Pasqal Integrates NVIDIA CUDA-Q into Hybrid Quantum Computing Environment
Summarize this article with:
Pasqal has integrated the NVIDIA CUDA-Q platform with its Quantum Resource Management Interface (QRMI) runtime, bringing quantum processing into standard high-performance computing workflows. This new capability allows CUDA-Q workloads to be scheduled on Pasqal’s neutral-atom quantum systems using the widely adopted Slurm system, effectively making quantum processors native accelerators in complex computing environments. The integration aims to lower barriers to quantum adoption for HPC users by embedding quantum tasks within their existing operational models for job submission, scheduling, and monitoring. “HPC users don’t want a new operational model to access quantum capabilities,” said Wasiq Bokhari, Pasqal’s Chief Executive Officer. “By integrating CUDA-Q into our HPC-native environment with the QRMI, we’re enabling Pasqal quantum processors to be used within hybrid GPU-QPU workflows.” This advancement will first be deployed at CINECA, integrating Pasqal’s QPU with the Leonardo supercomputer. Pasqal and NVIDIA Integrate CUDA-Q with QRMI Runtime Neutral-atom quantum computers are now directly compatible with standard supercomputing workflows thanks to a new integration between Pasqal and NVIDIA, which promises to accelerate the development of hybrid quantum-classical applications. This advancement avoids the need for entirely new operational models, a significant hurdle in broader quantum adoption. The integration centers on enabling CUDA-Q workloads to be scheduled using Slurm, a widely adopted HPC workflow manager, through QRMI. This means researchers and engineers already familiar with Slurm can submit jobs that seamlessly incorporate quantum processing power alongside traditional CPUs and GPUs. QRMI, initially developed by IBM with collaborative contributions from Pasqal, RPI, and STFC Hartree Centre, functions by exposing quantum processors as schedulable resources, handling authentication, allocation, and monitoring within the existing HPC infrastructure. This design philosophy, described as hardware-, modality-, and vendor-agnostic, is intended to minimize disruption to established HPC practices. The initial deployment of this integrated system will occur at CINECA, Italy’s leading supercomputing center, linking Pasqal’s quantum processing unit with Leonardo, a EuroHPC pre-exascale supercomputer. Currently, the integration is accessible via Pasqal’s cloud platform, offering immediate access for developers. This is a practical step toward making quantum acceleration usable at scale, alongside CPUs and GPUs, for real applications in optimization, simulation, and AI. NVIDIA’s Director of Quantum Product, Sam Stanwyck, further emphasized the accessibility benefits, stating, “CUDA-Q is designed to make hybrid quantum-classical computing accessible to developers by unifying quantum and HPC resources.” Sara Marzella, Responsible of Quantum Computing group at CINECA, anticipates that “Leonardo will be among Europe’s first supercomputers supporting hybrid HPC–QPU workloads in our standard Slurm environment.” This integration marks Pasqal’s progression toward production-grade hybrid workflows and lays the groundwork for future on-premises software components within their quantum computing stack. QRMI Enables Slurm-Based Scheduling for Quantum Processors The convergence of high-performance computing and quantum processing is steadily progressing beyond theoretical demonstrations, with a growing emphasis on practical integration within existing computational ecosystems. Currently, accessing quantum hardware often requires specialized workflows and expertise, creating a barrier for many researchers accustomed to traditional HPC environments. The company recently detailed a new integration between QRMI and the NVIDIA CUDA-Q platform, designed to streamline access to neutral-atom quantum systems for HPC users. This integration centers on embedding quantum workloads within standard Slurm-based HPC workflows. Slurm, a widely adopted workload manager, is the backbone of many large computing centers, and leveraging its existing infrastructure is a key strategy for Pasqal. QRMI functions by exposing quantum processing units as schedulable resources within Slurm, enabling secure allocation and monitoring alongside conventional hardware. Users can then submit jobs through familiar HPC interfaces, with QPUs automatically provisioned when quantum routines are called for. This collaboration will allow researchers to explore Slurm-native hybrid GPU and QPU workloads, pushing the boundaries of what’s possible with combined classical and quantum computation. NVIDIA’s involvement, through the CUDA-Q platform, further solidifies this interoperability. CUDA-Q is designed to make hybrid quantum-classical computing accessible to developers by unifying quantum and HPC resources. Sam Stanwyck, Director of Quantum Product at NVIDIA CINECA’s Leonardo Supercomputer to Support Hybrid GPU-QPU Workloads Pasqal is extending the reach of its neutral-atom quantum computing platform through a collaboration with CINECA, one of Italy’s largest high-performance computing centers. The partnership will integrate Pasqal’s quantum processing units (QPUs) with Leonardo, the EuroHPC pre-exascale supercomputer co-funded by the Italian Ministry of University and Research. This integration allows for the creation of hybrid GPU-QPU workflows managed within the standard Slurm workload scheduler, a widely used system in HPC environments. The move signifies a practical step toward incorporating quantum acceleration into existing computational infrastructure, rather than requiring entirely new operational paradigms. QRMI is designed to be hardware-, modality-, and vendor-agnostic, allowing quantum processors to function as schedulable resources alongside traditional CPUs and GPUs. By exposing QPUs through QRMI, Slurm can automatically provision them for quantum workloads, streamlining the process for HPC users. The system handles secure authentication, allocation, and monitoring, ensuring seamless operation within established HPC workflows. Pasqal’s on-premises stack, beginning with this CINECA deployment, is intended to provide a robust foundation for future software components and wider accessibility. The core principle driving this integration is minimizing disruption for existing HPC users. HPC users don’t want a new operational model to access quantum capabilities. By integrating CUDA-Q into our HPC-native environment with the QRMI, we’re enabling Pasqal quantum processors to be used within hybrid GPU-QPU workflows leveraging the existing resource management systems HPC teams already run in production. This is a practical step toward making quantum acceleration usable at scale, alongside CPUs and GPUs, for real applications in optimization, simulation, and AI. Wasiq Bokhari, Pasqal’s Chief Executive Officer Pasqal is streamlining access to its neutral-atom quantum processors for high-performance computing (HPC) users, potentially accelerating the integration of quantum capabilities into existing scientific workflows. This development bypasses the need for HPC facilities to overhaul their operational infrastructure to accommodate quantum computing, a significant barrier to entry for many institutions. The core of this advancement lies in treating quantum processing units (QPUs) as native accelerators within HPC environments. Leonardo will be among Europe’s first supercomputers supporting hybrid HPC-QPU workloads in our standard Slurm environment. Sara Marzella, Responsible of Quantum Computing group at CINECA Source: https://www.pasqal.com/newsroom/pasqal-introduces-new-integration-with-nvidia-cuda-q-to-enhance-its-hybrid-quantum-computing-environment-for-hpc/ Tags: