Many of the limitations imposed upon the users of high performance computing are about to be lifted, allowing even the most complicated system to be modelled in high fidelity by a scientist or engineer. That's according to Microsoft's Server and Tools Business president Bob Muglia, who recently introduced the company's Technical Computing Initiative with an open email to the HPC community. According to Muglia, the Technical Computing Initiative will be 'a new effort focused specifically on empowering millions of the world's smartest problem solvers.' This empowerment will be achieved through developments in three core areas: cloud computing, parallel programming, and technical computing tools such as databases and spreadsheets.
'Our goal is to unleash the power of pervasive, accurate, real-time modelling to help people and organisations achieve their objectives and realise their potential,' wrote Muglia in his introductory statement. 'We are bringing together some of the brightest minds in the technical computing community across industry, academia and science at www.modelingtheworld.com to discuss trends, challenges and shared opportunities.'
Bill Hilf, the general manager of platform strategy at Microsoft, spoke to Scientific Computing World about the initiative, and filled in some of the blanks: 'Over the past few years we've built out the HPC server business and we've been learning more and more about the broad technical community - the scientists, engineers, researchers, etc. As we learned more about this technical community, we realised that there is a lot more that we could be doing, from a business perspective, to build software within the spectrum of the scientific process. About 18 months ago we started a team internally, and pulled the HPC Server team into it, along with people working on a variety of things related to parallelism, and a large group of people working on things related to the cloud.'
What will the Technical Computing Initiative offer that is not already available? 'Collectively, it will be a set of products that range from the desktop up to the cloud, including the data centre and the cluster,' said Hilf. 'What we have on the market today are things like Windows HPC Server, and the Visual Studio tools for parallelism, but that's really just the beginning of what we're building out over time. So far, the industry has been using HPC and the idea of distributed computing as a method to break up a problem - solving it by distributing it across multiple resources. That same method is becoming critical on multicore systems via the necessity to build parallelised software that can leverage a multicore system. That same method is also becoming necessary when working on the cloud.'
Hilf stated that the Technical Computing Initiative will seek to generalise the approach of distributed computing to all types of platform and all types of HPC: 'We're taking the [distributed computing] method and approaching it holistically across the desktop, the cluster and the cloud. You'll be able to write an application or build a model that would seamlessly move across those different form factors, taking full advantage of the capability beneath each of them.' He went on to describe Microsoft's cloud services offering, Azure, which is optimised for HPC users: 'With Azure, we have hundreds of thousands of servers distributed around the world. The system is designed to allow for large scale, very extreme data sizes,' he said.
Muglia's introductory statement elegantly spelled out its goals in a general sense: 'One day soon, complicated tasks like building a sophisticated computer model that would typically take a team of advanced software programmers months to build and days to run, will be accomplished in a single afternoon by a scientist, engineer or analyst working at the PC on their desktop,' he wrote, adding that these models will become more complete and accurate in the way they represent the world. For scientists and engineers, this will certainly be an appealing statement, but only time will attest to its validity.
