Request: @bprofile macro for profiling

[quinnj]

A common issue with profiling micro-kernels is the need to gather enough samples to have significant results. This usually involves function setup, execution repetition, etc. which it turns out, BenchmarkTools.jl pretty much already provides! I don't know all the internal details, but it would be awesome to have a @bprofile function that was the same as @btime, but ran @profile on the inner kernel for all the execution repetitions.

[KristofferC]

@btime also does GC scrubbing between samples which you probably don't want for this?

[quinnj]

Hmmm, maybe; it might be nice to make that configurable

[tkf]

It looks like this can be closed now that we have #254?