Defense Advanced Research Projects Agency (DARPA), Contracts Management Office (CMO), 3701 N. Fairfax Dr., Arlington VA 22203-1714

A -- METHODS AND TOOLS FOR PERFORMANCE MODELING, MEASUREMENT, ANALYSIS, EVALUATION AND PREDICTION SOL BAA97-12 DUE 021197 POC Dr. Frederica Darema, DARPA/ITO, FAX: (703)522-7161. DARPA is sponsoring a competition to address research directions in performance methods and tools for modeling, measurement, analysis and evaluation of parallel and networked systems, and the applications executing on such systems. We use the term Performance Technology for this research thrust. There are in development, at any given time, future generations of advanced computing and information systems driven by the needs of applications of the highest leverage for science and education, national security, and national economic competitiveness. Realization of the benefit from these systems requires that they perform with high efficiency on high-end resource-intensive applications. Effective integration of enhanced capabilities of measuring, modeling, analyzing, and predicting the performance of computing systems into the hardware and software design process could lead to substantial decrease in the development time of future advanced computing systems and increase the effectiveness of these systems when they are delivered. Future advanced computing systems will be actualized by distributed/networked systems (heterogeneous processors, metacomputing systems) embodying parallel architectures and complex memory hierarchies, continuing the trend seen in today's high-end systems. Many factors affect the performance of such complex systems and the applications executing on them. These include the processor and hardware system architecture, the network architecture, various system software components, and the mapping of end-user applications and their algorithms to the architecture. There exists an array of isolated modeling and simulation methods and tools, aimed to understand the behavior of isolated components in these systems. The methods and tools range from queuing models to instruction-driven simulators. However, the existing methods do not have the capability to provide the well-defined hierarchical structuring needed to model systems of the level of complexity under consideration. To understand the behavior of such systems, substantial improvements in the current methods and tools are needed to measure, model, and analyze computer systems, at all levels (from the application, to the software, to the hardware level), and to develop a performance framework to permit analysis of the entire system. Major technical challenges include the development of multilevel methods and tools (application and application-enablers levels, system software, and hardware levels, including the network levels ), hierarchical or multiscale approaches (multiple levels of resolution for components at each level), and scalable approaches to modeling the behavior of the entire system, or the behavior at each level as affected by behavior of components in the other levels, and how the behavior of these components or the system scales as one moves from the present prototype to the future machine. The DARPA supported efforts in ECAD for multilevel modeling and design methods at the device, chip, and board levels have contributed to major advances in the design of such components. The present thrust is intended to enable equivalent capabilities at the computing system levels above the device and chip levels. This competition is aimed to foster research for developing performance methodologies that have predictive as well as evaluation capabilities. In particular, we are interested in methods that provide hierarchical or multilevel analysis of such systems, enable assessment of the effects of individual hardware and software layers and components of the systems, and have predictive capabilities in terms of scaling of the applications or the systems. Various approaches to developing tools that implement such modeling methods will be pursued. This competition will seek concept studies that lead towards the development of a layered, multi-level, scalable performance engineering capability spanning applications, systems software, and hardware. The development of this capability requires advances in the following areas: 1. FUNCTIONAL SPECIFICATION METHODS. The ability to create performance modeling methods and models, and validate such, at all levels of applications, systems software, and hardware in a coherent fashion that will allow coupling and integration. 2. MODEL VALIDATION AND INTEGRATION. The ability to create and validate models at multiple levels of detail and to construct total system models involving multiple levels of resolution without significant loss of accuracy. 3. MEASUREMENT AND ABSTRACTION. The ability to measure really complex systems and to produce meaningful and useful abstractions from those measurements. 4. PERFORMANCE SPECIFICATION LANGUAGES. Definition of the needed interfaces and a performance specification language for enabling interoperability of tools at the different system levels and different resolution levels. Some applications by which this research may be motivated and driven include at least the following: A. Design and validation of the technologies developed under the QUORUM, Active Nets, and Evolutionary Design of Complex Software DARPA programs. B. Adaptive "on the fly" control of application execution in large-scale meta-computing systems. C. Performance evaluation of future systems whose performance meets the needs of the most demanding DoD applications. Additional information may be found in the report on the DARPA/NSF Workshop on Performance, September 19-20, 1996, which is expected to appear on the web in the near future. Contact the administrative address for details. AWARDS SCOPE: Proposed research should investigate innovative approaches and techniques that lead to or enable revolutionary advances in the state-of-the-art. Specifically excluded is research which primarily results in evolutionary improvement to the existing state of practice or focuses on a specific system or hardware solution. This competition seeks to foster multidisciplinary collaboration of researchers in languages, compilers, problem solving environments, computer architects and networking researchers who will develop fundamentally new approaches and will implement these ideas on prototype performance frameworks and novel tools, NOT mere enhancements of existing tools. It is expected that proposed methods and ideas will be well developed and substantiated by implementation of the performance frameworks. Several such multidisciplinary projects will be selected to be funded for three years, commencing in Fiscal Year 1997. The awardees are expected to present a status report at an appropriate forum in the November 1997 -- April 1998 time frame. Since we are seeking methods and tools that analyze system behavior as well as multilevel analysis, interaction among researchers in relevant sub-areas is considered essential. GENERAL INFORMATION: Proposers must submit an original and six copies of full proposals by 4:00 PM (ET), Tuesday, February 11, 1997, in order to be considered. Proposers must obtain a pamphlet, BAA 97-12 Proposer Information, which provides further information on the areas of interest, submission, evaluation, funding processes, and full proposal formats. This pamphlet may be obtained by fax, electronic mail, or mail request to the administrative contact address given below, as well as at URL address http://www.ito.darpa.mil/Solicitations.html. Proposals not meeting the format described in the pamphlet may not be reviewed. This Commerce Business Daily notice, in conjunction with the pamphlet BAA 97-12 Proposer Information, constitutes the total BAA. No additional information is available, nor will a formal RFP or other solicitation regarding this announcement be issued. Requests for same will be disregarded. The Government reserves the right to select for award all, some, or none of the proposals received. All responsible sources capable of satisfying the Government's needs may submit a proposal which shall be considered by DARPA. Historically Black Colleges and Universities (HBCU) and Minority Institutions (MI) are encouraged to submit proposals and join others in submitting proposals. However, no portion of this BAA will be set aside for HBCU and MI participation due to the impracticality of reserving discrete or severable areas of this research for exclusive competition among these entities. Evaluation of proposals will be accomplished through a scientific review of each proposal using the following criteria, which are listed in descending order of relative importance: (1) overall scientific and technical merit, (2) potential contribution and relevance to DARPA mission, (3) offeror's capabilities and related experience, (4) plans and capability to accomplish technology transition, and (5) cost realism. All administrative correspondence and questions on this solicitation, including requests for information on how to submit a proposal to this BAA, must be directed to one of the administrative addresses below by 4:00 PM, February 4, 1997, e-mail or fax is preferred. DARPA intends to use electronic mail and fax for some of the correspondence regarding BAA 97-12. Proposals may not be submitted by fax or email, any so sent will be disregarded. The administrative addresses for this BAA are: Fax: 703-522-7161 Addressed to: DARPA/ITO, BAA 97-12, Electronic Mail: baa9712@darpa.mil, Electronic File Retrieval: http://www.ito.darpa.mil/Solicitations.html, Mail:DARPA/ITO, ATTN: BAA 97-12, 3701 N. Fairfax Drive, Arlington, VA22203-1714. (0348)

Loren Data Corp. http://www.ld.com (SYN# 0001 19961218\A-0001.SOL)