The Colossus Team Hub (CTH):
The primary power of Colossus is not the computer – yes, it is the computer’s name. It is the people. This site is meant to serve as a catalyst to bring a team of experts from around the world together in a space that supports free exchange, differing opinions, and mentoring in a place where the most important idea is team. A safe place for idea exchange and respectful discourse. And mentorship.
The project began as an idea to bridge the present into the near future so that teams from around the world, with the initial major groups being in India and the United States, could work toward porting Geant4 and the software project called TOPAS to work on field programmable gate array (FPGA) processors to speed up the calculations. Not just to work together now but to keep the team together into the future so new projects and tasks could be addressed via the power of the team. The Colossus Team.
The reason this is important to the NCI is because the next phase of radiation oncology depends on the ability to move from defining dose from the physical to the biological. This revolutionary step can only happen if we both do cutting edge biological research and create computing capacity to handle real-time clinical treatment planning with Monte Carlo (MC) methods that ultimately extend to include biological modeling.
TOPAS and the Geant4 project have some special features that make them a natural fit for this project. Firstly, Geant4 (and TOPAS) is written in 100% standard C++ and is fully multi-threaded. None of the other MC code is. Secondly, TOPAS is an NCI sponsored project so that working with the group is a natural extension of our support to date. NCI sponsored grants depend on peer review and rigor. Finally, the TOPAS-nBio extension to TOPAS that extends the human usability focus of TOPAS to radiobiology is a major capacity that will mesh well with the above primary goal of NCI – to develop the means to plan treatment biologically. Not just for radiation, ultimately for the entirety of the cancer intervention for the whole host.
FPGAs were chosen because of breakthrough work done at C-DAC and other centers on FPGAs that supports their capacity to calculate with very high precision, very rapidly, with very low power consumption. The FPGA is seen by some as the next step after the graphical processing unit (GPU) phase of high-performance computing (HPC) for this reason.
The real strength of this project is you, the user, the team member, the people, our team. One could write a novel about teamwork and achieving goals, but a proverb first told to me in Botswana (am not sure of its formal origin) says it well succinctly:
If you want to travel fast go alone, but if you want to travel far go together.
Our journey is broad and transcends any single country’s capacities; we cannot get there unless we work together. The CTH will strive to work to make teamwork as easy to achieve as the technology and rules allow. Constructive ideas are always welcome from everyone. We guarantee that we are not perfect – and we are very much aware that we don’t have all the answers.
To start with a start, the plan is to have Bruce Faddegon serve as the initial scientific rudder holder and set up teams for parts to do. Coding, testing, teaching, profiling, debugging, and more coding will be part of the job. Other things needed have yet to be defined but will come. Everyone on this team is as important as everyone else. And we’ll get their together. And the FPGAs are fast.