[Wrf-users] WRF and MS Windows (Michalakes/NCAR)

[Pat Hayes]

Hello,
I read the item below and am wondering if any of you have a working WRF-MS Windows CCS 2003 system?
Thank you!Pat Hayes
PatHayes at msn.com
Patrick.Hayes at ngc.com
 
----------
 
http://www.microsoft.com/presspass/features/2006/aug06/08-01wccsrtm.mspxQ&A: Taking High-Performance Computing MainstreamCustomers report that Microsoft Windows Compute Cluster Server 2003 makes it easier and more cost effective to develop, deploy and use HPC systems, and they report that a Windows-based HPC platform integrates smoothly with existing IT resources.REDMOND, Wash., Aug. 1, 2006 -- Today Microsoft announced the general availability of Microsoft Windows Compute Cluster Server 2003, the company's first product designed specifically for high-performance computing (HPC). With Windows Compute Cluster Server 2003, Microsoft aims to make it easier to create, integrate and operate HPC clusters within organizations, thereby expanding the technology beyond traditional supercomputing centers by bringing the value of computational clusters within reach of more people.To understand the impact of today's milestone, PressPass convened a roundtable of customers who have been test driving Microsoft Windows Compute Cluster Server 2003 in demanding applications, including biomedical research and scientific modeling. Providing their insight are:John Michalakes, senior software engineer at the National Center for Atmospheric Research (NCAR) in Boulder, Colo.PressPass: Would each of you begin by briefly describing the work you're doing as it relates to Microsoft Windows Compute Cluster Server?Michalakes: About eight years ago, NCAR and a number of partner organizations involved in atmospheric research and operational forecasting began working on a next-generation community weather model and data assimilation system to eventually replace aging model codes in use for forecasting and research. This new model, called the Weather Research and Forecast (WRF) model, is basically all new software, designed from the outset for HPC systems. WRF is maintained and freely distributed as a community model and is being run at hundreds of institutions across the range of systems, from individual workstations to large supercomputers. Thus, portability and portable performance has been a key concern in the design and implementation of WRFPressPass: What made you decide to use the Microsoft Windows Compute Cluster Server 2003, and what benefits do you think it offers to your organization and your work?Michalakes: We strive to maintain WRF [NCAR's Weather Research and Forecast model] on as many systems deployed in our user community as possible. Until now, that meant systems running some flavor of UNIX or Linux. With the emergence of Microsoft Windows as a viable HPC operating system, and given that we receive on average one user request per month asking if WRF will work on Windows, we see Windows CCS as an opportunity for further broadening the range of computational resources available to the WRF user community.PressPass: What business needs are you solving with high-performance computing?Michalakes: Although the WRF model is used primarily in public sector institutions -- atmospheric research departments and government-run research and forecast centers -- a growing number of commercial weather forecast companies use WRF as well. These companies make specialized weather forecast products for customers in construction, agriculture, energy and other businesses.PressPass: As HPC becomes more of a mainstream technology, what key opportunities and challenges do you foresee for commercial and research application developers?
Michalakes: One shift we've seen as HPC has matured is from thinking about the performance-at-any-cost of HPC systems to thinking more about cost-performance of such systems. I believe operational numerical weather prediction is solidly terascale, but it remains to be seen whether operational centers will move to petascale systems for their day-to-day, real-time forecasting production schedules. The issue will be cost -- the cost of petascale systems themselves, as well as the cost to operate such systems and the cost to retool and maintain modeling software to run on such systems, weighed against some hoped-for improvement in forecast quality. In the near term, I believe petascale computing will be used for non-real-time, very-high-resolution simulations for research to improve understanding of atmospheric processes that will, in turn, provide improvements to lower-resolution operational real-time prediction runs.Having said all this about petascale computing, my sense in the context of this discussion is that Microsoft is not currently targeting frontier computing systems for Windows Compute Cluster Server, but focusing instead on small- to mid-level clusters more widely deployed in the research and commercial areas of weather modeling. From this perspective, the challenges and opportunities are efficient integration and management of computing and data systems to allow for more seamless coordination and management of workflows for end-to-end computing and analysis of weather and climate applications. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20061016/730528ba/attachment.html