The following post was contributed by CCC Director, Ann Schwartz Drobnis.
The Networking and Information Technology Research and Development (NITRD) National Coordination Office (NCO), on behalf of Federal agencies and the NITRD National Science and Technology Council’s (NSTC) Fast-Track Action Committee (FTAC) on Strategic Computing (SC), put out a Request for Information (RFI) from the public to update for the Strategic Computing R&D goals and approaches.
The Computing Community Consortium (CCC) responded to the RFI on behalf of the community.
Some snippets from CCC’s response:
Many—if not most—of the benefits that information technology has provided to society have in turn depended on tremendous progress in technology (Moore’s Law) and in hardware designs for compute, storage, and communication. Future information technology benefits cannot assume further transparent technology/hardware progress. Thus, we face three options: seek to provide benefits without technology/hardware change, work with hardware for synergies to affect change (e.g., reduced precision, specialized accelerators), or await new (post-CMOS) technologies to develop. Prudence suggests all three options should be exploited.
-
We recommend expanded efforts, especially with respect to AI’s and deep neural network’s unsustainable demand for compute, memory/storage, and bandwidth, which can perhaps be aided by an increased focus on coordination (co-design) from algorithms to hardware.
-
We encourage industrial involvement to inform and perhaps support pre-competitive efforts to complement industrial development to aid industry and American competitiveness in the long run.
The CCC hopes that the NITRD NCO and the FTAC on Strategic Computing will view the full recommendations in the response as valuable input from the computing research community that can be used to update the Strategic Computing Objectives.
