Profile-Guided Optimization (PGO) benchmark result

Hi! I do ongoing PGO research on different applications - all results are available at https://github.com/zamazan4ik/awesome-pgo . I performed some PGO benchmarks on the `ada-url` library and want to share my results here. ## Test environment * Fedora 39 * Linux kernel 6.7.6 * AMD Ryzen 9 5900x * 48 Gib RAM * SSD Samsung 980 Pro 2 Tib * Compilers - Rustc 1.76 and Clang 17.0.6 * Ada-url version: Rust bindings on `main` branch on commit `be5c26098f3cec1679dbb453d22542c9a7b10902` * Disabled Turbo boost for improving consistency across runs ## Benchmark The release benchmark is done with `taskset -c 0 cargo bench`, PGO training phase - with `CXX=clang++ CXXFLAGS="-fprofile-generate=pgo_profiles_clang" cargo pgo bench`, PGO-optimized results - with `export CXX=clang++ && export CXXFLAGS=ada.profdata && taskset -c 0 cargo pgo optimize bench`. `taskset -c 0` is used for better benchmark consistency. PGO profiles for Clang I got with `llvm-profdata` tool (more details could be found in the Clang documentation). In all tests, the Clang compiler is used. All PGO-related routines are done with [cargo-pgo](https://github.com/Kobzol/cargo-pgo). All benchmarks are done on the same machine, with the same hardware/software during runs, with the same background "noise" (as much as I can guarantee, of course). ## Results Here are the results: * Release: https://gist.github.com/zamazan4ik/445260656ae67dafce7c2c4e06f83154 * PGO-optimized: https://gist.github.com/zamazan4ik/5a82715d041e849d73dc9dfb1f814544 * (just for reference) PGO instrumented: https://gist.github.com/zamazan4ik/f5959d668063103f5a957af6b4ac206c For anyone wondering, the improvement comes from optimizing with PGO the Rust part or C++ part, I also performed the PGO test only for the C++ part. It's done via passing `-fprofile-use` flag via `CXXFLAGS` but running benchmarks with `cargo bench` (so no PGO-optimization for the Rust part). The results: https://gist.github.com/zamazan4ik/975036b1cd4ede6e4e6eeab2146934e1 . The benchmark confirms that C++ performance is improved with PGO. At least in the provided by project benchmarks, there are measurable improvements. Not sure if should I create the PGO performance report in the main `ada-url` repo or not - it's up to the maintainers :) Please do not treat the issue as a bug - it's just a performance report. If maintainers agree that building the library with PGO can be valuable for the users, maybe mentioning PGO building somewhere in the README will be a good idea to consider.