What's Changed
- Upgrade
eguito 0.34.1 by @KonaeAkira in #332 - Add fast
FinishSolversuccess path for states with high enough CP by @KonaeAkira in #328 - Replay actions instead of storing current simulation state in search nodes by @KonaeAkira in #334
This significantly reduces memory usage for configurations with high node count which also means fewer crashes due to out-of-memory on the web version. - Add solver runtime and memory usage measurements via running
raphael-cliby @KonaeAkira in #335 - Use bump allocator for
QualityUbSolverandStepLbSolverby @KonaeAkira in #336 - Skip precompute for
StepLbSolverif it will never be queried by @KonaeAkira in #337 - Reduce size of candidate states in main search loop by @KonaeAkira in #339
- Update TW game data to patch 7.1 by @KonaeAkira in #340
Full Changelog: v0.27.0...v0.28.0