Ten year Anniversary of The Workshop on Monitoring and Analysis for HPC Systems Plus Application (HPCMASPA)

HPCMASPA, held in conjunction with IEEE Cluster, is the workshop focused on monitoring, observability, and response in large-scale applications that support scientific computing, at-scale databases, financial trading, web traffic, and similar distributed workloads.

Ten year Special Topic:

HPC, cloud, and data-analysis-focused computing are converging. New insight and assessment are needed to understand the dynamic resource demands of these computing paradigms, what constitutes figures of merit of performance for a diverse and/or integrated workload, what opportunities are there for co-scheduling and dynamic scheduling, and more. 

HPCMASPA in its 10 year solicits research and position papers (both full and short) as well as lightning talks meant to enable the revolutionary change in monitoring and analysis needed to ensure performance of this converging computational ecosystem.

Research/Position Papers (Full or Short/Work In Progress)

Full or short peer-reviewed papers describing novel work that is complete or in progress.  Potential topics include, but are not limited to:

• How HPC system and application monitoring can be influenced by how cloud and web companies are monitoring their environments and applications
• What data, analytics, figures of merit for performance apply in different computing cases and how monitoring and analysis will need to change to address these concurrently
• Challenges in monitoring and analysis in converged environments (e.g., microservices vs tightly-coupled applications)
• Changes to systems and software to support feedback from and dynamic response to monitoring and analysis
• Traditional HPC monitoring and analysis with consideration of relevancy to changing architectures and workflows

Review criteria for technical papers will include technical soundness, significance and potential of work/results, and originality of work. Review criteria for position papers will include soundness of principles, new insights presented, and potential for transformative directions.

Lengths:

• Full: 8 pages plus 2 additional reference-only pages
• Short/WIP: 4 pages plus 2 additional reference-only pages

Lightning Talks

Short, non-peer-reviewed talks that will be selected by the organizing committee.  Potential topics include, but are not limited to:

• Breaking news, new positions, and stimulating insights on the potential topics above
• Driving coordinated change in monitoring and analysis tooling, software, and architectures to enable interoperability of HPC, cloud, and data-analytic workflows

Lightning talk proposals should include a few sentences describing the topic, project, or activity to be covered as well as a short outline of the proposed talk.  Note that the Lightning Talk length is very limited, so proposals should describe a single core idea. Review criteria for lightning talks will include perceived innovation, interest, and ability to engage the audience.

Lengths:

• Lightning Talks: 6 minutes

Dates

• Full Papers Due: June 5, 2023
• Full Paper Notifications: July 10, 2023
• Short Papers Due: June 26, 2023
• Short Paper Notifications: July 17, 2023
• Full and Short Camera Ready Papers Due: July 25, 2023
• Lightning Talk Proposals Due: Mid-September

Submissions

• Web-based submissions through EasyChair.
• PDFs only.
• Submissions must be compliant with the format used by IEEE Cluster. LaTex and Word templates can be found here.
• No additional pages can be purchased.
• Submissions must be in English.
• Submissions must adhere to the IEEE standards for AI-generated content.
• Submission implies the willingness of at least one of the authors to register and present the work associated with submission in accordance with the policies of IEEE Cluster.

For more information: https://sites.google.com/view/hpcmaspa2023