This is to keep track of the various TODO about perf optimizations in the Tracing V2 codebase. Once the code is functional and upstreamed we should look into these:

1) use a templated version of the hotpath WriteBytes() method of ScatteredStreamWriter. memcpy() becomes super optimized if size is known at compile time.

2) Look into optimize_max GN config (which IIRC makes the code -O2)