# LoopVectorization: Best way to have a multi and single threaded version?

**URL:** <https://discourse.julialang.org/t/loopvectorization-best-way-to-have-a-multi-and-single-threaded-version/98862>\
**Category:** General Usage\
**Tags:** question, package\
**Created:** [May 15, 2023, 6:51am UTC](https://discourse.julialang.org/t/loopvectorization-best-way-to-have-a-multi-and-single-threaded-version/98862 "2023-05-15T06:51:04Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![jw3126](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/jw3126/32/3086_2.png) [@jw3126](https://discourse.julialang.org/u/jw3126)\
**Post date:** [May 15, 2023, 6:51am UTC](https://discourse.julialang.org/t/loopvectorization-best-way-to-have-a-multi-and-single-threaded-version/98862/1 "2023-05-15T06:51:04Z")

</div>

Sometimes when I use LoopVectorization, I want to have both a single and multi-threaded variant of the function available. For instance:

```julia
using LoopVectorization

function add!(out, x, y; thread)
    if thread
        add_multi_thread!(out, x, y)
    else
        add_single_thread!(out, x, y)
    end
end

function add_single_thread!(out, x, y)
    @turbo thread=false for i in eachindex(out, x, y)
        out[i] = x[i] + y[i]
    end
    out
end

function add_multi_thread!(out, x, y)
    @turbo thread=true for i in eachindex(out, x, y)
        out[i] = x[i] + y[i]
    end
    out
end

x = randn(10)
y = randn(10)
out1 = randn(10)
out2 = randn(10)
add!(out1, x, y, thread=true)
add!(out2, x, y, thread=false)
@assert out1 ≈ out2

```

What is a good way to reduce code duplication here? Should I use`@eval`? What if I also want to control the number of threads at runtime, would I need to use `@generated`?