Accelerated Modeling of Laplacian Diffusion via GPU
Budget: $250 – $750 USD
• Implement the one-dimensional model (metal rod) and demonstrate the numerical accuracy of your solution. Enter a specific partitioning of the rod (number of slices), a specific location along the rod, and a specific amount of time (steps). Run your model to output the exact temperature at the specified location at the specified time.
o # Slices = 2500
o Location = 0.7 m
o Time steps = 10,000,000
• Graph the change in temperature over time at a particular point along the rod in order to “observe” heat diffusion occurring.
• Conduct a performance evaluation of your CUDA code executing on a GPU vs. your sequential code on a CPU.
• Implement the two-dimensional model (baseboard heater).
o # Slices = 2500
o Location = 0.7 m
o Time steps = 10,000,000
• Graph the change in temperature over time at a particular point along the rod in order to “observe” heat diffusion occurring.
• Conduct a performance evaluation of your CUDA code executing on a GPU vs. your sequential code on a CPU.
• Implement the two-dimensional model (baseboard heater).