Abstract
The operation of modern tokamaks is impossible without an effective system for controlling the plasma boundary during the discharge. In this paper, we consider a method for determining the plasma boundary on the basis of integral equations. A high-speed parallel code using GPUs was developed for this purpose. Different methods of parallelization of the algorithm have been considered. The possibility of processing the magnetic measurements in real time during experiments is shown. The effect of the technical parameters of GPUs (the number of cores, the data bus width, the amount of internal memory) on the performance is also shown. The simulation results for the T-15MD tokamak, currently under construction at National Research Center Kurchatov Institute, are presented.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
