Parallel binomial American option pricing on CPU-GPU hybrid platform

We present a novel parallel binomial algorithm to compute prices of American options. The algorithm partitions a binomial tree into blocks of multiple levels of nodes, and assigns each such block to multiple processors. Each processor in parallel with the others computes the option’s values at the assigned nodes. The algorithm is implemented and tested on a heterogeneous system consisting of an Intel multi-core processor and a NVIDIA GPU. The whole task is split and divided over the CPU and GPU so that the computations are performed on the two processors simultaneously. In the hybrid processing, the GPU is always assigned the last part of a block, and makes use of a couple of buffers in the onchip shared memory to reduce the number of accesses to the off-chip device memory. The performance of the hybrid processing is compared with an optimised CPU serial code, a CPU parallel implementation and a GPU standalone program.