General optimisations for dict lookup/insert performance

Proposal:

There's room for tiny refactorings in the Objects/dictobject.c that may allow speedups, for example via better CPU cache behavior, compiler hints, cheap checks & fast exits for common cases. In my guess-timation, they may bring about upto 10% overall lookup/insert performance gains. That's more of an un-scientific intuition at best right now, but I'm laying some ideas out here:

• compare_unicode_unicode(): can benefit from pointer-equality and branch prediction, and reducing unnecessary function calls

• compare_generic(): can use early rejections on hash-mismatch, avoiding RichCompareBool calls

• unicodekeys_lookup_unicode(): pointer-equality for the common case can precede do_lookup which today is invoked everytime for all cases

• find_empty_slot(): early termination / separate path for first iteration can speed up insertions, especially for lightly-populated dicts

• build_indices_generic / build_indices_unicode (): breaking out the first-probe success case may speed up dict resize operations

I welcome feedback on how to raise this proposal differently, but otherwise I'd like to work on these ideas and quantify them. I feel modern CPUs have deep pipelines and these optimizations may help us avoid some cache-thrashing. These changes are purely performance oriented and don't change functionality / API / ABI and should be thread-safe too.

Has this already been discussed elsewhere?

This is a minor feature, which does not need previous discussion elsewhere

Links to previous discussion of this feature:

No response

Commencez dans Objects/dictobject.c et inspectez les fonctions nommées de recherche, de comparaison, de gestion des emplacements vides et de construction d’index. Quantifiez chaque optimisation proposée à l’aide de mesures de performance, tout en confirmant que les fonctionnalités du dictionnaire ainsi que le comportement de l’API, de l’ABI et de la sécurité des threads restent inchangés.