2022 INCITE Call for Proposals
New Proposal Preparation Instructions for Authors
Up to sixty percent of the allocable time on the IBM/AC922 Summit and the Intel/Cray Theta machine at the ALCF will be allocated for calendar year (CY) 2022 through the INCITE program. For each resource, allocations are anticipated to be between 500K and 1M Summit node-hours and 1.5M and 2.5M Theta node-hours. Individual awards may be higher. ALCF is introducing Polaris, a new accelerate resource for 2022. More information on Polaris will be forthcoming – check Resources for updates.
INCITE seeks research enterprises for capability computing: production simulations – including ensembles – and data analytics and artificial intelligence (AI) applications that necessarily require our leadership class resources. Please note that requests smaller than 20% of the average award noted above may be considered too small for the INCITE program, unless the proposal is (i) for algorithm development or other computer science activities requiring a large fraction of the machine resource (one or more of computing, memory, network, etc.) but not a large amount of time or (ii) clearly articulates the need for other aspects of the architecture and/or infrastructure associated with leadership computing resources. Please contact the INCITE program manager (INCITE@DOEleadershipcomputing.org) if you have questions regarding the allocation request for your proposed project allocation size.
INCITE provides several resources to help you prepare your proposal.
- The INCITE Overview and Policies includes a description of the basis for award decisions, among other things.
- The questions used by reviewers to assess proposals are available, as are templates for the narrative sections of the proposal.
- The slides from previous Informational Webinars about the INCITE program are available.
- Early access on leadership computing facility (LCF) resources can be requested through the Director’s Discretionary program if you wish to generate benchmarking data on these systems.
Experts will carry out scientific and technical reviews of the potential impact of the proposal and the ability of the applicant’s team to effectively use the Summit and Theta systems. Potential scientific impact is the predominant determinant for awards. INCITE awards are large—often 100 times greater than more generally available allocation programs—and a limited number of projects are selected each year. Campaigns chosen by the INCITE program typically cannot be performed anywhere else and require extremely large high-performance computing systems, large awards of time, high performance storage or networking capabilities or the unique LCF architectural infrastructure to succeed.
Revisions for 2022 INCITE Call for Proposals:
- The Early Career Track has been added for those PIs who are within 10 years from earning their PhD (PhD on or after December 31, 2011).
- Polaris, a new resource at the ALCF, has been added for 2022.
It is strongly recommended that authors comply with the guidelines established below, because they will be used to assist in the review of proposals. Templates for all sections are available.
The proposal must be clear and readily legible and must conform to the following requirements:
- Header/Footer/Page Number: Each section of the proposal must be paginated. Footers should be used for paginating all files. Also, headers should be used to indicate the title of the proposal and the lead PI.
- Title: Proposal titles may not be longer than 80 characters.
- Font: One of the two typefaces must be used: Arial or Times New Roman (font size 11). A font size of less than 11 points may be used for mathematical formulas or equations; for figure, table, or diagram captions; or when using a symbol font to insert Greek letters or special characters. PIs are cautioned, however, that the text must still be readable.
- Margins: Margins must be at least 1 inch in all directions. These requirements apply to all sections of the proposal, including supplementary documentation.
- Spacing: Proposals should be prepared using single line spacing. The proposal elements should not exceed the specified page count limits.
- References: References should be gathered at the end of the narrative. References are not included in the total page count.
- File Sizes: The total file size should be limited to 15 MB.
- Support Information: No letters of collaboration or letters of support will be accepted with the application. Current and pending support documentation is not required.
Adherence to type size and spacing requirements is necessary to ensure readability and that no proposer will have an unfair advantage by using smaller type or spacing to accommodate more text.
All proposals must be submitted electronically via the INCITE submission site.
Electronic applications will be accepted starting Monday, April 12, 2021, through 8:00 pm EDT on Friday, June 18, 2021. All proposals must follow these instructions. INCITE reserves the right to decline consideration of proposals not compliant with these instructions and guidelines. Hardcopies will not be accepted for review.
Any questions should be directed to the INCITE program manager at INCITE@DOEleadershipcomputing.org.
- Project Executive Summary (1-page limit): The executive summary should accurately describe the proposed research and the high-impact scientific or technical advances you will realize with the proposed INCITE allocation. Industry organizations should also summarize the potential economic or strategic business impact of the proposed research.
- Project Narrative: The narrative should not exceed 15 pages. Visual materials—such as charts, graphs, pictures, etc.—are included in the 15-page limit. References do not count toward the 15-page limit and should be included at the end of the Project Narrative. URLs that provide information related to the proposal should not be included. The 15-page limit will be strictly enforced. The Project Narrative should address the following points:
- Significance of Research: Explain what advances you expect to be enabled by an INCITE award that justifies an allocation of petascale resources (e.g., anticipated impact on community paradigms, valuable insights into or solving a long-standing challenge, etc.). Place the proposed research in the context of competing work in your discipline or business. List any previous INCITE award(s) received and discuss the relationship to the work proposed. The information should be sufficient for peer review in your area of research and also appropriate for general scientific review, comparing your proposal with proposals in other disciplines. Data analytics and artificial intelligence (AI) proposals may find the DOE AI for Science report valuable as a guide when writing the Significance of Research. Potential scientific or business impact is the predominant determinant for awards. This factor will be assessed by a peer review panel. This section is typically about 4 pages.
- Research Objectives and Milestones: Describe the proposed research, including its goals and milestones and the theoretical and computational methods it employs. Goals and milestones should articulate simulation and developmental objectives and be sufficiently detailed to assess the progress of the project for each year of any allocation granted. Milestones should correlate with those in Section 4, “Milestone Table.” It is especially important that you provide clear connections between the project’s overarching milestones, the planned production simulations, and the compute time expected to be required for these simulations (e.g., should correlate with Section 2.3.i, “Use of Resources Requested”) in the research proposal. You should also make clear any dependencies of milestones on other milestones. This section is typically about 6 pages.
- Computational Readiness: Proposals will be assessed on the need for, readiness to use, and reasonableness of the request for resources. Proposals should summarize the requirement(s) that best exemplifies the proposed computational work. Leadership targets in the INCITE program typically include one or both of the following categories:
- Simulation, data analytics and/or AI projects should use a significant fraction (of the order of 20% or more) of one or more of the LCF leadership class resources; compute, memory, network or disk, for example. Parameter sweeps, ensembles, design of experiments, and other statistical methods that require large numbers of discrete or loosely coupled simulations may be considered capability-class campaigns. See the FAQs for details and qualifiers.
- Specific architectural needs that can only be met by the LCF.
This section, including the following subsections, is typically about 5 pages.
- Use of Resources Requested: Describe your proposed production simulations and state how the runs are tied to each of your project’s goals and milestones (Section 4, “Milestone Table”). Also describe the data requirements of your production simulations. If at any point during your project the sum of your data storage needs in the scratch file systems exceed 1 petabyte, specific justification is required. See the Project Narrative Template for a detailed listing of information required in this section.C
- Computational Approach: Provide a detailed description of your computational approach, including a discussion of the state of the art in the field. The description should also mention:
- Particular libraries required by the production and analysis software, algorithms and numerical techniques employed (e.g., finite element, iterative solver), programming languages, and other software used.
- Parallel programming model(s) used (e.g., MPI, OpenMP/Pthreads and vector intrinsics [AVX-512] for Xeon Phi; MPI, OpenMP/Pthreads, CUDA, OpenACC, HIP, ROCm or AVX intrinsics for AC922).
- Project workflow, including the role of analysis and visualization. Identify where the analysis will be done and any potential bottlenecks in the analysis process. Describe any analysis and/or data reduction tools used.
- Software workflow solution (e.g., pre- and postprocessing scripts that automate run management and analysis) to facilitate this volume of work.
- I/O requirements (e.g., amount, size, bandwidth) for restart, analysis, and workflow. Highlight any exceptional I/O needs.
- Optimizations for the resources requested (in terms of efficiency, scalability, throughput, data input/output, workflow tools for ensemble runs, checkpointing etc.) which may be relevant for data analytics and AI proposals.
- Parallel Performance: Provide direct evidence, including supporting quantitative data, for your production application’s parallel performance for the intended research simulations. Ideally, the proposing team will have generated the data and this data will be representative of the entire workflow of the project proposed. If you cite work by others, explain why it is applicable here. You should use the application code you intend for the production work, not a related code. Data for sample systems not related to the intended research is undesirable. Performance benchmarking should reflect all I/O and workflow requirements. Scaling and performance data in either strong or weak scaling mode should be provided. Explain how the strong or weak scaling applies to the proposed work. For data analytics and AI proposals, a description of the compute, memory, networking, and storage needs in the context of the LCF resources will help reviewers understand the strong or weak scaling characteristics. See the examples at the end of this page.
NOTE: You may apply for a startup account at one of the centers to conduct performance studies. Director’s Discretion applications are available at
- Developmental Work: For the computational approach above, describe what, if any, development work has been carried out to-date, especially on the architecture of the requested resource. Describe what development work will be executed during the proposed INCITE campaign and when it will be executed. Provide an estimate of the computational resources required for this work. If applicable, identify the milestones and production activities in Section 2.3.i that are dependent on the developmental work and provide a plan for validating this developmental work.
- Personnel Justification and Management Plan: A personnel justification and management plan must be included in the proposal. This section is typically 1 to 2 pages, but may be longer for community proposals. (Does not count toward the 15-page Project Narrative limit).
- Personnel Justification: What personnel are already in place and what are their roles on the project? If applicable, describe (i) personnel that will be hired for the project in the future and their responsibilities and (ii) potential personnel turnover that may occur during the project and a strategy for replacing them. The INCITE program welcomes proposals from individual PIs or teams of collaborators.
- Management: Describe the project’s leadership team and how decisions are made to allocate time to individuals or, for proposals with multiple collaborating teams, subgroups within the project. Describe the focus of each individual or subgroup. Successful proposals will include a management plan that conveys to reviewers the interrelationship between subgroups and how the sum of the parts will lead to scientific discovery or engineering solutions that are the overarching goal of the project. Also identify points of contact who will provide updates on the status of the work, including publications, awards, and highlights of accomplishments.
- Milestone Table: Proposals must be accompanied by a summary table of planned milestones for each year of the proposed work. Milestones should be clearly articulated and appropriate for the size and length of the requested award (e.g., large requests should have sufficient milestones to allow reviewers to assess the planned project workflow). Future modification to the project scope and milestones will be tracked in part through the milestone table. Use the template provided. Below is a description of what should be provided for each column in the table. (Does not count toward the 15-page Project Narrative limit).
Milestones: Clearly state the scientific and technical (e.g., development) milestones for each year of the proposed work.
Details: Include the following details as appropriate for each milestone.
- Resource: List the primary computing resource.
- Node-hours: List the number of node-hours associated with this milestone.
- File system storage: Provide an estimate in terabytes (TB) of the required temporary storage needed on the file system while the data is analyzed and reduced or before it is moved to archival storage. Also provide the estimated time period when this storage is needed.
- Archival storage: Provide an estimate in TB of the required archival storage associated with this milestone for the duration of the project. Also provide the estimated time period when this storage is needed.
- Application: List the software application or code needed for this milestone.
- Tasks: Briefly summarize the work that will be done by listing the subtasks, computational runs, and data analysis and reduction tasks associated with each milestone.
- Dependencies: Note dependencies between the scientific milestones and on proposed development work.
Date: List the projected start and completion date for each milestone.
Status: For renewal proposals only.
- Publications Resulting from INCITE Awards: (Only required for projects with prior INCITE awards). Provide a list of publications resulting from previous INCITE awards to this project team for work related to the proposal under consideration. Only those publications that include an acknowledgement to INCITE and/or the LCF may be included. This list may not be used in lieu of references in Section 2, “Project Narrative.” If applicable, list the citations in both Sections 2 and 5. (Does not count toward the 15-page Project Narrative limit).
- PI/Co-I Biographical Sketch: Provide a biographical sketch for each PI and Co-I listed on the proposal. The biographical sketch appendix will not count in the project narrative page limitation and, for each PI or Co-I, must not exceed two (2) pages.
As part of the sketch, provide information that can be used by the reviewers to evaluate the team’s ability to perform the proposed work and utilize the resources effectively. The biographical sketch must include:
- Education and Training: Undergraduate, graduate and postdoctoral training, provide institution, major/area, degree and year.
- Research and Professional Experience: Beginning with the current position list, in chronological order, professional/academic positions with a brief description.
- Publications: Provide a list of the 5 publications most closely related to the proposed project. For each publication, identify the names of all authors (in the same sequence in which they appear in the publication), the article title, book or journal title, volume number, page numbers, year of publication, and website address if available electronically.
- Research Interests and Expertise: Provide a paragraph describing the research interests and expertise related to the proposed project.
- Synergistic Activities: List no more than 5 professional and scholarly activities related to the effort proposed.
- Collaborators: List in alphabetical order all persons, including their current organizational affiliation, who are, or who have been, collaborators or co-authors with you on a research project, book or book article, report, abstract, or paper during the 5 years preceding the submission of this proposal. For publications or collaborations with more than 10 authors or participants, only list those individuals in the core group with whom the PI interacted on a regular basis while the research was being done.
- Personally Identifiable Information (PII): Do not include sensitive and protected PII including social security numbers, birthdates, citizenship, marital status, or home addresses. Do not include information that a merit reviewer should not make use of.
Examples of Performance Required Materials
Quantitative data for production application performance should be provided in either tabular or graphical form or both. This data should reflect the performance of the application for the production simulations proposed, and should include all I/O requirements of the production simulations.
Where appropriate, characterize the production application’s single-node performance (e.g., percent of peak). For example, describe the most computationally expensive portion of your algorithm and describe the on-node parallelization scheme employed. Any optimization strategy to improve single-node performance—including code restructuring, GPU strategy, and exploiting OpenMP/Pthreads and vectorized instructions—should be described. For Summit, provide any evidence of speedup using GPUs compared to using all CPU cores. For Theta, demonstrate the efficient use of many CPU cores per node, and if not all cores can be used effectively, explain why. For all architectures, indicate the degree of utilization of any architecture-specific features. Data analytics and AI proposals may provide performance data which demonstrates scalability, efficiency, node or GPU utilization, network bandwidth estimation, or power profiling for node, CPU, GPU and/or memory.
Weak-scaling behaviors are probed by holding per-node computational work constant (e.g., the size of the mesh on a processor is held constant) as the total problem size grows with processor count. Strong-scaling behaviors are probed by holding the total problem size constant as the processor count grows, thereby decreasing the per-processor computational work. Note that in the examples provided, a logarithmic scale is preferred.
For data analytics and AI applications that may weak scale for improved convergence, present a justification for the scientific impact of the expected improvements when running at large scale. Address, if necessary, overcoming difficulties of training with respect to the relevant hyperparameter list for deep learning applications.
In addition, where appropriate the proposal should describe the entire computational workflow of the proposed project. In particular, the project should demonstrate that the entire workflow is tenable on the proposed architecture or supporting analysis resources at the LCFs or discuss where any pre- and postprocessing or data analysis will be conducted and how relevant files will be transferred.