Abstract for Conference- already accepted
In recent times, there has been a considerable rise in numerical computer simulations in nearly every area of science and engineering. Many of these applications are very resource demanding and require powerful compute hardware to perform. As CPU clock speeds stall out around 3 GHz, many processor manufacturers have switched to multicore systems to increase performance. But even these styles have their limitations (power consumption, heat dissipation). By making use of GPU programming languages (OpenCL, CUDA), one can harness the power of hardware already contained in a desktop system, and obtain over an order of magnitude gains in performance, all things considered. In my work I will apply these aforementioned programming techniques to illustrate their power in a Gravitational Wave Source Modelling Application.
Adelaide said,
02/24/2010 at 6:01 pm
What area of technology or computation do you think would benefit most from increased efficiency and performance? Do you think you method could make a significant contribution to processing speeds in the future?
jmckennon said,
02/24/2010 at 6:10 pm
Really, any area of technology that relies on calculation based algorithms and any area of science (particularly numerical relativity and astronomy) that relies on mathematically intensive routines to obtain results would benefit from increased efficiency. Serialized approaches (aka, what the cpu does on its own) work very well for simply tasks but are extremely ineffective for large, multidimensional calculations and things of that nature. Do I think I can make a contribution to processing speeds? No. Processing speeds are determined by the layout of the silicon that the “processor” itself resides on. Billions of transistors switching in different ways form logic functions, which are the basis for the computers functionality. The only way to increase processing speeds at the moment is to add more transistors… this is bottlenecked though because the power consumption and heat dissipation increase much faster than the performance gain. This will cause silicon based processors to stall out around 3 GHZ. I do however, think that by implementing such gpu algorithmsinto the operating system… there is a massive amount of potential to increase performance
jmckennon said,
02/24/2010 at 6:03 pm
I get lost in it myself quite often too haha