ICS 2027 Call for Papers

The 41st ACM International Conference on Supercomputing

The ACM International Conference on Supercomputing is the leading international forum for presenting research results in high-performance computing systems. The 41st edition of ICS will be held as part of the ACM Federated Computing Research Conference (FCRC) 2027 in Atlanta, Georgia, USA.

ICS 2027 solicits papers across the full stack of high-performance computing: architectures, systems software, and applications. Strong work in any of these areas is welcome, with a deliberate effort this year to broaden the program's balance toward architecture and algorithms alongside the conference's traditional strength in runtimes and programming systems.

Scope and Areas

Papers are solicited on all aspects of the architecture, software, and applications of high-performance computing systems at all scales, from microarchitectures to full supercomputing systems. What distinguishes ICS is its focus on architecture in the context of supercomputing and relevance to the future of HPC. Topics include, but are not limited to:

Architecture and Emerging Technologies. Processor, accelerator, memory, storage, interconnect, and full-system architectures, including designs built on future and emerging hardware such as quantum, superconducting, photonic, analog, and neuromorphic systems. Submissions relying on simulation methodology when appropriate are in scope. ICS aims to be a home for premier architecture work of relevance to the future of supercomputing and HPC, including security, reliability, and other emerging topics.

AI for HPC and HPC for AI. AI and machine learning for HPC system optimization, including learned schedulers, autotuning, performance prediction, and intelligent resource management across large-scale systems. GPU and HPC techniques for accelerating ML workloads, hardware-software co-design for training and inference, and scalable systems for foundation model training and deployment. Cross-stack techniques such as hardware-aware ML compilers, near-memory or in-network acceleration, and communication-efficient distributed training are also encouraged.

Algorithms and Applications. Parallel and high-performance algorithms and end-to-end algorithmic pipelines for large-scale data analytics, along with implementing and deploying algorithms and applications on large-scale systems. This includes processing and analysis of data from large-scale experimental instruments and automated laboratories, such as DOE light sources, where the computational demands are those of supercomputing and HPC. Algorithm-focused and end-to-end algorithmic contributions are especially encouraged.

Systems Software and Programming. Compilers, runtime systems, and system software, including optimization and energy management; programming languages, paradigms, and execution models, including domain-specific languages and scientific problem-solving environments. Cross-cutting work spanning the increasingly blurred boundary between applications and systems software is welcome.

Performance Tools. Tools for measurement, modeling, analysis, and visualization of performance, energy, sustainability, and other quantitative properties of high-performance computing systems.

In Scope vs. Out of Scope

Papers must demonstrate relevance to HPC or supercomputing. For example, FPGA, ASIC, SNN, and photonic or analog designs are in scope when they target a clear supercomputing or HPC application. Processing data from experimental instruments, such as DOE light sources, the LHC, or comparable facilities, is in scope when the underlying application requires HPC-scale computation.

The same hardware is out of scope when it targets applications without supercomputing or HPC relevance, such as FPGA or ASIC designs for hashing or video encoding. Consumer-oriented embedded, edge, and mobile computing, including power- or form-factor-constrained deployment without supercomputing-scale computational demand, is out of scope.

Because the boundary depends on the application rather than the technology, the submission form requires authors to explain how their work fits the ICS scope and why it is relevant to supercomputing and HPC. Precedents do not dictate what is in scope this year. For example, a prior ICS paper on filesystems does not by itself establish that filesystems are in scope.

What Makes a Strong ICS Paper

Papers are evaluated on novelty, technical soundness, and potential impact. Beyond that, the committee expects rigorous, convincing evaluation; clear relevance to supercomputing and HPC in some form; and end-to-end systems or software implementation with rigorous evaluation methodology.

Review Process

Details of the review process will be provided with the submission system.

Important Dates

Task	Date
Abstract submission deadline	December 1, 2026
Paper submission deadline	December 7, 2026
Early Reject Decisions Notification	February 11, 2027
Author Revision Window	February 11 – February 23, 2027
Final Decisions Notification	March 12, 2027

Workshops, Posters, and Student Research Competition

ICS 2027 will host a program of affiliated workshops alongside the main conference. A separate Call for Workshops will be issued; proposals for new and returning workshops will be welcome. The conference will also feature a poster session and the ACM Student Research Competition, giving students a venue to present early-stage work and receive feedback from the community. Separate calls for both will follow.

Visas

To reduce visa and immigration-related delays, each paper submission form must include a checkbox asking whether the author anticipates needing a visa letter to attend the conference, for example: “If accepted, do you anticipate needing a visa letter to attend the conference?” The form must include a disclaimer that this information is not visible to reviewers and will not be used in any decision on the paper; it is collected purely to speed up the process. Our goal is to issue visa letters at the same time as Final Decision Notification.

Junior Program Committee

The ICS Chairs are excited to announce the 1st Junior Program Committee (Junior PC) for ICS 2027, an initiative to train the next generation of HPC reviewers. It offers early-career researchers, including senior PhD students and postdocs who have not yet served on the ICS technical research track PC, a chance to gain reviewing experience and learn from senior researchers in the field. Junior PC members participate fully in reviews and online discussions, but do not attend the PC meeting or take part in final decisions.

ICS 2027 Junior PC Self-Nomination Form . Please fill out the form by October 10, 2026. ICS 2027 PC Chairs will contact selected nominees only by October 30, 2026.