Advanced Simulations of Quantum Computations in conjunction with the IEEE International Conference on Quantum Computing and Engineering (QCE22)

Sep 19, 2022, 10:00 AM → Sep 20, 2022, 4:45 PM US/Mountain

The Advanced Simulations of Quantum Computations is a 2-day workshop on September 19-20, 2022 from 10:00 am – 4:45 pm MDT. 

Abstract: As quantum computing hardware is steadily evolving towards the quantum advantage regime, classical simulation of quantum circuits is becoming more and more challenging, yet crucial for the verification and validation of the new hardware and algorithms. In recent years, we observed fast progress in new advanced techniques enabling more efficient simulations of rather large quantum circuits. Importantly, these techniques and algorithms were also able to take better advantage of modern classical high-performance computing platforms based on the heterogeneous accelerated node architectures. This workshop will bring together participants from the national labs, academia, industry and open-source software community to share recent results in algorithms and software for large-scale quantum circuit and analog simulations across a broad range of methods, covering state-vector, tensor network, graphical model, stabilizer, and pulse-level simulations. We seek to provide an open platform for state-of-the-art development efforts, exchanging ideas and best practices, and fostering research collaboration in an attempt to stimulate the formation of an inclusive research community focused around this topic.

 

Target Audience: The workshop will be of interest to a broad research community from national labs, academia, and industry who deal with all aspects of quantum simulation and quantum algorithm development. The speakers will represent all mentioned research sectors in a balanced way. We expect workshop attendees with diverse backgrounds ranging from general computer science especially, scientific and high-performance computing, to quantum information science.

IEEE Intl. Conference on Quantum Computing & Engineering - QCE22

Monday, 19 September

10:00 AM → 10:30 AM NWQSim: Scalable Simulation of Quantum Systems on Heterogeneous HPC Clusters

Despite fascinating developments in NISQ based quantum computing recently, simulations of quantum programs in classical HPC systems are still essential in validating quantum algorithms, understanding the noise effect, and designing robust quantum algorithms. To allow efficient large-scale noise-enabled simulation on state-of-the-art heterogeneous supercomputers, we developed NWQSim, a quantum circuit simulation environment that provides support for frontends such as Q#, Qiskit, OpenQASM, etc., and backends such as CPUs, NVIDIA/AMD GPUs, and Xeon-Phis, through state-vector and density matrix. In this talk, I will describe the recent development of NWQSim on OLCF Summit and NERSC Perlmutter supercomputers, particularly regarding to chemistry applications. NWQSim offers an HPC tool for large-scale, fast, and noisy-aware quantum system simulation. NWQSim is supported by the U.S. DOE Quantum Science Center (QSC).

Speaker: Ang Li (Pacific Northwest National Laboratory, USA)

10:30 AM → 11:00 AM Parallel Tensor Network Simulator QTensor

We present a parallel quantum circuit simulator QTensor for running on large supercomputers with the eventual goal to run at scale on exa-scale supercomputers Aurora and Frontier. The simulator is based on the tensor network representation of quantum circuits and designed to run efficiently on both CPUs and GPUs. We implemented NumPy, PyTorch, and CuPy backends and benchmarked the codes to find the optimal allocation of tensor simulations to either a CPU or a GPU. We also present a dynamic mixed backend to achieve optimal performance. To demonstrate the performance, we simulate QAOA circuits for computing the MaxCut energy expectation. Our method achieves 176 times speedup on a GPU over the NumPy baseline on a CPU for the benchmarked QAOA circuits to solve MaxCut problem on a 3-regular graph of size 30 with depth p= 4.

Speaker: Danylo Lykov (Argonne National Laboratory, USA)

11:00 AM → 11:30 AM Data-structures for Quantum Circuit Simulation: Pros and Cons of Decision Diagrams, Tensor Networks, etc.

Due to the underlying exponential complexity of quantum circuit simulation, efficient data-structures such as decision diagrams, tensor networks, etc. are key. Each of those emerged independently with differing perspectives, terminologies, and backgrounds in mind. In this talk, we provide an overview of the different data-structures for quantum circuit simulation and discuss their respective pros and cons. With this, we hope to provide an intuition on the complementary approaches as well as some guidance how to use their full potential.

Speaker: Robert Wille (Technical University Munich, Germany)

11:30 AM → 1:00 PM Break

11:30 am - 1:00 pm

1:00 PM → 1:30 PM Adiabatic State Preparation with Custom Quantum Gates

Quantum algorithms are most commonly implemented in terms of a universal set of quantum
gates. The decomposition of the unitaries defining these algorithms into those in the universal set often yields quantum circuits with depths that prohibit their execution on noisy intermediate-scale quantum hardware. We simulate the execution of an adiabatic-state-preparation algorithm with reduced depth achieved through the usage of custom gates realizing the small-time propagators at each Trotter step. In these simulations the target state was reached with fidelities above 95%, showing a considerable improvement over simulations implementing the same algorithm in terms of the quantum gates in the universal set used by IBM Quantum machines, as well as results from actual runs on IBM Quantum hardware.

Speaker: Eduardo Antonio Coello Perez (Oak Ridge National Laboratory, USA)

1:30 PM → 2:00 PM Optimal Control of Open Quantum Systems on HPC Platforms

Optimal quantum control can be used to shape the control pulses for realizing unitary and non-unitary transformations of quantum states. These control pulses provide the fundamental interface between the quantum compiler and the quantum hardware. Most current software for quantum optimal control (e.g. Qutip or Krotov) is restricted to run on shared memory platforms, limiting their applicability to smaller quantum systems, particularly if interactions with the environment are taken into account. In this talk, we will present an overview of the open-source code Quandary which is specifically designed to solve optimal control problems for large open quantum systems that require deployment on High-Performance Computing platforms. Quandary models the dynamics of quantum systems interacting with the environment using Lindblad's master equation, and optimizes for control pulses that drive the open system to a desired target state. Theoretical and numerical results for various application scenarios will be discussed, such as realizing a state-to-state transfer, unconditional quantum state preparation and quantum reset, as well as logical quantum gate transformations.

Implemented in C++, Quandary runs on distributed memory computers, enabling scalability to large numbers of compute cores using the message passing paradigm. We demonstrate parallel scalability on LLNL's supercomputing platforms, bringing open quantum control for large systems into the realm of the possible.

Speaker: Stefanie Guenther (Lawrence Livermore National Laboratory, USA)

2:00 PM → 2:30 PM Modeling and Simulations of Superconducting Circuits with Scqubits

Superconducting circuits have been gaining ground as one of the leading platforms in the development of hardware for quantum computing applications, making their accurate modeling an important aspect of ongoing research. In this talk, I will discuss scqubits: an open-source Python package for simulating and analyzing superconducting circuits. I will outline its core functionality, features, as well as limitations. I will also briefly present recently added facilities for automatic circuit quantization and talk about our ongoing efforts related to performance enhancements.

Speaker: Peter Groszkowski (Oak Ridge National Laboratory, USA)

2:30 PM → 3:15 PM Break

3:15 PM → 3:45 PM Accelerating the TN-QVM Quantum Circuit Simulator with the ExaTN + cuTensorNet Library Bundle

GPU-accelerated high-performance computing (HPC) provides an efficient way to increase the scale of numerical simulations of quantum circuits and quantum devices. In particular, this applies to quantum circuit simulation techniques based on tensor network formalism. In practice, there are two main challenges when building scalable HPC quantum circuit/device simulators based on tensor networks. First is scalable composability, that is, a simulator design based on expressive scalable data structures and algorithmic primitives capable of parallelization. Second is performance, in particular efficient execution on GPU accelerators. To address both challenges, we have integrated the tensor network processing library ExaTN with the cuTensorNet library from the cuQuantum package, where the former library provides expressivity, composability, and parallelism support that are necessary for implementing HPC quantum simulation methods based on tensor network formalism, while the latter library delivers high performance in contracting tensor networks on NVIDIA GPUs. We describe the capabilities of the ExaTN + cuTensorNet bundle and show its utility in implementing advanced quantum circuit simulation approaches.

Speaker: Dmitry Lyakh (NVIDIA Corporation, USA)

3:45 PM → 4:15 PM Fast Classical Simulation of Quantum Process and Computing

Software for fast evaluation of quantum protocols and programs is vital in exploring quantum advantages. In this talk, we will explain the technology and features of our quantum circuit simulator, Qulacs. We also present our recent work on evaluating high-level properties of quantum computing and discuss an expected use of HPCs for optimizing quantum computer architecture.

Speaker: Yasunari Suzuki (NTT, Japan)

4:15 PM → 4:45 PM Solving the Sampling Problem of the Sycamore Quantum Supremacy Circuits

In this talk, I will introduce two tensor-network methods targeting the sampling problem of the Sycamore circuits. The first method computes amplitudes and probabilities for a large number of correlated bitstrings. The obtained results verify the Porter-Thomas distribution of the large and deep quantum circuits of Google and can be used for spoofing the linear cross entropy benchmark (XEB) of Google. The second method can generate one million uncorrelated bit-strings from the sparse state with fidelity roughly 0.0037, for the Sycamore circuit with 53 qubits and 20 cycles. The whole computation has cost about 15 hours on a computational cluster with 512 GPUs. If our algorithm could be implemented with high efficiency on a modern supercomputer with ExaFLOPS performance, we estimate that ideally, the simulation would cost a few dozens of seconds, which is faster than Google's quantum hardware.

Speaker: Pan Zhang (Chinese Academy of Sciences, China)

Tuesday, 20 September

10:00 AM → 10:30 AM A Density-matrix Renormalization Group Algorithm for Simulating Quantum Circuits with a Finite Fidelity

In this talk, I will introduce a recently developed density-matrix renormalization group (DMRG) algorithm for the simulation of quantum circuits (https://arxiv.org/abs/2207.05612). This algorithm can be seen as the extension of time-dependent DMRG from the usual situation of hermitian Hamiltonian matrices to quantum circuits defined by unitary matrices. Like an actual quantum computer, the quality of the DMRG results is characterized by a finite fidelity. However, unlike a quantum computer, the fidelity depends strongly on the quantum circuit considered. For the most difficult possible circuit for this technique, the so-called "quantum supremacy" benchmark of Google Inc., we find that the DMRG algorithm can generate bitstrings of the same quality as the seminal Google experiment on a single computing core. For a more structured circuit used for combinatorial optimization (Quantum Approximate Optimization Algorithm or QAOA), we find a drastic improvement of the DMRG results with error rates dropping by a factor of 100 compared with random quantum circuits. Our results suggest that the current bottleneck of quantum computers is their fidelities rather than the number of qubits.

Speaker: Thomas Ayral (Atos Quantum R&D Program, Atos SE, France)

10:30 AM → 11:00 AM Advanced Pulse-level Simulations of Rydberg Atoms Platforms

Programmable arrays of hundreds of Rydberg atoms have recently enabled the exploration of remarkable phenomena in many-body quantum physics in recent years. Those devices now operate in regimes which are very hard to simulate classically. In addition, the development of high-fidelity quantum gates are making them promising architectures for the implementation of quantum circuits. Being able to faithfully simulate these devices for intermediate sizes represents an important challenge. Studying and exploring the outcome of quantum programs with optimized CPU and GPU-enabled simulations allows us to prototype and devise new procedures for our quantum processor. In this talk, I will present several advances in quantum control and numerical simulations applied to neutral-atom devices. By using powerful numerical simulations describing the precise dynamics of our processors, we are able to push our understanding of the system and devise noise-robust entangling protocols.

Speaker: Loic Henriet (PASQAL, France)

11:00 AM → 11:30 AM Techniques and Challenges for HPC Simulation of Noisy Near-term Quantum Devices

Accurate and high-performance simulation of both gate and pulse level noise models is an essential tool for the research of both algorithm performance, and the effectiveness of error characterization, mitigation, and correction techniques on near-term quantum devices. In this talk we will introduce several ways of modeling pulse and circuit level noise processes in HPC simulations; describe challenges in implementing accurate device-level noise models for simulator designers and users; and discuss trade-offs in accuracy vs performance and their suitability for different applications. While doing this we will also give an overview of the capabilities and performance of various noisy HPC simulation techniques available in Qiskit.

Speaker: Christopher Wood (IBM Quantum, USA)

11:30 AM → 1:00 PM Break

1:00 PM → 1:30 PM Simulations of QAOA Quantum Circuits Using a Machine Learning Approach

With the increasing size of NISQ devices, it is important to find novel approaches to classically simulate large quantum circuits. Recent works show that machine learning (ML) models allow the efficient simulation of variational quantum algorithms (like QAOA). However, although ML models typically achieve great success in simulating quantum circuits, there are cases where such models perform worse than expected.

In this talk, we will try to answer three questions: why does ML underperform? When does it underperform? And, how can we improve the simulation fidelity? We will show that the simulation quality is highly related to the sample quality and simulation path. We will also show that a circuit can be difficult to simulate when the entanglement entropy of the target state is high. Finally, to mitigate the quantum state approximation error, we proposed two heuristics: an MCMC sampler that utilizes the problem symmetry and a greedy strategy to choose the simulation path.

Speaker: Zichang He (University of California, Santa Barbara)

1:30 PM → 2:00 PM Maximum Likelihood Decoders of Stabilizer Codes Under Device Noise Using Tensor Networks

Quantum error correcting (QEC) codes promise arbitrary suppression of logical errors as the size of the code increases. This tradeoff between experimental resources and error suppression makes the prospect of scalable quantum computing possible. Decoders play a key role in QEC protocols. They infer what class of logical error is most likely to have occurred during a computation based on information coming from a few sparse measurements performed on the system and a model of the underlying error mechanisms. Decoding is in principle a hard classical problem, and heuristics have been developed to decode efficiently in practice. In this talk I will present our latest advances in close-to-optimal decoding of stabilizer codes using device-level error models and tensor networks. In addition, I will present the results of our tensor network decoder applied to Google’s quantum processors. These results were recently used to support the first ever experimental demonstration of error suppression using the surface code.

Speaker: Benjamin Villalonga (Google AI Quantum, USA)

2:00 PM → 2:30 PM Pushing the Limits—and Breaking the Rules—of Quantum Simulation

As quantum computing hardware gets better and better, we will continually need to reinvent classical simulation techniques. This naturally means pushing the capabilities of classical algorithms progressively further. But it also means taking a wider perspective of what we wish our simulators to do, and maybe even blurring the lines separating purely classical simulation from quantum hardware execution. In this talk, I will survey some recent developments by the Xanadu team across this spectrum: compilation strategies for quantum-classical hybrid algorithms, advanced simulation techniques using HPC and accelerator devices, new circuit-cutting techniques, as well as advocating for a more holistic view of the classical computing needs of quantum algorithms researchers.

Speaker: Nathan Killoran, (Xanadu, Canada)

2:30 PM → 3:15 PM Break

3:15 PM → 3:45 PM Simulating Quantum Circuits and Quantum Devices as Efficiently as Possible

In this talk we discuss two approaches to simulating quantum circuits and devices. First, we discuss current work on developing a massively parallel open quantum system emulator using tensor networks with a focus on neutral atom quantum computers. We demonstrate our scaling as a function of nodes; contrast and benchmark two different ways to represent the quantum state of the system; and explore how approximations inherent to tensor networks (i.e. limited bond-dimensions) affect the accuracy of the simulation. Secondly, we discuss a new paradigm for simulating quantum circuits which differs from state-vector and tensor-network approaches. In this alternative approach, we map the quantum state to a probability distribution and then variationally evolves this probability distribution using autoregressive models.

Speaker: Bryan Clark (University of Illinois Urbana Champaign, USA)

3:45 PM → 4:15 PM cuQuantum: High-Performance GPU Libraries For Quantum Circuit Simulation

Quantum circuit simulation is a critical step for developing, validating, and debugging novel quantum algorithms designed for the gate-based quantum computing model. It allows quantum computing researchers, scientists and engineers to harness the classical computing power for simulating an ideal, fault-tolerant quantum computer way before one is ready for use in production. However, it requires delicate design and implementation to bring satisfying performance, and graphic processing units (GPUs) are a natural candidate for tackling massively parallel computation. In this talk, I will present the cuQuantum SDK developed by NVIDIA, a collection of high-performance GPU libraries currently offering state vector and tensor network based simulation methods. Since its debut, the cuQuantum SDK has been adopted by major quantum computing frameworks such as Qiskit, Cirq, and Pennylane, as we ensure it is easily accessible to C/C++ and Python users, thus allowing straightforward integration into any existing workflow. I will discuss the design principles and key features that we bring to our GPU users. Finally, I will briefly comment on the future plan for cuQuantum, including multi-node and multi-GPU support.

Speaker: Leo Fang (NVIDIA Corporation, USA)

4:15 PM → 4:45 PM Pulse-level Control and Dynamics of Qudit Gates

Qudit gates for high-dimensional quantum computing can be synthesized with high precision using numerical quantum optimal control techniques. Large circuits are broken down into
modules and the tailored pulses for each module can be used as primitives for a qudit compiler. Application of the pulses of each module in the presence of extra modes may decrease their effectiveness due to crosstalk. We address this problem by simulating qudit dynamics for circuit quantum electrodynamics (cQED) systems. Our results show that the frequency shifts due to crosstalk yield extremely stringent bounds on interaction parameters and spectator mode occupations. Here, we provide an experimentally relevant fidelity scaling formula that is independent of the gate type and can be used as a bound on the fidelity decay. The estimated scaling of the fidelity matches the scaling calculated using our numerical results.

Speaker: Baris Ozguler (Fermi National Accelerator Laboratory, USA)