Add BufferedCustomFileEmission

[jackkoenig]

Uses virtual method .getBytesBuffered: Iterable[Array[Byte]] to optimize
file emission.

This is intended to enable faster annotation serialization. In particular this will be used to make .fir emission from Chisel support > 2 GiB .fir files. Currently. Chisel serializes the .fir to a String which has a max size of 2 GiB. We could use the CustomFileEmission API but it is 9-12x slower using just Iterable[Byte] so this PR is intended to make it much faster via buffering.

Contributor Checklist

• Did you add Scaladoc to every public function/method?
• [NA] Did you update the FIRRTL spec to include every new feature/behavior?
• Did you add at least one test demonstrating the PR?
• Did you delete any extraneous printlns/debugging code?
• Did you specify the type of improvement?
• Did you state the API impact?
• Did you specify the code generation impact?
• Did you request a desired merge strategy?
• Did you add text to be included in the Release Notes for this change?

Type of Improvement

• performance improvement
• new feature/API

API Impact

Expands the API with a new subclass of CustomFileEmission called BufferedCustomFileEmission

Backend Code Generation Impact

No impact

Desired Merge Strategy

• Squash

Release Notes

Introduce BufferedCustomFileEmission which enables Annotation writers to optimize serialization performance for their annotations.

Reviewer Checklist (only modified by reviewer)

• Did you add the appropriate labels?
• Did you mark the proper milestone (1.2.x, 1.3.0, 1.4.0) ?
• Did you review?
• Did you check whether all relevant Contributor checkboxes have been checked?
• Did you mark as Please Merge?

[jackkoenig]

I'm working on benchmarking serialization using this, will report back once I have numbers.

[seldridge]

Looks great!

[jackkoenig]

I'm working on benchmarking serialization using this, will report back once I have numbers.

The numbers look great, completely erases the 9-12x slowdown and actually for the specific case of emitting .fir from Chisel3, it appears to make emission faster over the Chisel 3 built-in emitter. See PR to follow in Chisel 3 using this new API.

[mwachs5]

Awesome!