3 ms·
I like the idea of exhaustive search, which the simplicity of the 6502 seems ideally suited for, but the search speed seems a bit limiting. I wonder if there's
by HarHarVeryFunny 7mo ago
I like the idea of exhaustive search, which the simplicity of the 6502 seems ideally suited for, but the search speed seems a bit limiting. I wonder if there's not potential for more generation restriction (e.g. code can only use a specific N bytes of zero page) and heavy search pruning to speed it up? If it could generate optimal 20-30 op sequences in semi-reasonable time that'd make it very useful.
- Aransentin 7mo agoHaving it only use operations that use a specific set of zero-page addresses is already supported, yep! 20-30 ops is probably impossible, unfortunately. The combinatorial explosion is just too enormous.
- peterfirefly 7mo agoe-graphs have been repeatedly reinvented across decades for many purposes. One of them is superoptimization. https://en.wikipedia.org/wiki/E-graph https://en.wikipedia.org/wiki/E-graph https://www.cs.cornell.edu/courses/cs6120/2025sp/blog/superopt/ https://www.cs.cornell.edu/courses/cs6120/2025sp/blog/supero... https://github.com/philzook58/awesome-egraphs https://github.com/philzook58/awesome-egraphs