Topic #: CW26-T15 · Branch: DOE · Status: Open
The Office of Science (SC) Office of Advanced Scientific Computing Research (ASCR) has spent decades on, and invested millions of dollars in, the development of high-performance computing (HPC) software that operates efficiently on large, heterogeneous supercomputers. Today, this hardware (e.g., CPUs, GPUs, TPUs, ASICs) has permeated society at large, finding its way into everything from smart phones to cloud computers. However, many of the software packages and libraries that can take advantage of this heterogeneity have remained solely within the HPC ecosystem.
Posted: 10/30/2026 · Opens: 10/30/2026
This button copies the full topic as plain text: the work, the dates, the apply steps, and the readiness list. Paste it into the AI you already use. Ask it if the work fits your firm, what you are missing, and what to skip.
This call starts with a short pitch, not a full proposal. Use the copy to judge the fit and shape the pitch. The call follows SBA policy: no limit on AI help, and no disclosure is required. Your firm signs the pitch and answers for every claim in it.
The first step is a short pitch, on the ConnectWerx site.
ApplySelect Apply. The official call page opens on the ConnectWerx site. The first step is a short pitch in the ATI portal. Full proposals are by invitation only.
Send this topic to a partner. The link opens this page.
Eligibility comes first. SBIR and STTR awards go to small businesses: for-profit, based in the United States, 500 or fewer employees counting every domestic and foreign affiliate, and at least 51 percent owned and controlled by U.S. citizens or permanent residents. A company owned by a large parent usually cannot apply on its own. It can take part as a subcontractor under an eligible small business.
An ATI portal account is set up for the person who will submit the pitch. Pitches go in through the portal, and the account takes time to set up. To pitch, the portal account is what you need.
If your pitch is invited, the full application needs SAM.gov entity registration active. Start it now: a new registration takes 10 or more business days to become active.
If your pitch is invited, the full application needs the SBA Company Registry complete. You get a SBC Control ID. Start it now: identity checks cannot be rushed.
The first step is a pitch: 4 narrative questions, 3 short-answer questions, and 7 structured selection questions, due by the date above at 2:00 p.m. Eastern. Full proposals are by invitation only. Plan to submit the pitch early.
Most topics need more than one small company can give. When people join a team, the members can cover the full scope of the topic together. Your email address stays private. Other members contact you through this site first.
Fill in the form to add your name to this list. Enter your role, a short text about what you bring, and your email address.
Send this page. It is the direct route: the link opens this topic, with the apply steps on it. The official listing opens a search page, and you need the topic number to find this topic there.
The Office of Science (SC) Office of Advanced Scientific Computing Research (ASCR) has spent decades on, and invested millions of dollars in, the development of high-performance computing (HPC) software that operates efficiently on large, heterogeneous supercomputers. Today, this hardware (e.g., CPUs, GPUs, TPUs, ASICs) has permeated society at large, finding its way into everything from smart phones to cloud computers. However, many of the software packages and libraries that can take advantage of this heterogeneity have remained solely within the HPC ecosystem.
While many ASCR-supported software packages are open-source, they are often complicated to use, distributed primarily in source-code form targeting scientific HPC systems, and potential adopters lack options for purchasing commercial support, training, and custom-development services. The expertise required to install and use these software packages poses a significant barrier to many organizations due to the levels of complexity built into them to facilitate scientific research. Moreover, without a commercial interest in broadly marketing the capabilities of the software, possibly including in markets beyond HPC, adoption is limited by a lack of exposure within the wider technology ecosystem. Providing simpler interfaces targeted for specific markets or offering a spectrum of commercial services around the underlying open-source software would make these software packages more usable for commercial, industrial, and nonscientific applications.
Goals and Metrics • Work proposed under this topic must critically depend on one or more software packages that has been funded through ASCR or through the Exascale Computing Project (ECP, 2016-2024, https://www.exascaleproject.org/). Applications should include a reference (web page or other citation) to show that the relevant software has been supported by ASCR or ECP. Relevant ASCR-funded software packages include, but are not limited to: o Mathematical Libraries: SuperLU (https://portal.nersc.gov/project/sparse/superlu/), STRUMPACK (https://portal.nersc.gov/project/sparse/strumpack/), HYPRE (https://www.llnl.gov/casc/hypre/), Trilinos (https://trilinos.github.io/), PETSc (http s://www.mcs.anl.gov/petsc/), SUNDIALS (https://computing.llnl.gov/projects/sundials), MFEM (https://mfem.org/). o Programming Models: Kokkos (https://github.com/kokkos/kokkos), RAJA (https://github.com/LLNL/RAJA), Umpire (https://github.com/LLNL/umpire), Legion (https://legion.stanford.edu/)
o I/O: ADIOS2 (https://github.com/ornladios/ADIOS2), Parallel NetCDF (https://parallel-netcdf.github.io/), HDF5 (https://www.hdfgroup.org/) o Compilers and Runtimes: LLVM (https://llvm.org/), Argobots (https://www.argobots.org/) o MPI: OpenMPI (https://www.open-mpi.org/), MPICH (https://www.mpich.org/) o Package Management: Spack (https://spack.io/) o Software Stacks and SDKs: E4S (https://e4sproject.github.io/), xSDK (https://xsdk.info/). Please note that E4S and xSDK include many ASCR-funded software packages that are not separately listed in this document. o Super Intelligence: DeepHyper (https://deephyper.github.io/), LBANN (https://github.com/LLNL/lbann). ASCR-funded software that enables an application to SI is in-scope even if it is not specific to SI. o Software for Quantum Computing and Information: Proposed products stemming from ASCR-funded software and algorithms for quantum information science are in-scope. Examples of such software and algorithms include BQSKit (https://bqskit.lbl.gov/), the quantum compiler framework, Jaqal (https://www.sandia.gov/quantum/quantum-informationsciences/projects/qscout-jaqal/), the assembly language for trapped ion quantum computers, Superstaq (https://infleqtion.com/quantum-software/), quantum compilation and control tools, and QubiC (https://gitlab.com/LBL- QubiC/software), qubit control system. See https://science.osti.gov/ascr/Research/Quantum-Information-Science- QIS for more information. ASCR-funded software that enables an application to quantum information is in-scope even if it is not specific to quantum information science. o Performance Tools: HPCToolkit (https://hpctoolkit.org), PAPI (https://icl.utk.edu/papi/), Dynist (https://github.com/dyninst/), TAU (http://www.cs.uoregon.edu/research/tau/home.php), Darshan (https://www.mcs.anl.gov/research/projects/darshan/). o Software maintained by the members of the Consortium for the Advancement of Scientific Software, https://cass.community/. o Software packages that have received ASCR funding as part of SciDAC (https://www.scidac.gov/) and other computational partnerships are also eligible. Please see the references for an additional, partial listing of available software packages and examples of their uses. Applications without a critical dependence on one or more ASCR-funded software package are out of scope. • Applications are sought to take one or more ASCR-funded software packages and make them easier to use or to incorporate ASCR-funded libraries into existing or new commercial software packages, or some combination of these adaptations. Example projects could include adding new capabilities, providing graphical user interfaces, making APIs for additional programming languages, hardening codes, and offering user support. Successful applications will address the potential both to broaden audiences for ASCR-funded software and to contribute to its stewardship. • A diverse community of stakeholders contributing to the maintenance and evolution of a software package lowers the long-term risks associated with commercialization of that software, including through the timely correction of software defects. Accordingly, ASCR encourages contributing bug fixes, changes needed to make the ASCR-funded software
function on generally available platforms, and other non-proprietary enhancements of general utility to the ASCR-funded software back to the project in a manner consistent with any applicable licensing requirements and other project policies. While not required, applicants are encouraged to provide letters of support from at least one developer of each ASCR-funded software package that plays a significant role in the proposed work. This letter should outline the mutually understood procedure via which any relevant contributions will be reviewed for acceptance into the project and briefly outline any anticipated prerequisites to initiating that procedure (e.g., the future execution of a Contributor License Agreement). • ASCR will consider collaborative applications from teams of small businesses under this topic, with up to three small businesses forming a team. Each institution in such a team must be a small business. Each institution may include one or more academic or lab partners as subcontractors. Each institution must submit an application that contains an identical narrative section and a common statement describing how any intellectual property issues will be addressed by the collaboration. Each application must have an institution-specific budget and budget-justification forms, biographical data for the PI and senior personnel involved in the project, and a commercialization plan. The budget proposed for each participating business must separately comply with the ceiling, floor, and other requirements in the Funding Opportunity Announcement. The cover sheet for each submission must clearly show all institutions involved in the collaboration.
References 1. U.S. Department of Energy, Software, Scientific Discovery through Advanced Computing (SciDAC), U.S. Department of Energy, June 12, 2025, https://www.scidac.gov/softwarelist.html. 2. U.S. Department of Energy, SciDAC Feature, Scientific Discovery through Advanced Computing (SciDAC), U.S. Department of Energy, June 12, 2025, http://www.scidac.gov. 3. ECP, ECP Software Technology Capability Assessment Report–Public, ECP-RPT-ST- 0002-2020-Public, February 2020, https://www.exascaleproject.org/wpcontent/uploads/2020/02/ECP-ST-CAR-V20-1.pdf. 4. U.S. Department of Energy, Exascale Computing Project, June 12, 2025, https://www.exascaleproject.org/. 5. U.S. Department of Energy, Office of Scientific and Technical Information, DOE CODE, June 12, 2025, https://www.osti.gov/doecode/. 6. Jenny Servo, Market Research Study: Applications of ASCR HPC Software, 2022, https://www.osti.gov/biblio/1874151/.