Improvements to EFIT in Preparation for the Burning Plasma Era*

EFIT was the first and is the most extensively used equilibrium reconstruction code in the world. Although robust, the burning plasma regime will bring new challenges which include adapting to novel operating regimes and incorporating diagnostics that can withstand a harsh, radioactive environment. This regime has motivated the exploration of machine learning techniques to improve the quality of equilibrium reconstructions that can be produced in real-time. To support this development, we are upgrading the core Grad-Shafranov solver. The improvements include clearly separating out the device-specific coding, enhancing code portability, and ensuring thread-safety in preparation for GPU developments. To aid in the development, we have created a test suite for use with continuous integration workflows.