This project simulates temperature distribution on a two-dimensional metal plate using a discrete model where the plate is represented as a grid. The edges of the plate are fixed at 0°C, while random heat sources are set at 100°C. The simulation iteratively calculates the temperature of each grid element based on its neighbors until the system reaches a stable state. The program is implemented in three modes—sequential, parallel, and distributed—to compare performance. A graphical interface visualizes the temperature gradient in real-time, with color coding from blue (cool) to red (hot), although the MPI-based implementation focuses solely on computation. Extensive testing is performed to analyze runtime scalability based on grid size and the number of heat sources.