Add module_roots

[timholy]

module_roots makes it easier to discover precompile-worthy
MethodInstances. The promise of @snoopi_deep is that it is not
restricted to just the entrance calls to inference; consequently,
we can find the most-expensive-to-infer calls in a specific module and
precompile those calls.

We may want to rewrite the entire parcel infrastructure around this, but perhaps one step at a time. We might want to create some kind of aggregate report that combines analysis of excessive specialization, inference breaks, and an improved set of precompile statements.

[timholy]

@NHDaly, if we were to generate some sort of "latency report and amelioration plan" for users & developers, here are the pieces I'm envisioning in very rough terms:

• for methods that get a huge number of specializations that lead to long compilation times, recommend that users try to figure out a way to streamline them through the approximate relationship compilation_time ∝ method_complexity * number_of_specializations, and that decreasing either factor can help. This is mostly what Add accumulate_by_method #152 was about.
• for methods that have a reasonably small number of specializations and where the types they're being called with are intrinsic to the package or its dependencies, generate precompile directives. That's mostly what this PR is about---with @snoopi_deep we can sometimes find something good to precompile even if it's not the entrance point to inference.
• for methods with big latencies that have relatively few specializations but which can't plausibly be precompiled in this package, recommend mechanisms to get them precompiled in downstream packages. This requires that methods be sufficiently type-stable that they can be inferred across package boundaries, as dump.c only stashes MethodInstances that either correspond to a method defined in the package or that have a backedge to the package. Identifying "breaks" in the chain of inference, where Julia uses runtime dispatch, will enable precompilation of bigger chunks of code. That's basically what Add inference_triggers #159 is about.

When you get a chance, I would love a review on this PR, and subsequently #159, so we can get the API moving towards completion.

[timholy]

I'm just going to start merging these, likely after taking another pass at reworking them.

[timholy]

I suspect this was superseded by #168, but I'll leave it here for a while at least.

[Sacha0]

Fantastic documentation additions! :)

[timholy]

But also almost certainly unnecessary! #168 has more powerful functionality. We need to add docs, but I am still feeling my way forward.

If you've not seen it, check out the new "visualizations.jl" and runtime_inferencetime. The aim is to get PGO, but statically rather than dynamically. There are advantages to either, but we can do the static one today...and it might actually be a better fit for the Julia mindset.

[timholy]

Yeah, this isn't really necessary after #168.