FS3 Project: Website & Project Team
Several recent online articles have highlighted the need to quickly move ahead on the FireSeed Streaming Supercomputer (FS3) and to focus on two specific project areas.
The FS3 project will design and build a GPU-cluster streaming supercomputer for less than $100,000 whose performance will place it in the Top 500 Supercomputers list.
The window of prime opportunity for the FS3 project in terms of a burning need for small and medium businesses was highlighted by a Business Week article, "Supercomputing Comes to the Rust Belt." Business Week covers several real life examples of small companies using supercomputing resources to compete in the global knowledge economy. One small company does fluid dynamics simulations for the transportation industry, while another one does laser scanning and digitization of large spaces, such as buildings. By definition, the knowledge economy will bring with it more and more opportunities related to manipulating data to manage and improve our knowledge. Therefore it makes sense that low-cost supercomputing resources are becoming increasingly a needed resource for new or innovative companies.
Numerous other articles have talked about the computing industry's desperate need for an effective programming model for the multi and many-core processors now being designed and produced. Intel, IBM, AMD and other CPU manufacturers are no longer developing top of the line processors with only a single core. Moore's law ran into a physics brick wall, and the computing hardware industry has shifted over to multi-core and many-core processors for both CPUs and GPUs. The problem is, the computing software industry has not been able to shift to multi-core or many-core processing. When you're using your brand new top of the line laptop with a dual core processor, you're running single-core programs and not taking advantage of the dual core design. And the NVIDIA 8800 GTX video card has 128 processing cores on it... The FS3 project has a huge opportunity to design and develop a programming paradigm for many-core processing. With the right project team members and a few innovation breakthroughs, the FS3 programming model could define many-core processing for the future. Someone has to do it, and the large IT corporations have already proven that just throwing money at the problem isn't the solution.
It's clear there is a window of opportunity here for the FS3 project. However, many other people and companies realize the window is open, so the FS3 project needs to move forward quickly.
Two specific areas where FS3 needs to move forward quickest are:
- FS3 project website
- FS3 project team
The FS3 project team is currently a small group of people who feel passionately enough about the project to launch the project and commit a significant amount of their time to it. To make the project succeed, the current team must identify additional core project members who are needed for various parts of the project. Once these core project members are identified, they will be invited to join the project.
These core members will cover many aspects including hardware, software, journalism, intellectual property/legal, marketing, and grant writing. If you are intrigued by the concept of the FS3 project and have skills in one of these areas or another area you feel is useful to the project, please contact me at bwaldron [att] gmail dott com.
The next month or two will be spent taking care of these project basic needs. After those needs are addressed, we'll really start having some fun!