IBM recently announced updates to Qiskit Runtime, its containerized quantum computing service, and programming model. Users can optimize workloads and efficiencies and execute them on quantum systems at scale with the service. Qiskit Runtime also makes it easier for non-physicists to interface and experiment with quantum computers by deploying complete programs than rather circuits. The new updates help simplify quantum computing even more and should save developers many hours of detailed programming, freeing them for more interesting and creative work.
Qiskit Runtime is equipped with two new primitives as part of its service. Primitives are predefined programs that make it easy to create quantum-classical workloads needed to build and customize applications.
IBM has also added a much-needed third pay-as-you-go consumption pricing option for access to IBM cloud service and its 27-qubit Falcon processors. The cost of the new pricing option is $1.60 for each second of Qiskit Runtime use. That makes it an ideal plan with manageable costs for startups, small businesses, and lightly funded university programs.
The new Qiskit Runtime updates, new pricing options, and the addition of primitives create significant competitive advantages for IBM.
Qiskit Runtime innovation
Before developing Qiskit Runtime, early IBM quantum research focused on increasing the execution speed of quantum circuits and quantum operational sequences. Speed was, and still is, important to quantum computing for several reasons:
- Qubit quantum states rapidly deteriorate, which means quantum operations must be initiated and completed before the quantum state collapses.
- Qubits are fragile and subject to errors caused by environmental factors such as noise, cabling, other qubits, and galactic radiation. Quantum error correction research indicates promising methods that may eventually be exploited for low error quantum computing. However, a fully fault-tolerant quantum computer is probably at least five to seven years away.
- Most algorithms require the interaction of quantum and classical computers. It is not uncommon for some applications to have millions of back and forth looping interactions between the two processors that create latency with each successive computational loop.
In addition to its early pursuit of circuit speed, IBM research scientists also recognized the benefits of accelerating the execution of entire quantum programs. Early in 2021, IBM introduced Qiskit Runtime to fill this need.
Qiskit Runtime established two new quantum computing paradigms:
- It was the first application to use containers in the quantum cloud to increase programming efficiency and speed. Containers are executable software packages that make applications more portable by carrying application code and necessary libraries and dependencies.
- IBM changed its infrastructure and architecture to collocate quantum and classical computers to reduce hybrid computing latency for algorithms
Without automation, setting up an algorithm and running it is a complicated and lengthy process. The developer must first determine which error mitigation, classical, and quantum algorithms will be used before the problem is recast to fit the quantum computer. It is then necessary to decide how many times the program should be run (shots) and how to interpret results within the proper expectations. If this process is to be continued as a hybrid classical-quantum solution, then recurring full loop calculations are required on the classical computer so it can use data from the quantum computer.
IBM has greatly simplified this process. In the Qiskit Runtime execution environment, the IBM hybrid cloud handles much of the work using its software architecture and containers.
The first two primitives in Qiskit Runtime – Sampler and Estimator – optimize how code is sent to a quantum computer. By sampling quantum circuits, Sampler generates outputs that help determine a solution to the computation. Estimator is a program interface that estimates expected values of quantum operators so that users can calculate and interpret quantum operator expected values needed for many algorithms.
Classical computers are deterministic and provide precise answers. On the other hand, quantum computers are probabilistic and provide non-classical probability distributions that, depending on how many times the program is run, give a good idea of the answer. Since almost all quantum algorithms use probability distributions, Sampler and Estimator are likely to have broad applicability across the entire spectrum of quantum algorithms.
Qiskit Runtime improvements created a 100x speedup in iterative circuit execution workloads. Depending on the problem, computations that took a month can now be solved in days or hours by Qiskit Runtime. But as good as this seems, according to IBM, Qiskit Runtime will eventually be able to run 200,000 times faster than it does now.
1. To eliminate any confusion, Qiskit by itself is a hardware-agnostic software development kit that simplifies the ability to build, compile, run, and analyze quantum circuits and quantum programs. Qiskit makes it easy to control the interaction between quantum software and quantum hardware.
2. According to IBM, three major factors affect a quantum computer’s performance: quality, speed, and scale. IBM uses a holistic measurement called quantum volume to determine and improve circuit quality. It measures scale by qubit count. IBM’s roadmap shows qubits scaling substantially every year in the future. By the end of 2022, IBM plans to replace its current 127-qubit Eagle processor with a 433-qubit processor. In 2023, IBM expects to replace that with a newer quantum computer with over 1000 qubits. IBM’s ultimate goal is to deliver a fault-tolerant quantum computer with over a million qubits sometime beyond 2026.
3. By 2023, IBM plans to offer groups of pre-built runtimes tailored to specific domains and callable from a cloud-based API.
4. Qiskit Runtime finally puts frictionless quantum computing on the horizon, making IBM’s projection of 2025 seem doable.
Follow Paul Smith-Goodson on Twitter for current information and insights on quantum and AI