Feature request: Use multithreading for simple operations on larger dfs

[Chuck321123]

Description

So polars seems to be equally as fast as pandas on simple operations on larger dfs. Is there nothing that can be done to make the operations faster?

import numpy as np
import polars as pl

# Define the number of rows
n_rows = 200_000_000

df = pl.DataFrame({
    "col1": np.random.rand(n_rows)
})

%timeit -r 1 -n 7 df.select(pl.col("col1")+1)

df=df.to_pandas()
%timeit -r 1 -n 7 df["col1"] + 1

504 ms ± 0 ns per loop (mean ± std. dev. of 1 run, 7 loops each)
484 ms ± 0 ns per loop (mean ± std. dev. of 1 run, 7 loops each)