We are excited to announce several new products and capabilities, as well as enhancements to our existing offerings, that are designed to advance all three pillars of our technology roadmap — performance, scale, and enterprise-grade — with a particular focus on the “enterprise-grade” pillar.
First, we are relaunching our quantum operating system as IonQ Quantum OS. While IonQ has always built its own control software, we undertook enabling Forte Enterprise to be our first generation of enterprise-grade hardware. The additional reliability, flexibility, and performance we’ve added make it worthy of a new name and a fresh look at its capabilities.
Second, we are excited to share more details about a collection of new capabilities we are calling the IonQ Hybrid Services suite. First previewed at last week’s SC24 conference for partners like NVIDIA, Google Cloud, Amazon Braket, and Dell Technologies, the IonQ Hybrid Services suite improves the speed, performance, and usability of hybrid quantum-classical workloads on IonQ systems today, while also laying the groundwork for a roadmap of future enhancements.
Finally, we are proud to announce that we expect to bring our first IonQ Forte Enterprise system online before the end of this year, enhanced and supported by the software innovations higher in the stack. In addition to being a quantum computer with world leading performance and resilience, it is also the first system in our new European data center, and our first system to be manufactured in our Seattle manufacturing facility and assembled in a separate location.
Together, these advancements represent a major expansion and upgrade of the IonQ quantum computing stack. They are designed to provide meaningful application-oriented benefits now and set the stage for continued enterprise-grade enhancements to help enable our customers in the quantum-accelerated future.
Reintroducing our quantum operating system (OS): IonQ Quantum OS
Since our founding, we have viewed building our own control software and firmware as a strategic advantage. We believe this has consistently paid off, enabling rapid advancements in control automation, error mitigation, gate compilation, and more. For years, having our own software and firmware has been a crucial component of our systems’ consistently leading performance.
We've continued to iterate and improve upon our OS. Years ago, we updated the OS to enable IonQ Harmony to be the first quantum computer on all three major public clouds. Then, we leveraged our OS to reach a new frontier of performance and stability in IonQ Aria and IonQ Forte. By continually shipping updates to push performance and enabling ever-more-advanced customer workloads, IonQ is helping ensure world class performance for our customers.
As our next generation systems like Forte Enterprise become more performant, we’re continuing to engage with a growing base of commercial customers. Through this process, it has become clear that our original OS is no longer the right platform to enable our ambitious roadmap or our customers’ needs.
So, we’ve been quietly pursuing a nearly ground-up rewrite of that software stack, not only to ensure that Forte Enterprise could truly be the enterprise-grade system we wanted to be but also to help ensure that we have a platform that is ready for the next five or more years of advancements. This includes features like Barium qubits, RMQA, on-prem deployments, deep HPC integration, error correction, photonic networking, and ultimately, commercial quantum advantage.
Today, we’re excited to announce that IonQ’s latest quantum operating system, IonQ Quantum OS, is now powering our flagship quantum computers IonQ Forte and IonQ Forte Enterprise. IonQ Quantum OS has been running commercial workloads on IonQ Forte since the summer and will soon be controlling IonQ Forte Enterprise as it comes online in our Basel Switzerland data center later this month.
At IonQ, we envision a future where our quantum computers are integrated with existing compute resources within production environments at a scale that can solve the world’s most complex and impactful problems. This integration could be co-located with classical HPC resources on customer sites, enabled by remote cloud resources, in one of our own quantum data centers, or a combination of the three. This future requires enterprise-grade software that is reliable, resilient, and performant, while also being flexible, interoperable, and modular enough to meet our customers’ needs and keep up with the rapid pace of the rest of the quantum market. This is in stark contrast to the software running many of the experimental systems in today’s market, which often do not meet these enterprise-grade requirements. The result is IonQ Quantum OS, a containerized, modular, and flexible architecture that is designed to grow with our systems and deliver customer value in the hybrid quantum computing ecosystem.
Optimizing System Performance
The core job of any quantum operating system is application-driven performance. We aren’t done when we make individual qubits and gates as good as they can be, or even when we make algorithmic fidelity as good as it can be, but when the all-in, end-to-end performance is the best it can possibly be. There are two main ways we measure this: the quality of the solution, and the time it takes to achieve that solution.
With the help of an improved processing pipeline and better system observability and profiling, IonQ Quantum OS has been able to dramatically improve time-to-solution on IonQ Forte. To date, we’ve been able to achieve an average reduction of over 50% in on-system classical overhead, and an additional 85% reduction in processing and network overhead for workloads submitted.
These optimizations impact workloads based on their circuit structure, shot count, and overall size, so there is no single number for end-to-end speedup in all cases. However, in our tests, we have a 2x or better improvement in overall time to solution for a wide variety of workloads, and in certain extreme cases, have improved time to solution by a whopping 46x over the same workload run on IonQ Aria.
For specific applications, we have been able to combine these speedups with other optimizations at the application level to help achieve even greater performance. For the Matchgate Shadows technique we demonstrated at SC24 in collaboration with NVIDIA, we saw a speedup of 7.3x, pulling the total time to solution for an example case from 2.5 hours down to just over 20 minutes. Similarly, for the QITE algorithm we demonstrated in collaboration with Oak Ridge National Lab, we were able to show a speedup of 6.9x, taking the per-iteration time from 3.9 seconds down to 566 milliseconds.
Our analysis indicates that more improvements are likely possible and that already implemented speedups may become even more powerful as workloads grow in qubit count and classical complexity.
In addition to improving time-to-solution, IonQ’s Quantum OS also has the potential to enhance algorithmic results through several other avenues.
An upgraded error mitigation and compilation suite has demonstrated improvements of up to 100x1 in the accuracy of several commercially relevant applications. This is supported by a rebuilt error mitigation software framework allowing us to more effectively test and deploy additional error mitigation techniques in the future.
Additionally, enhanced calibration, automation, and control software/firmware combo have also allowed us to move to more and more advanced continuous calibration techniques that help continually optimize qubit-gate performance and adjust for systematic drifts. This results in better algorithmic performance and reduced performance decay, while also reducing the total amount of time we spend calibrating the system per week by nearly a factor of 6 when compared to this time last year.
A Focus on Interoperability
Our commitment to interoperability and the support of an open quantum ecosystem has long been evident. We are the first and only quantum computing company with systems available on all three major public clouds, we provide industry-leading support for open-source SDKs and programming models, and our robust APIs make it easy for platform and application providers like Classiq, QCWare, Q-CTRL, and others to easily integrate with our systems and tools. As we rearchitected IonQ Quantum OS, we used it as an opportunity to extend our focus on interoperability and openness to even more of our stack. This is especially true in relation to integration with on-prem classical resources to enable more advanced hybrid quantum computing paradigms. With a modular, containerized architecture, and clean and well-documented IRs and APIs, IonQ Quantum OS is well positioned to easily integrate with other on-prem hardware and software solutions. This helps enable a variety of new hybrid deployment and integration models now, and, as our customers continue to specialize their quantum applications and more deeply integrate quantum computing into their stack, will likely lead to the development and integration of customer and use case specific software and tools. This potentially includes integrating third party partner tooling at a deeper level than ever before.
Prioritizing Resilience, Deployability, and Security
Resilience, deployability, and security may not be as exciting as massive speedups in execution or new ways to integrate into customer environments, especially when most of our customers and partners are more focused on application research than production deployment of those applications, but we believe they are just as critical to our vision of the enterprise-grade, quantum-accelerated data center of the future.
Building these qualities from the beginning allows both IonQ and our customers to innovate at a much higher speed and with much lower risk than we likely could if working with experimental quantum computing software.
In a system as complex as a quantum computer, the classical compute resources and the software stack that enables it should be reliable. We want our customers to worry about results, not things like network issues, runaway processes, or memory leaks. IonQ Quantum OS provides a resilient, high-stability, high-availability foundation to help enable just that, and the robust automation system that sits on top of that helps ensure that the quantum resources are just as stable and available.
Thanks to these unseen but critical investments, we are able to deploy updates, upgrades, and security patches to our quantum systems just like we would do with any other server in a contemporary data center. We expect to be able to deploy an entirely new build of the IonQ Quantum OS to any system in the fleet in under five minutes, with high confidence, turning what typically can take days or even weeks in the lab into a simpler, more reliable process.
IonQ Quantum OS offers enterprise-grade security capabilities for similar reasons: we worry about it so our customers and users don’t have to. We expect the production workloads of the future to come with significantly higher security requirements than current-era workloads. IonQ Quantum OS is designed from the ground up with observability, stability, security, and reproducibility at its core. As a result, it’s designed to support a diverse set of customer security requirements and common standards to meet even the most security-minded enterprises’ needs with ease.
A Scalable Platform For The Future
As IonQ continues to execute on our ambitious hardware roadmap, IonQ Quantum OS represents a modular, extensible, scalable platform that is designed to adapt to and help enable future hardware generations and capabilities, including IonQ Tempo and our future photonically networked solutions. By continuing to invest in developing software in parallel with our hardware and alongside our customers, IonQ is helping pave the way for success in an ever evolving and expanding quantum computing market.
IonQ Hybrid Services: Powering the Quantum-Accelerated Future by Integrating Classical and Quantum Computing
We have to continue to build the world’s best quantum computers to achieve commercial quantum advantage, but that’s not all we need to do. To unlock problems that are currently impossible (or at least impractical) to solve with classical computing, we believe we will need those quantum computers to effectively and efficiently work alongside classical computing resources.
At IonQ, we believe there are many computational tasks where quantum computers may never provide a commercial quantum advantage over classical systems. As a result, we believe that the most effective problem-solving machine of the future is a hybrid system, where users can take advantage of a variety of computing paradigms together, leveraging the best of all of them to unlock new value across industries and applications.
We think about enabling this future in two main ways that start as separate efforts, but eventually blend into one seamless package:
Improving the ability of quantum computing to accelerate business-critical application workflows by acting as an accelerator in a tightly integrated hybrid computing runtime.
Using classical computing to help improve or enhance what can typically be done with quantum computers themselves, especially as it relates to error correction and qubit virtualization.
Last week at SC24, we announced that we have made significant headway in advancing both parts of this vision. Designed to help enable the seamless development and deployment of hybrid workloads that combine IonQ quantum computers with high-performance classical resources over the cloud, IonQ Hybrid Services is a crucial tool for delivering value to our customers and serves as a foundation for a variety of future enhancements.
Quantum Accelerates Classical
Our vision for quantum acceleration is to make using a quantum computer boring. We want our customers to be over-the-moon excited by the outsized value quantum-accelerated computing creates but using quantum resources to compute with or upon quantum information should be as simple and ubiquitous as GPU acceleration is today.
We are embarking on this journey by introducing a series of new capabilities in the IonQ Quantum Cloud, designed to enhance the speed, performance, and usability of executing quantum workloads while paving the way for even more advanced acceleration in the future.
At the core of these capabilities is our new Workload Management & Solver Service. Built on a flexible, containerized, cloud-based infrastructure, it helps make building and deploying hybrid workloads simple and reliable. By providing quantum developers with a toolkit to execute on an IonQ-managed runtime hook into a custom hosted runtime, these capabilities can lead to improvements in developer experience, time to value, and time to solution by moving hybrid workloads into the cloud with minimal infrastructure and configuration overhead.
The service introduces two new layers of functionality on top of a basic quantum circuit: a quantum function and a hybrid solver designed to simplify the process of manually building circuits and iterative algorithms and package them into portable workloads that can be executed and enhanced in a variety of ways.
Quantum functions allow developers to describe their quantum algorithms at a higher level of abstraction than a single circuit. This is accomplished by specifying classical pre- and post-processing logic that maps an input that is relevant to their domain area — like a hamiltonian, graph, or even just a parametrized circuit — to one or several quantum circuits to be executed on QPU. Quantum functions then map the results of those executed circuits to an application-relevant output format, such as an observable, eigenvalue, or error-mitigated circuit result.
Hybrid solvers allow developers to compose individual quantum functions into larger applications with a variety of pre-built and customizable solvers and optimizers, simplifying common algorithmic approaches, and allowing application researchers to focus on exploration and innovation, not infrastructure.
To help ensure that these workloads can be run as quickly and efficiently as possible, we are also launching hybrid Sessions, a new scheduling feature built on top of the Fair Share scheduling engine that we rolled out earlier this year. Sessions help enable the streamlined provisioning of QPU time for large, iterative workloads, ensuring that iterative and resource-intensive workflows run with minimal wait and interruption without completely blocking out other users.
Not only is the speed of execution enabled by Sessions valuable in its own right, but it is also expected to drive algorithmic performance improvements by ensuring more consistent system behavior across the duration of the hybrid workload.
A variety of new ecosystem integrations and tools make these new features even more valuable and easy to use: the all-new IonQ SDK helps make integrating with the IonQ Quantum Cloud API and IonQ Hybrid Services simple, seamless, and reliable. Our improved scheduling and workload management features are designed to provide expanded support for the hybrid-related offerings of our best-in-class cloud partners, including Amazon Braket Hybrid Jobs and Azure Quantum Sessions.
Classical Enhances Quantum
A variety of classical computing techniques can also meaningfully improve or enhance what we can do with the quantum computers themselves, helping enhance the capabilities and performance of the portion of these quantum-accelerated workloads that we normally think of as purely quantum.
Many of these techniques require mid-circuit measurement and dynamic circuit execution, capabilities that we plan to first enable in our classical simulation tools, to help ensure that customers are ready to take advantage of them by the time they are brought to life on IonQ Tempo systems.
IonQ Hybrid Services is already accelerating how we build real-world, commercial quantum hybrid applications with our customers. Our recent work with Oak Ridge National Laboratory was enabled by a beta version of these capabilities, and the numerous hybrid applications created with our world-class customers, such as Airbus, DESY, and Hyundai, helped inform the service’s capabilities and design.
Launching Forte Enterprise Direct Access
Last year, we announced our plans to build our first European data center in Basel, Switzerland, in collaboration with QuantumBasel. We are excited to share that our first quantum computer in the Basel data center, Forte Enterprise, is expected to be online on schedule, later this year. Once online, QuantumBasel and their team of experts will have dedicated access to a large portion of the system’s capacity, and IonQ will make the remainder available to customers via the IonQ Quantum Cloud and through our access partners.
Forte Enterprise represents the first delivery of our next generation system, built for the modern data center with a standard rack form factor, low energy profile, and minimal environmental isolation requirements. In the coming weeks, Forte Enterprise will begin running IonQ Quantum OS bringing enterprise-grade software features to our first generation of enterprise-grade hardware. Together, IonQ Forte Enterprise running IonQ Quantum OS will help advance IonQ’s vision for production-ready quantum computing for the enterprise.
IonQ Forte Enterprise Commissioning - QuantumBasel
IonQ delivers, installs and commissions our first quantum computer in Europe.
This instance of Forte Enterprise in our European data center is not only the first Forte Enterprise to come online for customer access, but also the first computer IonQ has built and commissioned in a customer’s facility. Thanks to longstanding investments in production engineering and manufacturing, IonQ is leading the way in designing and deploying data center ready quantum computers. We look forward to additional deployments of IonQ computers in the future in both IonQ’s data centers and partner facilities.
To learn more about Forte Enterprise and to explore access options please reach out to our sales teams at https://ionq.com/get-access.
1 https://arxiv.org/abs/2301.07233↫