Quantum Processing Unit Slicing for Concurrent Job Execution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hybrid computing systems, quantum processing units (QPUs) often have idle resources due to inefficient usage, leading to increased wait times and reduced productivity, as users select QPUs based on the number of qubits required for their quantum circuits, resulting in underutilization of available qubits.
Innovation Solution
Implementing a quantum slicing mechanism that allows multiple quantum jobs to be executed concurrently on a single QPU, where the QPU is divided into partitions to accommodate jobs with varying qubit requirements, and results are separated accordingly, enabling more efficient use of resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users select a QPU based on the number of qubits required for their quantum circuit, then the quantum circuit can be executed, but the QPU has idle resources and reduced productivity
Solution Approach 1:
The QPU is segmented into multiple virtual QPUs, each capable of handling quantum circuits with different qubit requirements. This segmentation allows the physical QPU resources to be divided and allocated to multiple users simultaneously, ensuring that each user gets adequate resources while maximizing overall utilization.
Solution Approach 2:
A single physical QPU is made universal by enabling it to serve multiple users and execute multiple quantum circuits concurrently. The system allows the QPU to function as multiple virtual instances, each tailored to specific user requirements, thereby eliminating idle resources and improving productivity.
2Productivity
If multiple quantum jobs are executed concurrently on a single QPU, then resource utilization improves, but the system complexity increases
Solution Approach 1:
A distribution engine acts as an intermediary between users and the QPU, managing the complexity of concurrent job execution. This intermediary component handles job scheduling, resource allocation, and result distribution, thereby enabling multiple quantum jobs to run concurrently without directly complicating the QPU architecture.
Solution Approach 2:
Virtual copies of the QPU are created through the slicing mechanism, allowing multiple users to interact with what appears to be separate QPU instances while actually sharing the same physical resources. This copying approach simplifies user interaction while enabling efficient resource utilization.
3Reliability
If a QPU has more qubits than required by a user's quantum circuit, then the user's circuit can execute, but wait times increase due to idle resources
Solution Approach 1:
The QPU maintains continuous useful action by eliminating idle time between quantum circuit executions. Through virtual slicing and concurrent job execution, the system ensures that the QPU resources are continuously utilized by multiple users, thereby reducing wait times while maintaining reliable circuit execution.
Data Source
AI summary
Quantum processing unit slicing is disclosed. The qubits of a quantum processing unit are sliced or grouped to accommodate multiple independent and separate quantum jobs. The quantum jobs are matched, based on user tolerances related to at least number of shots, and the quantum jobs are then merged and performed. The results for each of the specific quantum jobs concurrently performed are extracted from the overall results of the quantum processing unit.