Parallel implementation of multiple-precision arithmetic and 2,576,980,370,000 decimal digits of π calculation

We present efficient parallel algorithms for multiple-precision arithmetic operations of more than several million decimal digits on distributed-memory parallel computers. A parallel implementation of floating-point real FFT-based multiplication is used, since the key operation for fast multiple-precision arithmetic is multiplication. The operation for releasing propagated carries and borrows in multiple-precision addition, subtraction and multiplication was also parallelized. More than 2.576 trillion decimal digits of π were computed on 640 nodes of Appro Xtreme-X3 (648 nodes, 147.2 GFlops/node, 95.4 TFlops peak performance) with a computing elapsed time of 73 h 36 min which includes the time required for verification