PyCLAW's fluid solver #6

Activity

1. shreyashkumar01 commented on Mar 15, 2025

   

   shreyashkumar01

   on Mar 15, 2025

   More actions

   Hello @aterrel

   Thank you for highlighting this use case. I'm interested in contributing to the development of more complex stencils for PyCLAW's 1D fluid solvers. After reviewing the Stegoton 1D application, I would like to take a shot at benchmarking the PyCLAW 1D fluid solvers mentioned in this issue. As a beginner, I see this as a very great opportunity for me to deepen my understanding of computational fluid dynamics and stencil operations beyond the standard Laplace equation.

   After reviewing, I would like to propose the following contribution with the below plan:

   1. Set up the stegoton1d example: I'll begin by cloning the PyCLAW repository and running the stegoton1d example provided in the
      link:
      https://github.com/clawpack/pyclaw/tree/master/apps/stegoton1d. This will give me a baseline understanding of the solvers' behavior.

   2. Develop benchmarking scripts: I'll create Python scripts using the time module or a dedicated benchmarking tool like timelit to measure the execution time of the solvers for different problem sizes and configurations.

   3. Implement various stencil operations: I'll explore different stencil implementations, potentially using NumPy or Numba for optimization, and compare their performance. This will help me to identify the most efficient methods for complex stencil
      operations.

   4. Compare against a known implementation: If possible, I'll compare the performance against a known, optimized implementation of a similar fluid solver (e.g., in Fortran or C) to assess the efficiency of the pure Python approach.

   5. Documentation and Reporting: I'll document my findings and provide the benchmarking code and results in a pull request for review. This will include detailed performance metrics and potential areas for further optimization.

   References:

   • PyCLAW Repository:
     https://github.com/clawpack/pyclaw
     (For cloning and running the example)
   • Python time module:
     https://docs.python.org/3/library/timeit.html
     (For basic timing)
   • pytest-benchmark:
     https://pytest-benchmark.readthedocs.io/en/stable/
     (For more advanced benchmarking)
   • NumPy:
     https://numpy.org/
     (For efficient array operations)
   • Numba:
     https://numba.pydata.org/
     (For potential JIT compilation of stencil operations)

   Let me know if you have any specific requirements or suggestions for this benchmarking effort. I am eager to receive your feedback and guidance to ensure my contribution aligns with the project's goals.

   Best regards,
   [Shreyash Kumar]