# Sepals<\/p>\n
An extremely radical and experimental optimization mod for Minecraft server performances.<\/p>\n
We recommend using Sepals with Lithium and Async for optimal performance.<\/p>\n
## Foreword<\/p>\n
First,
\nthe Sepals is a collection of highly experimental server-side optimizations that target very specific performance bottlenecks in Minecraft\u2019s AI, entity processing, and task systems. Instead of providing broad \u201cgeneral\u201d optimizations like Lithium, Sepals focuses on deep rewrites of a few extremely expensive vanilla mechanisms \u2014 especially villager brains, frog behavior, entity cramming, target filtering, and nearby-entity sensing.<\/p>\n
### So what does it do in practice?<\/p>\n
If you run large farms, villager halls, dense mob enclosures,
\nor any scenario with hundreds\/thousands of entities in a small area,
\nSepals drastically lowers tick time by avoiding Vanilla\u2019s most expensive AI operations.<\/p>\n
If you\u2019re running a normal SMP with small farms, you probably won\u2019t notice much of a difference.<\/p>\n
It is experimental, and not every optimization is guaranteed to have Vanilla parity,
\nbut in performance-stress tests, the gains can be huge.<\/p>\n
## Compatibility<\/p>\n
Currently, Sepals is compatible with nearly all mods.<\/p>\n
Here is the verified mod list with the latest Sepals version:<\/p>\n
| Target mod | Required target version |
\n|——————————————————:|:———————–:|
\n| Sodium | all |
\n| Iris | all |
\n| FerriteCore | all |
\n| Krypton | all |
\n| Lithium | >=0.18.0 |
\n| C2ME | >=0.3.4.0.0 |
\n| Moonrise | >=0.6.0-beta.1+45edfd7 |
\n| Async | >=0.1.7+alpha.7-1.21.8 | <\/p>\n
If you don’t know which version to choose, please use the latest version of optimization mods, as Sepals will
\nensure compatibility with them.<\/p>\n
### Important<\/p>\n
Sepals won’t support any old versions and snapshot versions of Minecraft; all bug fixes and features are only available in the newest version.<\/p>\n
When using it with Async,
\nplease manually compare performance with Sepals and Async, and edit their configurations, to find which optimizations are worthwhile,
\nas each machine has different results.<\/p>\n
Sepals feature related: “`enableSepalsEntitiesCramming“`.<\/p>\n
## Config<\/p>\n
| Config name | Allowed value | Default value |
\n|:——————————————-:|:————-:|:————-:|
\n| forceEnableSepalsPoi | bool value | false |
\n| enableSepalsVillager | bool value | true |
\n| enableSepalsFrogLookAt | bool value | true |
\n| enableSepalsFrogAttackableSensor | bool value | true |
\n| enableSepalsLivingTargetCache | bool value | true |
\n| nearestLivingEntitiesSensorUseQuickSort | bool value | true |
\n| enableSepalsBiasedLongJumpTask | bool value | true |
\n| enableSepalsEntitiesCramming | bool value | true |
\n| enableSepalsItemMerge | bool value | true |
\n| enableSepalsQuickCanBePushByEntityPredicate | bool value | true |<\/p>\n
## Performance<\/p>\n
A test was done in server of Mars provided by feimia, CPU: Intel i7-14700K; Game memory: 4G; OS: Ubuntu 24.04.1 LTS;
\nMinecraft: 1.21.<\/p>\n
### Entities cramming<\/p>\n
“`
\nUse box cache to prevent too much ‘getOtherEntities’ calls<\/p>\n
— Notice —
\nThis feature ignores scoreboard\/team predicates for every single entity
\nand may cause unexpected behavior.<\/p>\n
This problem will not affect anything in Vanilla, except when using
\ncommand blocks.<\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
1390 villagers cramming in a 7×7 space:<\/p>\n
| Environment | tickCramming | Percent (Avg.) |
\n|——————–:|:————:|:————-:|
\n| Vanilla | 53.6 ms | 100 % |
\n| Lithium | 54.4 ms | 101 % |
\n| Sepals | 10.2 ms | 19 % |
\n| Sepals + Lithium | 8.5 ms | 15 % |<\/p>\n
### Weighted random<\/p>\n
“`
\nUse binary search to replace vanilla weight random<\/p>\n
— Warning —
\nThis feature may not be worthwhile,
\nas Spark proved that it is slower than Vanilla when constructing range tables,
\neven if the binary search is close to 0 ms.<\/p>\n
— Status —
\nDisabled by default<\/p>\n
— Warning —
\nNot tested.
\n“`<\/p>\n
### Biased long jump task<\/p>\n
“`
\nUse Sepals long jump task implementation to replace Vanilla<\/p>\n
As mentioned above, the binary search is almost no costs
\nsepals impletation will construct the range table at the same time as generating targets
\nand used Catheter to replaced java stream<\/p>\n
The same time, rearrange should jump conditions in frog brain, make a good performance<\/p>\n
— Notice —
\nThis feature is unable to change in game runtime
\nrequired restart the server to apply changes<\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
800 frogs cramming in a 7×7 space:<\/p>\n
| Environment | keepRunning | Percent(Avg.) |
\n|:———————————————–:|:———–:|:————-:|
\n| Vanilla
(LongJumpTask) | 43.1 ms | 100 % |
\n| Lithium
(LongJumpTask) | 7.5 ms | 17 % |
\n| Sepals
(SepalsLongJumpTask) | 0.2 ms | 0.4 % |
\n| Sepals + Lithium
(SepalsLongJumpTask) | 0.05 ms | 0.1 % |<\/p>\n
| Environment | getTarget | Percent(Avg.) | Percent(in “`keepRunning“`) |
\n|:———————————————–:|:———:|:————-:|:—————————–:|
\n| Vanilla
(LongJumpTask) | 43.1 ms | 100 % | 100 % |
\n| Lithium
(LongJumpTask) | 3.6 ms | 9 % | 48 % |
\n| Sepals
(SepalsLongJumpTask) | N\/A ms | 0 % | 0 % |
\n| Sepals + Lithium
(SepalsLongJumpTask) | N\/A ms | 0 % | 0 % |<\/p>\n
### Quick sort in NearestLivingEntitiesSensor<\/p>\n
“`
\nUse quick sorting from FastUtil to replace Java TIM’s sorting<\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
800 frogs cramming in a 7×7 space:<\/p>\n
| Environment | sort (NearestLivingEntitiesSensor#sense) | Percent(Avg.) |
\n|——————–:|:—————————————-:|:————-:|
\n| Vanilla | 3.8 ms | 100 % |
\n| Lithium | 3.6 ms | 94 % |
\n| Sepals | 2.2 ms | 57 % |
\n| Sepals + Lithium | 2.2 ms | 57 % |<\/p>\n
### Frog attackable target filter<\/p>\n
“`
\nRearrange the attackable conditions
\nto have the least costly conditions first,
\nand therefore reduce the probability of high-costs calculating<\/p>\n
— The complaints —
\nMojang’s attackable predicate is:<\/p>\n
!entity.getBrain().hasMemoryModule(MemoryModuleType.HAS_HUNTING_COOLDOWN)
\n && Sensor.testAttackableTargetPredicate(entity, target)
\n && FrogEntity.isValidFrogFood(target)
\n && !this.isTargetUnreachable(entity, target)
\n && target.isInRange(entity, 10.0)<\/p>\n
in this case, ‘Sensor#testAttackableTargetPredicate’ calls ‘TargetPredicate#test’
\nwhich causes lots of raycast calculating when there are too many entities in an area,
\nmaking it even worse, considering Minecraft’s raycast is painfully slow<\/p>\n
in this case, ‘TargetPredicate#test’ (800 frogs) costed 9.8ms per game tick,
\nand ‘BlockView.raycast’ contributed 7.3ms<\/p>\n
therefore, I changed it to:<\/p>\n
FrogEntity.isValidFrogFood(target) &&
\n entity.getBrain().hasMemoryModule(MemoryModuleType.HAS_HUNTING_COOLDOWN) &&
\n target.isInRange(entity, 10.0) &&
\n Sensor.testAttackableTargetPredicate(entity, target) &&
\n isTargetUnreachable(entity, target);<\/p>\n
the ‘isValidFrogFood’ is a simple condition, it checks the entity’s ‘frog_food’ tag
\nand an extra check when entity is slime then skip it when it size not 1<\/p>\n
‘isInRange’ and ‘hasMemoryModule’ are also fine, as they only calculate some simple things<\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
800 frogs cramming in a 7×7 space:<\/p>\n
| Environment | time | Percent(Avg.) |
\n|————————————————-:|:——:|:————-:|
\n| Vanilla (FrogAttackablesSensor#matches) | 10 ms | 100 % |
\n| Lithium (FrogAttackablesSensor#matches) | 5.7 ms | 57 % |
\n| Sepals (SepalsFrogBrain#attackable) | 0.1 ms | 1 % |
\n| Sepals + Lithium (SepalsFrogBrain#attackable) | 0.1 ms | 1 % |<\/p>\n
### Frog look-at target filter<\/p>\n
“`
\nUse ‘SepalsLivingTargetCache’ to improves target search performance
\nYou must have that option enabled, otherwise this is identical to Vanilla<\/p>\n
— Notice —
\nThe raycast is in TargetPredicate test
\nat the ‘findFirst’ in LivingTargetCache when input predicate is success<\/p>\n
but if subsequent conditions are failures, it is useless to calculate this further
\nbecause even if the findFirst has found (raycast success)
\nbut we don’t used this result in subsequent contexts <\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
800 frogs cramming in a 7×7 space:<\/p>\n
| Environment | findFirst | Percent |
\n|————————————————————————————-:|:———:|:——-:|
\n| Vanilla
(LookAtMobWithIntervalTask$$Lambda#findFirst) | 2.7 ms | 100 % |
\n| Lithium
(LookAtMobWithIntervalTask$$Lambda#findFirst) | 2.5 ms | 92 % |
\n| Sepals
(SepalsLookAtMobWithIntervalTask$$Lambda#findFirstPlayer) | 0.1 ms | 3 % |
\n| Sepals + Lithium
(SepalsLookAtMobWithIntervalTask$$Lambda#findFirstPlayer) | 0.1 ms | 3 % |<\/p>\n
### Villager miscellaneous optimizations<\/p>\n
“`
\nThis is what Sepals does:<\/p>\n
1. Use Catheter to replaced Java’s Stream, since it has much better performance and scalability compared to Stream
\n2. Cache tasks, activities, running tasks and memory to improve starting and updating task times
\n3. Use sepals composite task to replaced vanilla composite task
\n4. Whenever possible, find opportunities to skip more raycasts and useless predicates
\n5. Use ‘SepalsLivingTargetCache’ to replaced vanilla cache, At the cost in sensors tick make less cost in finding interaction target or look at mob task
\n6. Rearranged predicates and extra lower cost predicate, The purpose of this is do higher cost predicate later or best don’t do that, skip the remaining high cost predicates in advance
\n7. Copied and modified ‘SerializingRegionBasedStorage’ optimizations from lithium
\n8. With more targeted task, don’t use the generics to reduce useless operations
\n9. Use binary search list to replace hashset search<\/p>\n
— Notice —
\nIt is recommended to use this with Lithium and C2ME for optimal performance<\/p>\n
— Warning —
\nNot long-term stability tested, only a month running shown it’s ok currently
\nthis feature has not been proved to be identical to Vanilla,
\nbut also hasn’t appeared to be statistically different<\/p>\n
— Status —
\nEnabled by default
\n“`<\/p>\n
800 villagers cramming in a 7×7 space at noon:<\/p>\n
| Environment | Brain#tick (Total) | Percent | Brain#startTasks | Percent(startTasks) | Brain#tickSensors | Percent(tickSensors) | Brain#updateTasks | Percent(updateTasks) | Brain#tickMemories | Percent(tickMemories) |
\n|:———————–:|:——————:|:——-:|:—————-:|:——————-:|:—————–:|:——————–:|:—————–:|:——————–:|:——————:|:———————:|
\n| Vanilla | 18 ms | 100 % | 9.3 ms | 100 % | 5.2 ms | 100 % | 3 ms | 100 % | 0.5 ms | 100 % |
\n| Lithium | 12.4 ms | 68 % | 4.8 ms | 51 % | 5.9 ms | 113 % | 1.2 ms | 40 % | 0.5 ms | 100 % |
\n| Sepals | 9.7 ms | 53 % | 3.6 ms | 38 % | 3.7 ms | 71 % | 2 ms | 66 % | 0.4 ms | 80 % |
\n| Sepals + Lithium | 10 ms | 55 % | 3.4 ms | 36 % | 3.7 ms | 71 % | 2.5 ms | 83 % | 0.4 ms | 80 % |<\/p>\n
800 villagers cramming in a 7×7 space at night:<\/p>\n
| Environment | Brain#tick (Total) | Percent | Brain#startTasks | Percent(startTasks) | Brain#tickSensors | Percent(tickSensors) | Brain#updateTasks | Percent(updateTasks) | Brain#tickMemories | Percent(tickMemories) |
\n|:———————–:|:——————:|:——-:|:—————-:|:——————-:|:—————–:|:——————–:|:—————–:|:——————–:|:——————:|:———————:|
\n| Vanilla | 16.7 ms | 100 % | 8.2 ms | 100 % | 6 ms | 100 % | 2 ms | 100 % | 0.5 ms | 100 % |
\n| Lithium | 10.2 ms | 61 % | 3.2 ms | 24 % | 6 ms | 113 % | 0.5 ms | 25 % | 0.5 ms | 100 % |
\n| Sepals | 9 ms | 53 % | 3.3 ms | 16 % | 4.7 ms | 78 % | 0.7 ms | 35 % | 0.3 ms | 60 % |
\n| Sepals + lithium | 8.7 ms | 52 % | 2.9 ms | 11 % | 4.6 ms | 76 % | 0.7 ms | 35 % | 0.5 ms | 100 % |<\/p>\n
### Predicate optimization<\/p>\n
1172 frogs cramming in a 3×3 space:<\/p>\n
| Environment | time | Percent(Avg.) |
\n|——————————————————————————————————————:|:——–:|:————-:|
\n| Vanilla (java.util.function.Predicate.lambda$and$0()) | 49.01 ms | 100 % |
\n| Sepals (com.github.cao.awa.sepals.entity.predicate.SepalsEntityPredicates$$Lambda\/0x000002d8f116e000.test()) | 22.6 ms | 46 % |<\/p>\n","protected":false},"excerpt":{"rendered":"
An extremely radical and experimental optimization for Minecraft server performances.<\/p>\n","protected":false},"featured_media":137196,"template":"","meta":{"_minecraft_slug":"sepals","_minecraft_project_id":"hH1OHXZD","_minecraft_author":"cao-awa","_minecraft_license":"GPL-3.0-only","_minecraft_downloads":34613,"_minecraft_follows":146,"_minecraft_client_side":"unsupported","_minecraft_server_side":"required","_minecraft_storage_type":"zip","_minecraft_archive_name":"sepals.zip","_minecraft_size_bytes":139676125,"_minecraft_version_count":75,"_minecraft_updated_at":"2026-05-16T14:36:34.050862Z","_minecraft_date_created":"2024-10-10T22:11:03.617286Z","_minecraft_date_modified":"2026-03-27T13:20:05.291155Z","_minecraft_project_url":"https:\/\/modrinth.com\/mod\/sepals","_minecraft_icon_attachment_id":137196,"_minecraft_imported_at":"2026-06-08T02:36:12+00:00","_minecraft_import_hash":"eb5756efb19aa14bf6a3c7c916883d7c","_minecraft_versions":"","_minecraft_links":"","_minecraft_gallery":"","_minecraft_api_metadata":""},"mod_category":[97,158,90],"mod_loader":[99,160],"minecraft_version":[62,37,40,41,73,74,38,78,79,80,81,82],"class_list":["post-137195","mod","type-mod","status-publish","has-post-thumbnail","hentry","mod_category-fabric","mod_category-neoforge","mod_category-optimization","mod_loader-fabric","mod_loader-neoforge","minecraft_version-1-21","minecraft_version-1-21-1","minecraft_version-1-21-10","minecraft_version-1-21-11","minecraft_version-1-21-2","minecraft_version-1-21-3","minecraft_version-1-21-4","minecraft_version-1-21-5","minecraft_version-1-21-6","minecraft_version-1-21-7","minecraft_version-1-21-8","minecraft_version-1-21-9"],"yoast_head":"\n