Switch to truncated casting when converting floating point types to integer. This ensures that vectorized casts are consistent with scalar casts
2 files changed