function sumofsins1(n::Integer)  
    r = 0  
    for i in 1:n  
        r += sin(3.4)  
    end  
    return r  
end  

function sumofsins2(n::Integer)  
    r = 0.0  
    for i in 1:n  
        r += sin(3.4)  
    end  
    return r  
end

Timing the above two functions shows major differences in terms of time and memory allocations.

julia> @time [sumofsins1(100_000) for i in 1:100];
0.638923 seconds (30.12 M allocations: 463.094 MB, 10.22% gc time)

julia> @time [sumofsins2(100_000) for i in 1:100];
0.163931 seconds (13.60 k allocations: 611.350 KB)

This is because of type-unstable code in sumofsins1 where the type of r needs to be checked for every iteration.