If the results are non-deterministic, unless I'm happy with literally any possible result, then yeah, I'm going to be unhappy with all of the other results. That's what non-determinism means. What would being right always with non-deterministic results even look like in terms of code output?

▲

hansvm 10 hours ago | parent [-]

Compilers have a ton of hard optimization passes, and many of those are solved well enough via stochastic algorithms. Nobody actually fully optimizes register allocation or whatever, and a few random restarts can go a long way toward some better performance.

▲

saghm an hour ago | parent | next [-]

Every professional compiler I've ever used has given me the ability to choose what optimization level to use. No professional LLM I've ever used has given me any choice whatsoever to modify the temperature.

▲

leonidasrup 4 hours ago | parent | prev [-]

No compiler does optimal register allocation, because optimal register allocation is an NP-complete problem.

Chaitin et al.: "Register allocation via coloring".

https://en.wikipedia.org/wiki/Register_allocation

	▲	hansvm 2 hours ago \| parent [-]
		Yes? And stochastic methods are one way to approximate optimal NP solutions. In the context of register allocation, you'd be most likely to see that when modeling actual runtime optimality rather than minimizing spill or whatever. Mainstream compilers don't do that, but mainstream compilers don't incorporate many SOTA techniques.