TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
An IBM patent application describes software that would schedule compatible users’ quantum programs together and reuse stored information about processor noise to reduce repeated error-mitigation measurements. The filing outlines possible scheduling and billing features, but provides no measured results for throughput, waiting times or costs.
IBM’s patent application for “Quantum Queue Optimization for Error Mitigation” describes software that could run compatible quantum jobs from multiple users at the same time while reusing stored information about processor noise. The application, filed May 5, 2025 and approved on Nov. 5, 2026, sets out a proposed scheduling method; it does not report measured gains in hardware throughput, user wait times or costs.
The proposed scheduler would inspect jobs waiting in a queue and identify programs that could use different parts of the same quantum processor concurrently. It would check whether enough suitable qubits are available, whether the needed connections between qubits are supported, and whether users’ preferences allow their jobs to run near other activity. It would also account for job timing and priority.
A central feature is reuse of noise models: stored descriptions of errors associated with particular operations and qubit connections. Some error-mitigation methods require extra measurements to characterize those errors. The filing says that, when a combined set of operations matches information the system has already collected, the scheduler could reuse that information rather than repeat the measurements. Models would need regular updates because processor noise changes over time; outdated models could be discarded.
The application illustrates sharing on a 127-qubit processor: one program uses 52 qubits and another uses 30, with four unused qubits between the groups to help limit crosstalk. Other qubits could remain available for additional work. The filing also describes stricter and more permissive approaches to sharing, including taking new measurements or deriving needed noise information from existing models. It mentions machine learning as a possible aid to placement decisions, rebates for users who allow sharing, and separating the results of each job after execution.
Quantum processors are limited resources, and a job that occupies only part of a chip can leave other qubits unused if it receives exclusive access. Running compatible programs together could let providers schedule more work on available hardware. Reusing noise information could also reduce repeated preparation measurements, one source of extra work in error mitigation.
Those potential benefits are not established results. The filing does not quantify how much additional work the method could support, whether it would shorten queues or lower costs, or how often compatible jobs and sufficiently current noise models would be available. Its significance for users will depend on whether the proposed coordination can manage interference and security without undermining the reliability of results.
Popular right now
Dyson Airsmooth Styling Brush
Dyson V8 Cyclone Cordless Vacuum
DREAME A3 AWD 1000 LiDAR Robot Lawn Mower
As an affiliate, we earn on qualifying purchases.
Noise Models Shape Scheduling
Sharing a processor is not simply a matter of placing two programs on separate qubits. Operations in one region can affect another through crosstalk, changing the error patterns that need to be accounted for. Noise information is tied to particular active connections, the filing says, so a model for one arrangement does not automatically apply to another.
The application describes scheduling as a balance between available qubits, circuit connections, program duration, user preferences, interference and job priorities. It could move a waiting program forward when doing so allows it to use a noise model before that information becomes outdated. The proposal addresses a related but distinct problem from Columbia Engineering’s HyperQ, which The Quantum Insider previously reported was tested on IBM processors and uses quantum virtual machines to assign portions of a processor to different programs. IBM’s filing focuses on coordinating concurrent jobs with error-mitigation measurements.
““Quantum Queue Optimization for Error Mitigation””
— IBM patent application, “Quantum Queue Optimization for Error Mitigation”
Performance and Safeguards Unreported
The application does not provide measured comparisons of the proposed scheduler against existing scheduling methods. It reports no figures for throughput, waiting times or costs, so the practical size of any efficiency gains remains unknown. The filing also does not establish how often existing noise models would match combined jobs or how frequently new measurements would be needed.
IBM’s application identifies challenges including interference between programs, possible information leakage between users and the coordination of jobs with different requirements. The supplied report does not explain what safeguards would be implemented or provide test results showing how those risks would be handled. Patent approval also does not by itself establish that the method has been deployed in a commercial service.
Evidence Needed From Implementation
The next evidence readers would need is a technical demonstration or deployment report comparing this scheduling approach with existing practice. Useful results would show how many jobs can run concurrently, how often models can be reused, how model updates affect scheduling, and whether results remain reliable when programs share hardware.
No implementation date or product launch is specified in the supplied material. It is also not clear whether IBM plans to use the method in a service or how any rebates would be calculated. Until those details or performance measurements are available, the filing should be read as a description of a proposed system, not proof that quantum-computing access will become faster or cheaper.
Key Questions
What does IBM’s patent application propose?
It describes software that would identify compatible quantum jobs, run them together on different portions of a processor, and reuse stored noise information where applicable.
Does the application show that quantum jobs will run faster?
No. The filing reports no measured improvements in throughput, waiting times or costs. It describes a proposed method and its intended efficiencies.
Why does the scheduler need noise models?
Noise models describe errors associated with operations and qubit connections. Reusing a suitable, current model could avoid repeating some measurements used in error mitigation.
What problems could arise when users share a processor?
The application identifies interference between programs, potential information leakage and the challenge of coordinating jobs with different requirements. The supplied material does not report test results resolving those concerns.
Does the patent mean IBM has launched this system?
No launch or deployment is identified in the supplied report. Approval of the application does not establish that the proposed scheduler is in use.
Source: thequantuminsider.com
