Now, I have nice code for the 4d integral quaternionic and the 8d integral octonionic
Eisenstein Series. The fastest code is JAX/gpu, when it works, then openACC, nvc++/gpu,
While, over the cpu Numba/cpu and openACC/cpu … I am now coding with GPT/Claude .
Shortly, I shall provide the code with some of the math. Now, i hope with fast Riemann
Theta … I hope it shall open new fields for better understanding computational Riemann
Surface … I hope to fix the 8d JAX code into a batched powerfull JAX. So, with symmetry
reduction you can beat Amdahl’s law. Long Live the new king, symmetry reduction … What
is the correct forum for discussing these codes and improve them … if any one is interested in
nvidia about symmetry reduction.