Ramanujan Machine v2 (Asymmetric-Degree CF Search)

Searches for polynomial continued fraction formulas with asymmetric degrees. Evaluates CF candidates and matches against known mathematical constants.

Usage

import torch
from kernels import get_kernel

kernel = get_kernel("cahlen/ramanujan-machine-cuda")
result = ramanujan.search(deg_a=1, deg_b=2, range_a=10, range_b=10)

Compile (standalone)

nvcc -O3 -arch=sm_90 -o ramanujan_machine ramanujan/ramanujan_v2.cu -lm

Results

All computation results are open:

Citation

@misc{humphreys2026bigcompute,
  author = {Humphreys, Cahlen},
  title = {bigcompute.science: GPU-Accelerated Computational Mathematics},
  year = {2026},
  url = {https://bigcompute.science}
}

Human-AI collaborative. Not peer-reviewed. All code and data open.

Downloads last month
-
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Dataset used to train cahlen/ramanujan-machine-cuda