Описание
This pack contains **only proven optimizations** that measurably improve **FPS/TPS**.
Redundant tweaks and overlapping features that may introduce regressions are intentionally excluded.
It is built for **NeoForge 1.21.x**, preferring NeoForge-native builds of major optimization mods to maximize compatibility between client and server.
## **Core Principles**
### **• De-duplication**
Avoid overlapping features that cause conflicts or “negative synergy.”
Every included mod has a distinct, non-overlapping role.
### **• Async \& Parallel-First**
Prioritize optimizations that offload or parallelize:
* rendering
* particle processing
* chunk IO
* entity \& block-entity logic
### **• Low Visual Footprint**
No aggressive visual downgrades.
Vanilla 12-chunk render distance, full-quality leaves — performance gains come from engineering, not uglifying.
## **Author Benchmarks** *(author-provided)*
* **5070laptop**: ~3× FPS in normal scenes; ~5–6× in extreme loads
* **i7-13th gen, heavy entity SP**: stable **~20 MSPT** @ vanilla view distance
* **Server tests**: With async pipeline + C2ME, **2048-furnace setups maintain ~20 TPS** even on weak CPUs
## **Included Major / Verified Mods**
### **• Flerovium**
General engine-level performance improvements.
~~**• AsyncParticles**~~
Offloads particle simulation and rendering to multi-threaded pipelines.
### **• [Iris (Async Fork)](https://github.com/Lonmo0208/Iris)**
_Licensed under LGPL-3.0_
An asynchronous, performance-tuned fork of Iris:
* reduces shader-related main-thread stalls
* smoother shader loading
* pairs efficiently with Sodium and async rendering stack
### **• Accelerated Rendering**
GPU compute-shader–based entity \& block-entity renderer.
Replaces per-frame CPU vertex transforms with GPU caching + parallel transforms, significantly reducing CPU load in entity-dense scenes.
### **• Sodium (NeoForge)**
Modern and stable client-side rendering optimizations.
### **• Lithium (NeoForge)**
Logic and tick performance improvements.
### **• C2ME**
Multi-threaded chunk IO and region management.
### **• Entity Culling**
Skips rendering of entities and block entities outside camera visibility.
### **• Additional Mods**
Reese’s Sodium Options, ImmediatelyFast, ModernFix, Noisium, MoreCulling, GPUTape, Krypton FNP, Ixeris, and all other provided mods are included with tuned configs and preserved filenames.
## **FAQ**
### **• Does this change vanilla gameplay?**
No — all vanilla semantics and mechanics are preserved.
### **• Server-friendly?**
Yes. Fully compatible with NeoForge servers and async-heavy environments.
---
本整合包仅包含**经过验证、确实能够提升 FPS / TPS 的性能优化**。
任何可能引入回退、冲突或负优化的冗余与重叠调整,均被有意排除。
整合包基于 **NeoForge 1.21.x** 构建,优先选用 **NeoForge 原生版本**的主流性能优化 Mod,以最大化客户端与服务端之间的兼容性。
## **核心原则**
### **• 模块去重(De-duplication)**
避免功能重叠、相互干扰或产生“负协同效应”的优化。
每一个被纳入的 Mod 都拥有清晰、独立且不可替代的优化职责。
### **• 异步 / 并行优先(Async & Parallel-First)**
优先采用将负载从主线程中卸载或并行化的优化方案,包括但不限于:
* 渲染流程
* 粒子处理
* 区块 IO
* 实体与方块实体逻辑
### **• 低视觉侵入(Low Visual Footprint)**
不采用激进的画面降级手段。
保持原版 **12 区块视距**、完整树叶与高画质表现 —— 性能提升来自工程优化,而非“丑化画面”。
## **作者实测数据** *(由整合包作者提供)*
* **5070笔记本**:常规场景下约 **3× FPS**,极端负载场景可达 **5–6×**
* **i7 第 13 代,单人模式,高实体密度**:在原版视距下稳定维持 **~20 MSPT**
* **服务端测试**:结合异步处理链路与 C2ME,在弱 CPU 环境下
**2048 熔炉阵列仍可维持 ~20 TPS**
## **主要 / 已验证 Mod 组成**
### **• Flerovium**
通用的引擎级性能优化。
### **• AsyncParticles**
将粒子模拟与渲染卸载至多线程执行,降低主线程负载。
### **• [Iris(异步 Fork)](https://github.com/Lonmo0208/Iris)**
_采用 LGPL-3.0 许可证_
一个以性能为导向的 Iris 异步分支版本:
* 显著减少 Shader 相关的主线程阻塞
* 提供更平滑的 Shader 加载体验
* 可高效配合 Sodium 与整合包内的异步渲染体系使用
### **• Accelerated Rendering**
基于 **GPU 计算着色器** 的实体 / 方块实体渲染加速方案。
通过 GPU 缓存与并行变换,取代逐帧的 CPU 顶点变换,在高实体密度场景中显著降低 CPU 占用。
### **• Sodium(NeoForge)**
现代、稳定的客户端渲染性能优化。
### **• Lithium(NeoForge)**
核心逻辑与 Tick 层面的性能优化。
### **• C2ME**
多线程区块 IO 与区块管理优化。
### **• Entity Culling**
跳过摄像机不可见的实体与方块实体渲染。
### **• 其他包含 Mod**
Reese’s Sodium Options、ImmediatelyFast、ModernFix、Noisium、MoreCulling、GPUTape、Krypton FNP、Ixeris 等所有其余 Mod 均按原文件名保留,并已针对兼容性进行配置调整。
## **FAQ**
### **• 是否会改变原版玩法?**
不会 —— 整合包完整保留原版语义与机制,不包含任何玩法改动。
### **• 是否适合服务器使用?**
是的。
完全兼容 **NeoForge 服务端**,并针对高并发、异步负载环境进行了设计。