ASSESSING HARDWARE-DRIVEN VARIATIONS FROM WORKSTATIONS TO PERSONAL COMPUTERS IN GATE SIMULATION TIME FOR RADIOEMBOLIZATION STUDIES

Abstract

Monte Carlo GATE (Geant4 Application for Tomographic Emission) is widely used in medical physics for tomographic emission simulations, particularly in nuclear medicine procedures like radioembolization for liver cancer treatment. Despite its user-friendly interface, GATE simulations are often criticized for their slow computational speed, which poses challenges for students and researchers. Factors such as computer hardware (RAM, CPU, GPU, storage type), simulation settings, and the complexity of physics modeling significantly influence simulation times. This study investigates how hardware configurations impact GATE simulation performance by simulating a Yttrium-90 (Y-90) radioembolization procedure using a cylindrical phantom and tumor inserts. Simulations were conducted on three computers: an HP workstation, a DELL, and an HP Envy, with varying hardware specifications. Results revealed that the workstation, equipped with higher RAM, CPU, and GPU capabilities, demonstrated significantly faster simulation times compared to the personal computers. This highlights the critical role of advanced hardware in reducing computational time for GATE simulations. The study provides valuable insights for young researchers, emphasizing the importance of hardware optimization to achieve efficient and timely results in Monte Carlo-based medical physics research.