Archiviert
13
0
Dieses Repository wurde am 2024-12-25 archiviert. Du kannst Dateien ansehen und es klonen, aber nicht pushen oder Issues/Pull-Requests öffnen.
Paper-Old/patches/removed/1.16/0530-Optimize-Villagers.patch

274 Zeilen
14 KiB
Diff

Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Aikar <aikar@aikar.co>
Date: Tue, 26 May 2020 21:32:05 -0400
Subject: [PATCH] Optimize Villagers
This change reimplements the entire BehaviorFindPosition method to
get rid of all of the streams, and implement the logic in a more sane way.
We keep vanilla behavior 100% the same with this change, just wrote more
optimal, as we can abort iterating POI's as soon as we find a match....
One slight change is that Minecraft adds a random delay before a POI is
attempted again. I've increased the amount of that delay based on the distance
to said POI, so farther POI's will not be attempted as often.
Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's.
We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them
all to the pathfinder, so that once we get a match we can return before even looking at other
POI's.
This benefits us in that ideally, a villager will constantly find the near POI's and
not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is
what causes significant lag.
Other improvements here is to stop spamming the POI manager with empty nullables.
Vanilla used them to represent if they needed to load POI data off disk or not.
Well, we load POI data async on chunk load, so we have it, and we surely do not ever
want to load POI data sync either for unloaded chunks!
So this massively reduces object count in the POI hashmaps, resulting in less hash collions,
and also less memory use.
Additionally, unemployed villagers were using significant time due to major ineffeciency in
the code rebuilding data that is static every single invocation for every POI type...
So we cache that and only rebuild it if professions change, which should be never unless
a plugin manipulates and adds custom professions, which it will handle by rebuilding.
diff --git a/src/main/java/net/minecraft/server/BehaviorFindPosition.java b/src/main/java/net/minecraft/server/BehaviorFindPosition.java
index 35eb3a5a61145e94d5b0c77c0eb13bfa46fac23b..6861b1a345496a83900b0ef702ba34315a030ef6 100644
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
--- a/src/main/java/net/minecraft/server/BehaviorFindPosition.java
+++ b/src/main/java/net/minecraft/server/BehaviorFindPosition.java
@@ -29,8 +29,55 @@ public class BehaviorFindPosition extends Behavior<EntityCreature> {
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
protected void a(WorldServer worldserver, EntityCreature entitycreature, long i) {
this.f = 0;
- this.d = worldserver.getTime() + (long) worldserver.getRandom().nextInt(20);
+ this.d = worldserver.getTime() + (long) java.util.concurrent.ThreadLocalRandom.current().nextInt(20); // Paper
VillagePlace villageplace = worldserver.B();
+
+ // Paper start - replace implementation completely
+ BlockPosition blockposition2 = new BlockPosition(entitycreature);
+ int dist = 48;
+ int requiredDist = dist * dist;
+ int cdist = Math.floorDiv(dist, 16);
+ Predicate<VillagePlaceType> predicate = this.a.c();
+ int maxPoiAttempts = 4;
+ int poiAttempts = 0;
+ OUT:
+ for (ChunkCoordIntPair chunkcoordintpair : MCUtil.getSpiralOutChunks(blockposition2, cdist)) {
+ for (int i1 = 0; i1 < 16; i1++) {
+ java.util.Optional<VillagePlaceSection> section = villageplace.getSection(SectionPosition.a(chunkcoordintpair, i1).v());
+ if (section == null || !section.isPresent()) continue;
+ for (java.util.Map.Entry<VillagePlaceType, java.util.Set<VillagePlaceRecord>> e : section.get().getRecords().entrySet()) {
+ if (!predicate.test(e.getKey())) continue;
+ for (VillagePlaceRecord record : e.getValue()) {
+ if (!record.hasVacancy()) continue;
+
+ BlockPosition pos = record.getPosition();
+ long key = pos.asLong();
+ if (this.e.containsKey(key)) {
+ continue;
+ }
+ double poiDist = pos.distanceSquared(blockposition2);
+ if (poiDist <= (double) requiredDist) {
+ this.e.put(key, (long) (this.d + Math.sqrt(poiDist) * 4)); // use dist instead of 40 to blacklist longer if farther distance
+ ++poiAttempts;
+ PathEntity pathentity = entitycreature.getNavigation().a(com.google.common.collect.ImmutableSet.of(pos), 8, false, this.a.d());
+
+ if (pathentity != null && pathentity.h()) {
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
+ record.decreaseVacancy();
+ GlobalPos globalPos = GlobalPos.create(worldserver.getWorldProvider().getDimensionManager(), pos);
+ entitycreature.getBehaviorController().setMemory(this.b, globalPos);
+ break OUT;
+ }
+ if (poiAttempts >= maxPoiAttempts) {
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
+ break OUT;
+ }
+ }
+ }
+ }
+ }
+ }
+ // Clean up - vanilla does this only when it runs out, but that would push it to try farther POI's...
+ this.e.long2LongEntrySet().removeIf((entry) -> entry.getLongValue() < this.d);
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
+ /*
Predicate<BlockPosition> predicate = (blockposition) -> {
long j = blockposition.asLong();
@@ -61,6 +108,6 @@ public class BehaviorFindPosition extends Behavior<EntityCreature> {
Optimize Villager Pathfinding - Massive Villager Improvement This change reimplements the entire BehaviorFindPosition method to get rid of all of the streams, and implement the logic in a more sane way. We keep vanilla behavior 100% the same with this change, just wrote more optimal, as we can abort iterating POI's as soon as we find a match.... One slight change is that Minecraft adds a random delay before a POI is attempted again. I've increased the amount of that delay based on the distance to said POI, so farther POI's will not be attempted as often. Additionally, we spiral out, so we favor local POI's before we ever favor farther POI's. We also try to pathfind 1 POI at a time instead of collecting multiple POI's then tossing them all to the pathfinder, so that once we get a match we can return before even looking at other POI's. This benefits us in that ideally, a villager will constantly find the near POI's and not even try to pathfind to the farther POI. Trying to pathfind to distant POI's is what causes significant lag. Other improvements here is to stop spamming the POI manager with empty nullables. Vanilla used them to represent if they needed to load POI data off disk or not. Well, we load POI data async on chunk load, so we have it, and we surely do not ever want to load POI data sync either for unloaded chunks! So this massively reduces object count in the POI hashmaps, resulting in less hash collions, and also less memory use. Additionally, unemployed villagers were using significant time due to major ineffeciency in the code rebuilding data that is static every single invocation for every POI type... So we cache that and only rebuild it if professions change, which should be never unless a plugin manipulates and adds custom professions, which it will handle by rebuilding.
2020-05-27 07:23:19 +02:00
return entry.getLongValue() < this.d;
});
}
-
+ */ // Paper end
}
}
diff --git a/src/main/java/net/minecraft/server/RegionFileSection.java b/src/main/java/net/minecraft/server/RegionFileSection.java
index a6d8ef5eb44f3f851a3a1be4032ca21ab1d7f2b2..c3e8a0145d63843736d2060f978cdf38df359563 100644
--- a/src/main/java/net/minecraft/server/RegionFileSection.java
+++ b/src/main/java/net/minecraft/server/RegionFileSection.java
@@ -51,29 +51,15 @@ public class RegionFileSection<R extends MinecraftSerializable> extends RegionFi
@Nullable
protected Optional<R> c(long i) {
- return (Optional) this.c.get(i);
+ return this.c.getOrDefault(i, Optional.empty()); // Paper
}
+ protected final Optional<R> getSection(long i) { return d(i); } // Paper - OBFHELPER
protected Optional<R> d(long i) {
- SectionPosition sectionposition = SectionPosition.a(i);
-
- if (this.b(sectionposition)) {
- return Optional.empty();
- } else {
- Optional<R> optional = this.c(i);
-
- if (optional != null) {
- return optional;
- } else {
- this.b(sectionposition.u());
- optional = this.c(i);
- if (optional == null) {
- throw (IllegalStateException) SystemUtils.c(new IllegalStateException());
- } else {
- return optional;
- }
- }
- }
+ // Paper start - replace method - never load POI data sync, we load this in chunk load already, reduce ops
+ // If it's an unloaded chunk, well too bad.
+ return this.c(i);
+ // Paper end
}
protected boolean b(SectionPosition sectionposition) {
@@ -117,7 +103,7 @@ public class RegionFileSection<R extends MinecraftSerializable> extends RegionFi
private <T> void a(ChunkCoordIntPair chunkcoordintpair, DynamicOps<T> dynamicops, @Nullable T t0) {
if (t0 == null) {
for (int i = 0; i < 16; ++i) {
- this.c.put(SectionPosition.a(chunkcoordintpair, i).v(), Optional.empty());
+ //this.c.put(SectionPosition.a(chunkcoordintpair, i).v(), Optional.empty()); // Paper - NO!!!
}
} else {
Dynamic<T> dynamic = new Dynamic(dynamicops, t0);
@@ -135,7 +121,7 @@ public class RegionFileSection<R extends MinecraftSerializable> extends RegionFi
}, dynamic2);
});
- this.c.put(i1, optional);
+ if (optional.isPresent()) this.c.put(i1, optional); // Paper - NO!!!
optional.ifPresent((minecraftserializable) -> {
this.b(i1);
if (flag) {
@@ -199,7 +185,7 @@ public class RegionFileSection<R extends MinecraftSerializable> extends RegionFi
if (optional != null && optional.isPresent()) {
this.d.add(i);
} else {
- RegionFileSection.LOGGER.warn("No data for position: {}", SectionPosition.a(i));
+ //RegionFileSection.LOGGER.warn("No data for position: {}", SectionPosition.a(i)); // Paper - hush
}
}
diff --git a/src/main/java/net/minecraft/server/VillagePlaceRecord.java b/src/main/java/net/minecraft/server/VillagePlaceRecord.java
index 1e9d7a3f902eb4571b93bb0e58cba966365f07b8..44535a3e2c3320aac472c5a7ee557fac7bab2530 100644
--- a/src/main/java/net/minecraft/server/VillagePlaceRecord.java
+++ b/src/main/java/net/minecraft/server/VillagePlaceRecord.java
@@ -32,6 +32,7 @@ public class VillagePlaceRecord implements MinecraftSerializable {
return dynamicops.createMap(ImmutableMap.of(dynamicops.createString("pos"), this.a.a(dynamicops), dynamicops.createString("type"), dynamicops.createString(IRegistry.POINT_OF_INTEREST_TYPE.getKey(this.b).toString()), dynamicops.createString("free_tickets"), dynamicops.createInt(this.c)));
}
+ protected final boolean decreaseVacancy() { return b(); } // Paper - OBFHELPER
protected boolean b() {
if (this.c <= 0) {
return false;
@@ -42,6 +43,7 @@ public class VillagePlaceRecord implements MinecraftSerializable {
}
}
+ protected final boolean increaseVacancy() { return c(); } // Paper - OBFHELPER
protected boolean c() {
if (this.c >= this.b.b()) {
return false;
@@ -52,14 +54,17 @@ public class VillagePlaceRecord implements MinecraftSerializable {
}
}
+ public final boolean hasVacancy() { return d(); } // Paper - OBFHELPER
public boolean d() {
return this.c > 0;
}
+ public final boolean isOccupied() { return e(); } // Paper - OBFHELPER
public boolean e() {
return this.c != this.b.b();
}
+ public final BlockPosition getPosition() { return f(); } // Paper
public BlockPosition f() {
return this.a;
}
diff --git a/src/main/java/net/minecraft/server/VillagePlaceSection.java b/src/main/java/net/minecraft/server/VillagePlaceSection.java
index 3f2602dbe0995f8d01d4a1428d919405d711a205..436b064c3b277143075386fc9a71027fb5962681 100644
--- a/src/main/java/net/minecraft/server/VillagePlaceSection.java
+++ b/src/main/java/net/minecraft/server/VillagePlaceSection.java
@@ -23,7 +23,7 @@ public class VillagePlaceSection implements MinecraftSerializable {
private static final Logger LOGGER = LogManager.getLogger();
private final Short2ObjectMap<VillagePlaceRecord> b = new Short2ObjectOpenHashMap();
- private final Map<VillagePlaceType, Set<VillagePlaceRecord>> c = Maps.newHashMap();
+ private final Map<VillagePlaceType, Set<VillagePlaceRecord>> c = Maps.newHashMap(); public final Map<VillagePlaceType, Set<VillagePlaceRecord>> getRecords() { return c; } // Paper - OBFHELPER
private final Runnable d;
private boolean e;
diff --git a/src/main/java/net/minecraft/server/VillagePlaceType.java b/src/main/java/net/minecraft/server/VillagePlaceType.java
index ab3e054cd2f38756a5d802d4d981022318ab047d..c1f293fc98d3efb4665cfb9036f208b842fc8e36 100644
--- a/src/main/java/net/minecraft/server/VillagePlaceType.java
+++ b/src/main/java/net/minecraft/server/VillagePlaceType.java
@@ -12,8 +12,14 @@ import java.util.stream.Stream;
public class VillagePlaceType {
+ static Set<VillagePlaceType> professionCache; // Paper
private static final Predicate<VillagePlaceType> v = (villageplacetype) -> {
- return ((Set) IRegistry.VILLAGER_PROFESSION.d().map(VillagerProfession::b).collect(Collectors.toSet())).contains(villageplacetype);
+ // Paper start
+ if (professionCache == null) {
+ professionCache = IRegistry.VILLAGER_PROFESSION.d().map(VillagerProfession::b).collect(Collectors.toSet());
+ }
+ return professionCache.contains(villageplacetype);
+ // Paper end
};
public static final Predicate<VillagePlaceType> a = (villageplacetype) -> {
return true;
@@ -89,11 +95,11 @@ public class VillagePlaceType {
}
private static VillagePlaceType a(String s, Set<IBlockData> set, int i, int j) {
- return a((VillagePlaceType) IRegistry.POINT_OF_INTEREST_TYPE.a(new MinecraftKey(s), (Object) (new VillagePlaceType(s, set, i, j))));
+ return a((VillagePlaceType) IRegistry.POINT_OF_INTEREST_TYPE.a(new MinecraftKey(s), (new VillagePlaceType(s, set, i, j)))); // Paper - decompile error
}
private static VillagePlaceType a(String s, Set<IBlockData> set, int i, Predicate<VillagePlaceType> predicate, int j) {
- return a((VillagePlaceType) IRegistry.POINT_OF_INTEREST_TYPE.a(new MinecraftKey(s), (Object) (new VillagePlaceType(s, set, i, predicate, j))));
+ return a((VillagePlaceType) IRegistry.POINT_OF_INTEREST_TYPE.a(new MinecraftKey(s), (new VillagePlaceType(s, set, i, predicate, j)))); // Paper - decompile error
}
private static VillagePlaceType a(VillagePlaceType villageplacetype) {
diff --git a/src/main/java/net/minecraft/server/VillagerProfession.java b/src/main/java/net/minecraft/server/VillagerProfession.java
index c38296165b33698bc15fe49a2de0d0d19cfb910a..f9d7a16c79a4e3ffe8b6e7ed469236a93892f01d 100644
--- a/src/main/java/net/minecraft/server/VillagerProfession.java
+++ b/src/main/java/net/minecraft/server/VillagerProfession.java
@@ -61,6 +61,7 @@ public class VillagerProfession {
}
static VillagerProfession a(String s, VillagePlaceType villageplacetype, ImmutableSet<Item> immutableset, ImmutableSet<Block> immutableset1, @Nullable SoundEffect soundeffect) {
+ VillagePlaceType.professionCache = null; // Paper
return (VillagerProfession) IRegistry.a((IRegistry) IRegistry.VILLAGER_PROFESSION, new MinecraftKey(s), (Object) (new VillagerProfession(s, villageplacetype, immutableset, immutableset1, soundeffect)));
}
}