprocessing Methods for Scalable QuantumAnnealing
2. QAApproaches to Graph Partitioning for Electronic
Structure Problems
3. Combinatorial Blind Source Separation Using “Ising”
4. Rigorous Comparison of “Ising” to Established B-QP
Solution Methods
Round 3 (January 2017)1. The Cost of Embedding
2. Beyond Pairwise Ising Models in D-Wave: Searching for
Hidden Multi-Body Interactions
3. Leveraging “Ising” for Random NumberGeneration
4. Quantum Interaction of Few Particle Systems Mediated
by Photons
5. Simulations of Non-local-Spin Interaction inAtomic
Magnetometers on “Ising”
6. Connecting “Ising” to Bayesian Inference ImageAnalysis
7. Characterizing Structural Uncertainty in Models of
Complex Systems
8. Using “Ising” to Explore the Formation of GlobalTerrorist
Networks
Copyright © D-Wave Systems Inc. 19Los Alamos National Laboratory 6/27/2017
Use Case 2016 2017 Total %Combinatorial Optimization 5 5 10 45%
Machine Learning, Sampling 2 2 4 18%
Understanding Device Physics 2 1 3 14%
Software Stack/Embeddings 1 1 2 9%
Simulating Quantum Systems 2 2 9%
Other (good) Ideas 1 1 5%
Total 11 11 22 100%
The LANL Rapid Response Project results for 2016 and 2017 are available as PDF’s at: http://www.lanl.gov/projects/national-security-education-center/information-science-technology/dwave/index.php
Copyright © D-Wave Systems Inc. 20
LANL
• 1000Q installed early 2017
• Upgraded to 2000Q in 2019
• Option for Blake/Pegasus upgrade
Copyright © D-Wave Systems Inc. 21
ORNL begins a second-year with the D-Wave
• D-Wave manages 20% of a remote system for ORNL
• QIS is now a part of the OLCF
• A growing community of users with access to the DW 2000Q processor
• Multiple active projects across optimization, machine learning, physics, and chemistry
• New starts in high-energy physics, basic energy sciences and applied mathematics
• QCI News
21
Copyright © D-Wave Systems Inc. 23
D-Wave Programming Class, February, 2019
https://dwavefeb2019.eventzilla.net
Copyright © D-Wave Systems Inc. 24
Examples of Machine Learning Apps
Election Modeling
Unsupervised Machine Learning
Reinforcement Modeling
Generative Machine Learning
Copyright © D-Wave Systems Inc. 26
Examples of Material Science and Simulation
Molecular DynamicsSimulation
Electronic StructureSimulation
Quantum Materials Simulation
Copyright © D-Wave Systems Inc. 27
Overview
• D-Wave Introduction and Background
• DOE Activity
• Future Systems
• Conclusion
Copyright © D-Wave Systems Inc. 28
Next Generation Machine: Blake / Pegasus
• Named after Bill Blake (VP of R & D, d. 2015)
• Pegasus architecture (successor to Chimera)
• Will be initially available for cloud activity, later for purchase
• Expected mid-2020 for customer installation
• Pegasus vs. Chimera
• Perhaps get to quantum advantage/supremacy
• Get started now – tools available
Copyright © D-Wave Systems Inc. 29
Pegasus
• Much improved connectivity (maximum 15 near neighbors vs. 6 in Chimera)
• It has been shown that a TSP graph needs fewer qubits (80 -> 50?)
• Lower noise
• More than 5000 qubits
• Embedding tools already available in Ocean software and Leap environment
Copyright © D-Wave Systems Inc. 33
2048 Qubits
2000Q Pegasus
Qubits 2048 5000 (+)
Couplers 6000 40,000(+)
Connectivity 6 Max 15
Noise
Temp 10-15mK <10mK
Power 20-25 kw 20-25 kw
Copyright © D-Wave Systems Inc. 34
Overview
• D-Wave Introduction and Background
• DOE Activity
• Future Systems
• Conclusion
Copyright © D-Wave Systems Inc. 35
Conclusions
• Quantum computers are new computers taking advantage of the physics theory quantum mechanics
• D-Wave has offered commercial quantum computers for several years, and IBM/Google/Intel and others are beginning to do so
• Incredibly competitive arena – stay tuned
• Watch for Pegasus in 2020
• Perhaps I have motivated you to get involved!
• Take advantage of our Leap offer – 1 minute/month http://cloud.dwavesys.com/leap
• jgottlieb@dwavesys.com
top related