Changelog for Adaptive Performance Tweaks 1.21.1

Note

This change log includes the summarized changes. For the full changelog, please go to the GitHub History instead.

12.6.0

• Fixed item merge losing the remainder when a drop only partially fits a nearby stack.
• Fixed arrow cleanup removing thrown tridents and pickup-able player arrows; protected projectiles now use a separate hard cap (tridents removed last).
• Fixed game rules not being restored on shutdown, persisting optimized values (randomTickSpeed, doFireTick, ...) into level.dat and corrupting the baseline on restart.
• Fixed /aptweaks feature not applying enable/disable transitions (stale view/simulation distance and game rules).
• Fixed monitoringIntervalSeconds=0 silencing monitoring instead of logging on load changes.
• Fixed dimensions.ignore in spawn presets falling back to global limits instead of exempting it.
• Fixed server load rising-check comparing against a two-cycles-old measurement.
• Fixed login protection clearing invisibility/invulnerability set by game modes or other mods.
• Added entity tag support (#namespace:tag) in spawn preset allow/deny lists.
• Added protection for named/enchanted/NBT items in world/type limits with a 2x hard cap.
• Added wiring for minOptimizationLoadLevel in simulation distance and game rules.
• Added validation for core load thresholds and damage reduction percentages.
• Improved AI throttling to skip only goal/navigation/sensing, keeping physics, item pickup and despawning intact.
• Improved game-rule movement warmup to work without an active distance feature.
• Improved config reload to log mod conflict warnings only once.
• Improved entity tracking report to write asynchronously and mob-farm cleanup to use live entity positions.
• Improved benchmark heap metric: GC-collapsed measurements now report "n/a (GC)" instead of a misleading 0KB delta.
• Removed unused NeoForge-style mixin entries from the Forge mods.toml.

12.5.0

• Fixed benchmark report not covering all features and adjustments.
• Added Exploration scenario and update benchmark scenarios.
• Improved simulation distance features and set default to AUTO with better load-based adjustments.

12.4.0

• Added player warmup for simulation distance and random tick speed on login, teleport and fast movement.
• Added config toggles for login and movement warmup independent of minOptimizationLoadLevel.
• Improved load-based simulation distance changes with step-by-step reductions.
• Improved warmup recovery to only restore values below normal server load.
• Improved reset handling and tests for restoring server default values.

12.3.0

• Added more detailed MSPT buckets and improved reporting.
• Added smaller performance optimizations and removed some redundant code.
• Added additional benchmark scenarios for better performance analysis and optimization.
• Improved benchmark reporting and save result to file.
• Improved entity report and store results outside of config folder.
• Improved memory usage and reduced string concatenation in maps.

12.2.0

• Fixed #89 compatibility for many popular mods with special entities.
• Fixed #88 and #84 by providing NeoForge version and fixed related issues.
• Fixed #49 Ars Nouveau mob spawn handling and compatibility.
• Fixed benchmark crashes with chunk optimization mods.
• Fixed features are not completely disabled and downgrading benchmark phase 2 results.
• Fixed the incorrect Forge "missing on client" log message on clients.
• Added entity exclusion metadata (mod, category, reason) and entity tracking reports.
• Added more third-party compatibility checks.
• Added stuck arrow cleanup.
• Added spawn-type-based handling for structures, chunk generation, reinforcements...
• Added movement-aware simulation distance throttling.
• Added more spawn presets for known mods and entities.
• Added the missing "Eco Stack Manager" integration for the 12.x feature set.
• Added unit and game tests for the new features and optimizations.
• Improved benchmark measurement accuracy and result presentation.
• Improved logging, server load summaries, and stats output while reducing noisy output.

12.1.0

• Fixed issue with server load measurement per world was not accurate enough.
• Fixed smaller memory leak in the server load measurement code.
• Fixed log spam issue with auto-detected mods.
• Fixed Forge Mixin issue with missing mixin configuration file.
• Added auto-benchmarking tools for better performance analysis and optimization.
• Improved stats command to provide more detailed information about server load and performance.
• Improved Fabric performance by optimizing existing code to Fabric entity pipeline.

12.0.0 🚀

Next major alpha release with refactored code and improved performance. Now available for Fabric, Forge and NeoForge with a single codebase and better maintainability.