Swayam Singh, along with Nathan Goldbaum, has developed numpy_quaddtype, a cross-platform quadruple precision datatype for NumPy to address the long-standing issues with np.longdouble. This blog post discusses the necessity of high precision in scientific simulations, the technical aspects of numpy_quaddtype, its unique dual-backend approach using SLEEF and Long Double, precision levels, Dragon4 algorithm for accurate printing, ULP analysis, casting operations, universal functions, and applications like Mandelbrot set visualization and quantum harmonic oscillator modeling. With numpy_quaddtype, high-precision arithmetic in Python is now achievable without the maintenance headaches associated with np.longdouble.
https://labs.quansight.org/blog/numpy-quaddtype-blog