mirror of
https://github.com/PaperMC/Paper.git
synced 2024-10-01 16:07:36 +02:00
01a13871de
Patch documentation to come Issues with the old system that are fixed now: - World generation does not scale with cpu cores effectively. - Relies on the main thread for scheduling and maintaining chunk state, dropping chunk load/generate rates at lower tps. - Unreliable prioritisation of chunk gen/load calls that block the main thread. - Shutdown logic is utterly unreliable, as it has to wait for all chunks to unload - is it guaranteed that the chunk system is in a state on shutdown that it can reliably do this? Watchdog shutdown also typically failed due to thread checks, which is now resolved. - Saving of data is not unified (i.e can save chunk data without saving entity data, poses problems for desync if shutdown is really abnormal. - Entities are not loaded with chunks. This caused quite a bit of headache for Chunk#getEntities API, but now the new chunk system loads entities with chunks so that they are ready whenever the chunk loads in. Effectively brings the behavior back to 1.16 era, but still storing entities in their own separate regionfiles. The above list is not complete. The patch documentation will complete it. New chunk system hard relies on starlight and dataconverter, and most importantly the new concurrent utilities in ConcurrentUtil. Some of the old async chunk i/o interface (i.e the old file io thread reroutes _some_ calls to the new file io thread) is kept for plugin compat reasons. It will be removed in the next major version of minecraft. The old legacy chunk system patches have been moved to the removed folder in case we need them again.
58 lines
3.2 KiB
Diff
58 lines
3.2 KiB
Diff
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
|
|
From: stonar96 <minecraft.stonar96@gmail.com>
|
|
Date: Sun, 17 Jan 2021 01:11:36 +0100
|
|
Subject: [PATCH] Optimize HashMapPalette
|
|
|
|
HashMapPalette uses an instance of CrudeIncrementalIntIdentityHashBiMap
|
|
internally. A Palette has a preset maximum size = 1 << bits.
|
|
CrudeIncrementalIntIdentityHashBiMap has an initial size but is
|
|
automatically resized. The CrudeIncrementalIntIdentityHashBiMap is created
|
|
with the maximum size in the constructor of HashMapPalette, with the aim
|
|
that it doesn't need to be resized anymore. However, there are two things
|
|
that I think Mojang hasn't considered here:
|
|
1) The CrudeIncrementalIntIdentityHashBiMap is resized, when its initial
|
|
size is reached and not the next time, when a further object is added.
|
|
2) HashMapPalette adds objects (unnecessarily) before checking if the
|
|
initial size of CrudeIncrementalIntIdentityHashBiMap is reached.
|
|
This means to actually avoid resize operations in
|
|
CrudeIncrementalIntIdentityHashBiMap, one has to add 2 to the initial size
|
|
or add 1 and check the size before adding objects. This commit implements
|
|
the second approach. Note that this isn't only an optimization but also
|
|
makes async reads of Palettes fail-safe. An async read while the
|
|
CrudeIncrementalIntIdentityHashBiMap is resized is fatal and can even lead
|
|
to corrupted data. This is also something that Anti-Xray is currently
|
|
relying on.
|
|
|
|
diff --git a/src/main/java/net/minecraft/world/level/chunk/HashMapPalette.java b/src/main/java/net/minecraft/world/level/chunk/HashMapPalette.java
|
|
index c5551c31597f5a7fe62ade325e36425e5f2df0b9..ba9b0f419b0785a0b1e3bc57f18bfe5edaa192bd 100644
|
|
--- a/src/main/java/net/minecraft/world/level/chunk/HashMapPalette.java
|
|
+++ b/src/main/java/net/minecraft/world/level/chunk/HashMapPalette.java
|
|
@@ -19,7 +19,7 @@ public class HashMapPalette<T> implements Palette<T> {
|
|
}
|
|
|
|
public HashMapPalette(IdMap<T> idList, int indexBits, PaletteResize<T> listener) {
|
|
- this(idList, indexBits, listener, CrudeIncrementalIntIdentityHashBiMap.create(1 << indexBits));
|
|
+ this(idList, indexBits, listener, CrudeIncrementalIntIdentityHashBiMap.create((1 << indexBits) + 1)); // Paper - Avoid unnecessary resize operation in CrudeIncrementalIntIdentityHashBiMap
|
|
}
|
|
|
|
private HashMapPalette(IdMap<T> idMap, int i, PaletteResize<T> paletteResize, CrudeIncrementalIntIdentityHashBiMap<T> crudeIncrementalIntIdentityHashBiMap) {
|
|
@@ -37,10 +37,16 @@ public class HashMapPalette<T> implements Palette<T> {
|
|
public int idFor(T object) {
|
|
int i = this.values.getId(object);
|
|
if (i == -1) {
|
|
- i = this.values.add(object);
|
|
- if (i >= 1 << this.bits) {
|
|
+ // Paper start - Avoid unnecessary resize operation in CrudeIncrementalIntIdentityHashBiMap and optimize
|
|
+ // We use size() instead of the result from add(K)
|
|
+ // This avoids adding another object unnecessarily
|
|
+ // Without this change, + 2 would be required in the constructor
|
|
+ if (this.values.size() >= 1 << this.bits) {
|
|
i = this.resizeHandler.onResize(this.bits + 1, object);
|
|
+ } else {
|
|
+ i = this.values.add(object);
|
|
}
|
|
+ // Paper end
|
|
}
|
|
|
|
return i;
|