Investigate ClangBuildAnalyzer to identify areas where we can improve build performance

[azrogers]

While benchmarking #1064, I found out about ClangBuildAnalyzer. Using this tool in conjunction with clang's -ftime-trace flag, we can generate a report of which parts of our codebase take the longest to build. The granularity is pretty impressive, allowing us to break down which files take longest to parse, which templates take the longest to generate, which functions take the longest to compile, and more. It would be nice to have this as an additional compile target we can use, like clang-tidy.