|
|
|
SCD has spent much time and energy to develop 11 major themes or initiatives around which we will plan for our future evolution as a major provider of computing resources to our research communities. More than just an exercise in planning, SCD has sought to include a willed future for each of these initiatives encompassing all aspects of our work. Our focus is oriented outward, centering on the needs of our constituents, as illustrated in this graphic:
Each of the 11 roadmaps has been carefully circumscribed and informed by input from scientific community needs, NSF Review Panels, SCD Advisory Panels, the NCAR Strategic Plan for High-Performance Simulation, market research, technology trends, community efforts, and potential partnerships. The division's roadmaps furnish both an overview and an outline of where we plan to invest our time and resources to advance scientific computing for the atmospheric and related sciences. The current versions of our roadmaps reside on the web at http://www.scd.ucar.edu/dig/atlas/
Early in March 2000, SCD began to formulate plans, in accordance with our overall timetable for substantially augmenting scientific computing at NCAR, to acquire the next generation of high-performance computing equipment to support NCAR's research agenda. SCD had already laid out an aggressive plan to expand the current computing capabilities afforded by the recently acquired IBM RS/6000 SP supercomputer (blackforest). This included upgrades to the hardware of that system, increasing its capacity to 151 nodes (604 CPUs), upgrades to the Compaq ES40 experimental system (named prospect), and a considerable increase to the capacity our Mass Storage System with the addition of three new StorageTek robotic tape handlers (for a total of five).In FY1999, SCD had planned for an additional increment in computing in late FY2000, such that the overall computing capability would, in aggregate, reach one teraflop. Accordingly, SCD established a working group of technical experts -- from SCD and including scientists from each of NCAR's science divisions -- to draft an official Request for Proposal (RFP) to install and periodically upgrade a high-performance Advanced Research Computing System (ARCS) at NCAR.
As stated in the RFP, the ARCS must provide a highly productive computational environment for the development and execution of complex, long-running, computationally intensive earth system models. SCD's paramount objective is to facilitate advances across a broad spectrum in geosciences. We seek the highest level of computing capacity and capability to address this central objective.
We will release the RFP in draft form for vendor review in late October, and the final RFP will be released in late November 2000.
SCD ASR - Table of contents