CUDA-Accelerated High-Effeciency Code for Ion Tracing in Mass Spectrometry

Understanding the intricate coupling of multiple physical fields in mass spectrometry (MS) and the ion motion within them is crucial in instrument development. As an important numerical tool for researchers to comprehend MS performance, the scale and effectiveness of ion tracing simulations are now restricted by the limited computing power of hardware. Here we present a highefficiency multi-physical field-affected ion tracing software accelerated by the Graphical Processing Unit (GPU). The program is coded in a hybrid manner using C++ QT and CUDA/C based on the CPU-GPU heterogeneous computing scheme. The computation procedures, including fast iteration of the electric field, ion-neutral collision computing, space charge estimation, and data recording strategy are optimized for the parallel execution requirements. The program is tested in two benchmark cases (the ion funnel and the ion trap) and its performance is compared to the SIMION 8.1 program runs on the CPU. The results suggest that the running speed of our program is at least 20 times faster than the SIMION while precision can be maintained. It can be concluded that the introduction of GPU is promising in the numerical modeling of MS devices and may accelerate the iteration of the next-generation instrumentation design.