Showing 1–1 of 1 results for author: Martinez, H E
-
SClib, a hack for straightforward embedded C functions in Python
Authors:
Esteban Fuentes,
Hector E. Martinez
Abstract:
We present SClib, a simple hack that allows easy and straightforward evaluation of C functions within Python code, boosting flexibility for better trade-off between computation power and feature availability, such as visualization and existing computation routines in SciPy. We also present two cases were SClib has been used. In the first set of applications we use SClib to write a port to Python o…
▽ More
We present SClib, a simple hack that allows easy and straightforward evaluation of C functions within Python code, boosting flexibility for better trade-off between computation power and feature availability, such as visualization and existing computation routines in SciPy. We also present two cases were SClib has been used. In the first set of applications we use SClib to write a port to Python of a Schrödinger equation solver that has been extensively used the literature, the resulting script presents a speed-up of about 150x with respect to the original one. A review of the situations where the speeded-up script has been used is presented. We also describe the solution to the related problem of solving a set of coupled Schrödinger-like equations where SClib is used to implement the speed-critical parts of the code. We argue that when using SClib within IPython we can use NumPy and Matplotlib for the manipulation and visualization of the solutions in an interactive environment with no performance compromise. The second case is an engineering application. We use SClib to evaluate the control and system derivatives in a feedback control loop for electrical motors. With this and the integration routines available in SciPy, we can run simulations of the control loop a la Simulink. The use of C code not only boosts the speed of the simulations, but also enables to test the exact same code that we use in the test rig to get experimental results. Again, integration with IPython gives us the flexibility to analyze and visualize the data.
△ Less
Submitted 19 December, 2014;
originally announced December 2014.
