Experiment with alternative values of `GOGC`

Today we run with a default of [`GOGC=100`](https://go.dev/doc/gc-guide#GOGC).

We should experiment with alternative values and see how it affects overall compile times, language server response times, and peak memory usage.

Start by locating where the TypeScript repository uses the default GOGC=100 and identify how compile times, language server response times, and peak memory usage are measured. Experiment with alternative GOGC values and compare each result with the default across all three metrics; done means the trade-offs are documented clearly.

Written by the indexing model from the issue text.