Evaluate using Profile-Guided Optimization (PGO) and Post-Link Optimization (PLO)

[zamazan4ik]

Hi!

Let's try to apply even more optimizations after enabling LTO ;) I checked Profile-Guided Optimization (PGO) CPU performance improvements on multiple projects (including many databases like PostgreSQL and network proxies like Envoy, HAProxy, Nginx, etc.) - the results are available in the awesome-pgo repo. Since the pgdog project cares about its performance - it could an interesting opportunity to use PGO.

I can suggest the following things to do:

• We need to perform PGO benchmarks on pgdog. If it shows improvements - add a note about possible improvements to the documentation. Providing an easier way (e.g. a build option) to build scripts to build pgdog with PGO can be useful for the end-users too.
• It will be a good idea to integrate building with PGO into your CI pipelines. In this case, users will get already PGO-optimized binaries from the website (when will they be available for the users, of course).

Here you can find different materials about PGO: many benchmarks for different apps, examples of how PGO is already integrated with various projects, and some PGO-related pieces of advice. For Rust-based applications, I recommend starting with cargo-pgo.

After PGO, I suggest evaluating the Post-Link Optimization (with tools like LLVM BOLT) - it can give more aggressive optimizations even after PGO. However, starting with regular PGO will be easier to do.

Thank you.

[levkk]

Worth a shot. The flamegraph currently is reasonably clean (I'm using samply), but I'm definitely open to some "free" optimizations from the linker.