Call for Contributions

Scope and Topics

The ACM/SPEC International Conference on Performance Engineering (ICPE) is the leading international forum for presenting and discussing novel ideas, innovations, trends, and experiences in the field of performance engineering.

Performance, energy efficiency, reliability, sustainability, and trustworthiness are central to the acceptance and long-term viability of modern computing systems. Digitalization brings computing technology closer to people and manages many aspects of their lives. In turn, human interaction, automation, and increasingly AI-enabled decision-making shape the behavior of computing systems. As a result, our systems become more complex, more adaptive, and more difficult to engineer, observe, and understand.

ICPE brings together researchers and practitioners to report on open problems, state-of-the-art solutions, and in-progress research in performance engineering of software and systems, targeting performance and associated quality attributes such as efficiency, reliability, scalability, sustainability, and cost in all phases of the computing system lifecycle, from specification and development to run time and maintenance.

Application Domains

ICPE deals with performance and associated quality attributes across application domains. Examples include, but are not limited to:

  • cyber-physical systems, IoT, industrial Internet, robotic systems, and autonomous systems
  • communication networks, wireless systems, mobile systems, and distributed systems
  • peer-to-peer environments and ad-hoc networks
  • service-oriented architectures, microservices, web-based environments, cloud-native systems, and serverless applications
  • big data systems, stream processing, graph processing, data analytics, and data science
  • machine learning, artificial intelligence, LLM-based systems, AI infrastructure, and agentic AI applications
  • HPC systems, cloud platforms, edge platforms, grid systems, fog computing, and accelerator-based platforms
  • multimedia systems and applications
  • resilient, adaptive, and self-managing systems and applications
  • production software systems, observability platforms, and large-scale operational environments

Types of Contributions

ICPE 2027 invites contributions across the following tracks:

  • Research track for novel research on the conference topics
  • Industry track for research involving industry and production performance-engineering practice
  • Industry Presentation track for presentations from industry practitioners showcasing their views and experiences in performance engineering (new)
  • Journal-first track for contributions from recent journal publications
  • Artifact track for sharing software, data, benchmarks, traces, and other reusable research artifacts
  • Emerging research track for preliminary, exploratory, or vision contributions
  • Poster and demo track for short interactive contributions
  • Tutorials track for presentations by leading scientists and practitioners
  • Workshops track for focused communities and specific aspects of performance engineering
  • Data challenge track for addressing challenges presented by performance datasets

To get more details on each track, please refer to the tracks and submissions page.

Topics of Interest

Topics of interest include, but are not limited to, the following.

Measurement and Empirical Evaluation

  • data collection techniques, simulation, measurement, instrumentation, profiling, tracing, telemetry, and observability
  • controlled experiment design, data-driven experiments, diagnostics, and performance debugging
  • data management and interchange, tool interoperability, and open performance datasets
  • statistical analysis, exploration, visualization, visual data mining, and uncertainty-aware analysis
  • measuring and evaluating performance and related attributes such as energy efficiency, carbon footprint, cost, stability, or resource usage
  • measuring and evaluating dependability and related attributes such as security, resiliency, availability, privacy, or confidentiality
  • measuring and evaluating algorithmic attributes such as learning efficiency, solution quality, model quality, and decision quality
  • measurement and evaluation of AI/ML systems, LLM inference pipelines, and agentic workflows

Modeling

  • approaches and metrics for describing performance and related quality attributes
  • modeling languages and formalisms, modeling methods and tools
  • explanatory and predictive models, online models, causal models, and uncertainty-aware models
  • model learning and extraction techniques
  • model validation, calibration, transfer, and generalization across workloads and environments
  • workload characterization, workload forecasting, and workload generation
  • models for cost, energy, carbon, dependability, and performance tradeoffs

Design and Development Processes

  • requirements engineering, agile and experiment-driven techniques, DevOps, MLOps, and LLMOps
  • performance-oriented design, software architectures, performance patterns, and performance antipatterns
  • performance testing, regression detection, root cause analysis, and bottleneck identification
  • humans in the loop, developer experience, operational usability, and ethical concerns
  • performance engineering for AI-enabled, autonomous, and safety-critical systems

Managing Systems at Runtime

  • adaptive systems, monitoring, autotuning, elasticity, autoscaling, power management, and carbon-aware control
  • virtualization and consolidation, resource scheduling, capacity management, and admission control
  • anomaly detection, service-level objectives, service-level monitoring, incident response, and production observability
  • runtime management of ML/AI inference, LLM serving, batching, routing, caching, and model selection
  • performance and reliability of agentic systems, multi-agent systems, and human-AI operational workflows

Platform-Related Optimizations

  • parallel programming, multi-core systems, and many-core systems
  • compiler optimizations, managed languages, and runtime optimization
  • performance and efficiency of hardware accelerators, GPUs, TPUs, NPUs, and novel memory systems
  • cloud, serverless, edge, container, microservice, Kubernetes-based platform optimizations, and lightweight virtualization technologies
  • cost-performance, energy-performance, and latency-throughput tradeoffs in modern platforms

Benchmarking

  • design and standardization processes
  • benchmark suites, investigative benchmarks, and benchmark governance
  • benchmark synthesis, benchmark workload generation, and realistic production workloads
  • benchmarking energy efficiency, carbon impact, resilience, stability, security, and related quality metrics
  • benchmarks for AI/ML systems, LLM serving, agentic systems, observability pipelines, and cloud-native systems
  • reproducibility, repeatability, and reusable benchmark artifacts

ACM Publication Policies and Open Access

By submitting an article to ICPE 2027, authors acknowledge that they and their co-authors are subject to ACM Publications Policies , including policies on authorship, research involving human participants and subjects, publication integrity, and author name changes.

AI Tools Policy: AI tools may not be listed as authors. Authors must follow the ACM Policy on Authorship regarding the disclosure of AI tool use where applicable.

Open Access Publishing: All ACM publications, including those from ACM-sponsored conferences such as ICPE 2027, are 100% Open Access. Authors have two primary options for publishing Open Access articles with ACM: the ACM Open institutional model or by paying Article Processing Charges (APCs).

  • Authors affiliated with an ACM Open participating institution will have their papers covered by the program (no cost to the author).
  • Authors from institutions not participating in ACM Open will be required to pay an APC, unless they qualify for a financial or discretionary waiver. To find out whether an APC applies to your article, please consult the list of participating institutions in ACM Open and review the APC Waivers and Discounts Policies (geographic) and the APC Waivers and Discounts Policies (discretionary) . Please note that waivers are rare and based on specific criteria set by ACM. A temporary subsidized APC will apply in 2027: USD 500 for ACM/SIG members and USD 750 for non-members. This represents a 29% discount off current APC list pricing, funded directly by ACM. Authors are encouraged to help advocate for their institutions to join ACM Open during this transition period.

ORCID: Authors of accepted papers are required to obtain and provide ORCID IDs to complete the ACM publishing process.