patch v8 float compute lose precision
github.com/dtboy1995/float-compute-patch
dtboy1995/float-compute-patch