IEEE Quantum-HPC Working Group
Technologies to engineer, operate, and use quantum computers and simulators (QCS) are still evolving to achieve reliable and scalable QCS systems for practical applications. Recent efforts to interface QCS and HPC systems highlight an opportunity to expand access and adoption of this rapidly developing capability by leveraging existing expertise and capabilities in conventional computing.
The primary objective for Quantum-HPC integration has been towards leveraging QCS systems as accelerators for HPC systems and applications. Other objectives include the use of HPC systems for operational control, compilation and optimization, as well as error correction management of QCS systems. An emergent mechanism for this has been at HPC centers across the world interfacing classical HPC systems with QCS systems - either co-located within the same datacenter or remotely accessible from an HPC cluster, possibly as a network of federated systems.
The requirements to interface HPC and QCS systems cover the full stack:
- hardware infrastructure deployment and operation
- networking and interconnectivity
- system software interoperability
- application programming and workflow integration
- hybrid use-case identification, demonstration and evaluation
- interdisciplinary knowledge transfer and skills development
The Quantum-HPC Working Group in the IEEE Quantum Initiative was established to systematically identify and address these open challenges by bringing together experts and stakeholders from the relevant disciplines.
Objectives & Activities
The Quantum-HPC Working Group will define and steer the development of methods, tools and skills for building, operating, using and supporting hybrid high-performance quantum computing and simulation systems. To achieve this, the working group will use the hub-spoke model illustrated below to systematically address multiple topics, stakeholders, and open challenges.
This spoke addresses different approaches and measures to interface and integrate QCS systems with HPC systems, including aspects related to hosting, co-location, federation, tight and loose coupling. Particularly, preparing data centers and HPC environments for the special requirements of QCS systems, power and communication network infrastructure, and operational considerations that cover security, SLA, management of consumables, and operational data analytics. Specific needs related to QCS modalities such as photonic, superconducting, ion-trapped, neutral atoms and others will also be considered.
This spoke covers the need for system software required to access QCS from HPC systems either directly or remotely, in the latter case possibly via a federation, to provide the service platform through which to access, allocate and manage HPC and QCS resources, to deploy compilers and to perform static and dynamic analysis, and to schedule, execute and manage jobs with interleaved HPC and QCS components. The goal is to enable seamless access between HPC and QCS systems for applications and end-users, especially when computation needs to be triggered from within HPC jobs, to ensure effective system utilization and resource sharing by building new scheduling approaches, and to deliver the system-level features for a continuum of static to dynamic compilation through efficient low-latency distributed toolchains.
Software tools & programming
This spoke covers interoperability, integration and extension of classical HPC-related software and programming tools to increase the productivity of software and application developers on hybrid HPC-QCS systems. This includes development and/or extension of language abstractions and support, programming models with high-/low-latency support and runtime software with HPC and QCS capabilities, compilers, optimisers, profilers and debuggers. This topic will also address software to support high- and low-latency hybrid workflows, cross-platform portability, unified development environments, HPC-enabled QCS emulation, and performance characterisation and analysis of hybrid applications.
Hybrid system use-cases
This spoke addresses the identification and development of use-cases that effectively combine conventional HPC and emerging QCS systems to solve computational problems and serve to co-design HPC-QCS systems. This includes how applications and workflows are combined physically and logically as well as how hybrid libraries and data flows are managed. This topic will consider how to measure and evaluate the performance of hybrid system use-cases and pursue best practices for hybrid system demonstrations.
This spoke addresses the need to bridge and develop a community across the HPC and QCS domains with interdisciplinary and cross-domain competencies and skills. This targets the operation, provisioning and development of HPC-QCS systems (covering data center & operations, engineering & administration of HPC, QCS, cloud and communication network systems, system software development) as well as the programming and exploiting through hybrid system use-cases (covering quantum information science, algorithmics, HPC parallel programming, quantum programming, software & devops engineering, domain expertise).
The Quantum-HPC Working Group is a community of
- representatives of HPC centers (involved in the interfacing and service provision of HPC and QCS systems),
- quantum technology developers (involved in the engineering and development of quantum computing, simulation and communication systems),
- computer system architects, administrators and electrical engineers (driving the host and control side of HPC and QCS systems),
- HPC and QCS software developers (involved in system software, runtime and programming tools),
- HPC and QCS application developers and use-case owners.
The members and activities of the working group will be organized based on the spokes to implement structured discussions and developments towards community-driven interoperable standards and white papers towards developing solutions for deploying, operating and using hybrid HPC-QCS systems, tools and applications.
The Quantum-HPC Working Group will also engage and collaborate with TCHPC, an IEEE Computer Society Technical Community that focuses on high performance computing technologies and applications.
The activities of the Quantum-HPC Working Group will include online meetings as well as in-person workshops at related events such as IEEE Quantum Week, SC and ISC-HPC and dedicated Quantum-HPC events.
- Birds of a Feather (BoF) session at IEEE Quantum Week (QCE) 2023
- Tue, Sep 19 2023, 15:00-16:30 Pacific Time (PDT)
- Birds of a Feather (BoF) session at Supercomputing Conference (SC) 2023
- Tue, Nov 14 2023, 17:15-18:45 Mountain Standard Time (MST)
Laura Schulz, LRZ (Leibniz Supercomputing Centre)
Martin Schulz, TUM (Technical University of Munich)
Scott Pakin, LANL (Los Alamos National Laboratory)
Travis Humble, ORNL (Oak Ridge National Laboratory)
Venkatesh Kannan, ICHEC (Irish Centre for High-End Computing)