# current way to benchmark
/BenchmarksGame.jl$ time julia nbody/nbody-fast.jl 50000000
-0.169075164
-0.169059907
real 0m2.639s
user 0m2.763s
sys 0m0.365s
#when compiled
/BenchmarksGame.jl$ time ./build/nbody-fast 50000000
-0.169075164
-0.169059907
real 0m2.484s
user 0m2.563s
sys 0m0.398s
# Reference when calling 2 times (first gets called on include)
julia> push!(ARGS, "100")
julia> include("nbody/nbody-fast.jl")
julia> @time NBody.perf_nbody(50000000)
-0.169075164
-0.169059907
2.334110 seconds (28 allocations: 976 bytes)
Shaves off a few ms from the benchmarks:
Need to discuss how to properly integrate this!