Faster Genetic Programming GPquick via multicore and Advanced Vector Extensions
This work provides a significant speedup for genetic programming applications, enabling more extensive evolutionary experiments in computational domains.
The researchers tackled the challenge of accelerating genetic programming for long-term evolution experiments by implementing a multicore and SIMD approach, achieving a performance of up to 139 billion GP operations per second on a single server.
We evolve floating point Sextic polynomial populations of genetic programming binary trees for up to a million generations. Programs with almost four hundred million instructions are created by crossover. To support unbounded Long-Term Evolution Experiment LTEE GP we use both SIMD parallel AVX 512 bit instructions and 48 threads to yield performance of up to 139 billion GP operations per second, 139 giga GPops, on a single Intel Xeon Gold 6126 2.60GHz server.