An Update on Google's Quantum Computing Initiative

An Update on Google’s Quantum Computing Initiative January 31, 2019 <a href="mailto:[email protected]" target="_blank">[email protected] </a>Copyright 2018 Google LLC Our Brains are Wired for Newtonian Physics Brains that recognize and anticipate behaviors of Heat, Light, Momentum, Gravity, etc. have an Evolutionary Advantage. Quantum phenomena contradict our intuition. Quantum Phenomena Contradict Intuition Interference, “Erasure”, etc. 0(1) Quantum Theory Explains Cleanly… ...but the Math looks Strange 1i1 √2 1 √2 1 i i 1 How can a Particle be On Two Paths at the Same Time? 1 1i√2 Pu 1 √2 1 i i 1 (P) rCopyright 2017-2019 Google LLC Superposed States, Superposed Information l0〉 ( |0〉+|1〉 )2 = |00〉+|01〉+|10〉+|11〉 |0〉 + |1〉 l1〉 Copyright 2017-2019 Google LLC 01 Macroscopic QM Enables New Technology Control of single quantum systems, to quantum computers <ul style="display: flex;"><li style="flex:1">1 nm </li><li style="flex:1">1 μm </li><li style="flex:1">1 mm </li></ul>H atom wavefunctions: Problem: Light is 1000x larger Large “atom” has room for complex control Copyright 2017-2019 Google LLC Xmon: Direct coupling + Tunable Transmons ● Direct qubit-qubit capacitive coupling ● Turn interaction on and off with frequency gcontrol <ul style="display: flex;"><li style="flex:1">“OFF” </li><li style="flex:1">“ON” </li></ul>f10 f21 ΔηCoupling g SQUID <ul style="display: flex;"><li style="flex:1">Qubit </li><li style="flex:1">Qubit </li></ul>Coupling rate Ωzz ≈ 4ηg2 / Δ2 <ul style="display: flex;"><li style="flex:1">Qubit A </li><li style="flex:1">Qubit B </li></ul>Copyright 2017-2019 Google LLC Logic Built from Universal Gates <ul style="display: flex;"><li style="flex:1">Classical circuit: </li><li style="flex:1">Quantum circuit: </li></ul>1 qubit rotation 2 qubit CNOT No copy 1 bit NOT 2 bit AND Wiring fan-out 1 Input Gates 2 Input Gates (space+time ) time Copyright 2017-2019 Google LLC Execution of a Quantum Simulation arXiv:1512.06860 Copyright 2017-2019 Google LLC Quantum Simulation Results, H2 Molecule Control Qubits arXiv:1512.06860 Copyright 2017-2019 Google LLC Office of the CTO Space-Time Volume of a Quantum Gate Computation 2 qubit gate fidelity = 99.5% Uncorrected Gate “Circuits” Limited by Fidelity of Operations and Decoherence Times Fidelity is the Third Dimension Copyright 2017-2019 Google LLC Gate Depth Goals for Near-Term Scaling 10-2 Error correction threshold ∗“Quantum Supremacy” 10-3 10-4 Useful error corrected QC Near-term applications Classically simulatable ✔✘?<ul style="display: flex;"><li style="flex:1">100 </li><li style="flex:1">101 </li><li style="flex:1">102 </li><li style="flex:1">103 </li><li style="flex:1">104 </li><li style="flex:1">105 </li><li style="flex:1">106 </li><li style="flex:1">107 </li><li style="flex:1">108 </li></ul>Number of Qubits Copyright 2017-2019 Google LLC Which computer is better at landing samples the “Bright Spots” ? Quantum “Supremacy” Do what classical CPUs Cannot do: 2n index k 0Ideal distribution Multiple errors >50 Qubits, >40 Steps 2n ordered index k 0Copyright 2017-2019 Google LLC Toward Universal Fault-Tolerant QC ● Qubit error rates ~10-2-10-3 per operation ● Universal QC requires ~10-10 ● Error correction: ○ Low error logical qubit made with many physical qubits ● Surface code error correction: ○ 2D array of qubits (n.n. coupling) ○ Modest error rates (1% threshold, 0.1% target) ○ Useful at 106 physical qubits Copyright 2017-2019 Google LLC 9 Qubit: Good performance, Limited Scaling 9 qubit device has good performance ● ErrCZ down to 0.6% ● ErrSQ < 0.1% ● ErrRO = 1% Limited to 1D connectivity (planar geometry) Scale-up strategy: move qubits, control to different planes Copyright 2017-2019 Google LLC Bump-Bond Architecture ● Bond together two separate chips ○ Qubits → “Chip” “Chip” bumps ○ Control → “Carrier” “Carrier” ● Superconducting interconnect ● Use lossless vacuum as dielectric In TiN Al 30 μm Copyright 2017-2019 Google LLC Qubits (chip) Control (substrate) Scaling to 2D Design must be “tileable” (control fits in qubit footprint) ● Readout resonator ● XY coupler ● SQUID coupler Need to shield qubits from interior wire routing ● Small coupling to 50Ω line will decohere qubit Copyright 2017-2019 Google LLC “Foxtail” 22 Qubit Device “Chip” ● Qubits “Carrier” Readout XY control Z control ●●●Copyright 2017-2019 Google LLC 2D Unit Cell ● Diagonal for surface code: all “measure” qubits on same line ● Condense footprint across 2 chips ● Introduce shielded wiring between qubits ● Tile unit cell for 2D array Unit cell: Condensed, diagonal linear chain Readout line RO XYZ RO XYZ RO XYZ <ul style="display: flex;"><li style="flex:1">RO </li><li style="flex:1">RO </li></ul>XYZ RO XYZ rotate XYZ Unit cell designed for surface code Copyright 2017-2019 Google LLC “Bristlecone” Architecture Tile 12 unit cells of 6 qubits = 72 qubits Tile for a 2D grid of n.n. coupled qubits Bonus: Looks like a pine cone! Copyright 2017-2019 Google LLC Photo: Erik Lucero “72 qubits cold in fridge” Copyright 2017-2019 Google LLC Early Quantum Computing Applications Quantum Numerical Simulation Optimization Copyright 2017-2019 Google LLC C’est quoi ce Cirq? Copyright 2017-2019 Google LLC Cloud Quantum Computing Workflow Application Frameworks OpenFermion (chemistry) IsingFlow (optimization) TensorFlow (ML) Quantum Hardware results quantum hardware language qhl qhl Quantum Engine results Python framework for writing quantum programs Quantum Simulator results Programs Jobs Results Copyright 2017-2019 Google LLC Is this a good quantum circuit? CZ a b CZ b c CZ c d CZ d e CZ e f CZ f g CZ g h CZ h i Copyright 2017-2019 Google LLC Is this a good quantum circuit? abcdefghiCZ a b CZ b c CZ c d CZ d e CZ e f CZ f g CZ g h CZ h i No: large depth Copyright 2017-2019 Google LLC Is this a good quantum circuit? Attempt #2 CZ a b CZ c d CZ e f CZ g h CZ b c CZ d e CZ f g CZ h i Copyright 2017-2019 Google LLC Is this a good quantum circuit? Attempt #2 abcdefghiCZ a b CZ c d CZ e f CZ g h CZ b c CZ d e CZ f g CZ h i Copyright 2017-2019 Google LLC Is this a good quantum circuit? Attempt #2 abcdefghiCZ a b CZ c d CZ e f CZ g h adgbehcfCZ b c CZ d e CZ f g CZ h i iNo: Does Not Map to Physical Topology Copyright 2017-2019 Google LLC Is this a good quantum circuit? Attempt #3 CZ a b CZ c f CZ e d CZ g h CZ b c CZ f e CZ d g CZ h i Copyright 2017-2019 Google LLC

An Update on Google's Quantum Computing Initiative

Details

Download

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

Support