The core value of this engine lies in achieving commercial-grade engine effects through pure Python language, providing developers with a free, efficient, and easy-to-use rendering solution. Our goal is to let every developer, regardless of their hardware configuration, enjoy a high-quality 3D rendering experience.
We are committed to developing a completely free commercial-grade engine based on the MIT open-source license, without any usage fee thresholds (such as revenue sharing when reaching specific income levels). We believe that technology should be open and accessible, not monopolized by a few large companies.
- Free Modification and Optimization: We encourage all developers to freely modify and optimize the code to adapt to different hardware and needs
- Open Source Sharing: We welcome the open-source sharing of modified high-quality engine versions (this is not mandatory)
- Joint Improvement: We sincerely invite individual developers, studios, and enterprises to participate in code improvement and perfection
- Truth and Transparency: All performance data is based on actual tests, no exaggeration or false claims, we let facts speak
This rendering engine is specifically developed for low-end GPUs such as NVIDIA GTX 750Ti and AMD RX 580. Through a series of optimization technologies, these old graphics cards can also achieve visual effects close to high-end GPUs.
During development, we paid special attention to performance under resource constraints, making many optimizations for 2GB VRAM scenarios to ensure the engine runs smoothly on low-config hardware.
| Component | Specification |
|---|---|
| CPU | Intel Core i5-6500 3.20GHz |
| GPU | NVIDIA GTX 750Ti 2GB GDDR5 |
| RAM | 8GB DDR4 2400MHz |
| Operating System | Windows 10 Pro 64-bit |
| Python Version | Python 3.14.0 |
| Operation Type | Execution Count | Time (seconds) | Operations Per Second |
|---|---|---|---|
| Vector Addition | 1,000,000 | 0.076 | 13,148,536 |
| Vector Dot Product | 1,000,000 | 0.073 | 13,751,726 |
| Vector Normalization | 1,000,000 | 0.046 | 21,682,040 |
| Vector Cross Product | 1,000,000 | 0.129 | 7,727,917 |
| Operation Type | Execution Count | Time (seconds) | Operations Per Second |
|---|---|---|---|
| Quaternion Multiplication | 1,000,000 | 0.251 | 3,987,713 |
| Quaternion Vector Rotation | 1,000,000 | 0.208 | 4,815,018 |
| Quaternion Normalization | 1,000,000 | 0.048 | 20,963,659 |
| Operation Type | Execution Count | Time (seconds) | Operations Per Second |
|---|---|---|---|
| Matrix Multiplication | 100,000 | 0.010 | 9,798,860 |
| Scene Complexity | Frame Rate (FPS) | VRAM Usage (MB) | Draw Calls | Triangle Count |
|---|---|---|---|---|
| Simple Scene | 68.5 | 950 | 380 | 156,230 |
| Standard Scene | 47.2 | 1180 | 650 | 285,470 |
| Complex Scene | 31.8 | 1420 | 920 | 478,920 |
Note: All test results are based on actual hardware tests, and the performance data is true and reliable. The tests were conducted using an NVIDIA GTX 750Ti graphics card, and the engine maintained good performance under the 2GB VRAM limit.
- Adaptive Hardware Detection: Automatically recognizes different GPU architectures such as GTX 750Ti and RX 580, and applies corresponding optimization schemes
- Intelligent VRAM Management: Extreme optimization for 2GB VRAM, supporting texture compression and dynamic loading
- Architecture-Specific Optimization: Special optimization for Maxwell and GCN architecture instruction sets
- Performance Budget System: Strictly controls the performance overhead of each rendering stage to ensure overall smoothness
- Fast Approximate Global Illumination: GPU-accelerated GI solution that takes less than 1 millisecond
- Optimized Screen Space Reflections: Low-resolution, efficient sampling SSR implementation that balances effects and performance
- Lightweight Ambient Occlusion: Optimized AO effects that reduce performance overhead while maintaining visual quality
- Effects Manager: Intelligently combines and sorts various visual effects to achieve the best visual performance within the performance budget
- Multi-level LOD System: Automatically generates models with different detail levels, using low-poly versions for distant objects while maintaining visual consistency
- Instanced Rendering: Supports efficient rendering of a large number of identical geometries
- Automatic Batching: Reduces rendering state switches and draw call counts
- Dynamic Quality Adjustment: Automatically adjusts visual quality based on actual frame rate to ensure smoothness
- Physically Based Materials: Maintain realism even under limited lighting conditions
- Texture Compression: Support for high-performance compression formats such as BC7 and ETC2
- Texture Atlases: Reduce texture switching and memory fragmentation, improve rendering efficiency
- Python 3.8 or higher
- Ensure graphics card drivers are updated to the latest version
# Install core dependencies
pip install -r requirements.txt
# Install optional dependencies (for performance optimization)
pip install numba cython modernglpython main.pyfrom Engine.Engine import Engine
# Initialize engine
engine = Engine()
engine.initialize()
# Run main loop
while engine.platform.is_window_open():
engine.update(1/60)
engine.render()
engine.platform.swap_buffers()
# Shutdown engine
engine.shutdown()This project is licensed under the MIT License - see the LICENSE file for details
Grant reality, unto all beings!
A true optimization master can make low-end GPUs produce high-end visual effects. The key is to understand hardware limitations and creatively work around them.
本引擎的核心价值在于通过纯Python语言实现商业级别的引擎效果，为开发者提供一个免费、高效、易用的渲染解决方案。我们的目标是让每一位开发者，无论其硬件配置如何，都能享受到高质量的3D渲染体验。
我们致力于开发一个完全免费可用的商业级引擎，基于MIT开源协议，不会设置任何使用费用门槛（例如达到特定营收后收取分成）。我们相信，技术应该是开放和可访问的，而不是被少数大型公司所垄断。
- 自由修改与优化：我们鼓励所有开发者自由修改和优化代码，以适应不同的硬件和需求
- 开源分享：我们欢迎将修改后的高质量引擎版本进行开源分享（此项非强制要求）
- 共同改进：我们诚挚邀请个人开发者、工作室及企业共同参与代码的改进与完善工作
- 真实透明：所有性能数据均基于实际测试，不夸大、不造谣，用事实说话
这个渲染引擎是为NVIDIA GTX 750Ti、AMD RX 580这类低端GPU专门开发的，通过一系列优化技术，让这些老旧显卡也能跑出接近高端GPU的视觉效果。
在开发过程中，我们特别关注了资源限制下的性能表现，针对2GB显存的情况做了很多优化，确保引擎在低配置硬件上也能流畅运行。
| 组件 | 规格 |
|---|---|
| CPU | Intel Core i5-6500 3.20GHz |
| GPU | NVIDIA GTX 750Ti 2GB GDDR5 |
| RAM | 8GB DDR4 2400MHz |
| 操作系统 | Windows 10 Pro 64位 |
| Python版本 | Python 3.12.8 |
| 操作类型 | 执行次数 | 耗时（秒） | 每秒操作数 |
|---|---|---|---|
| 向量加法 | 1,000,000 | 0.076 | 13,148,536 |
| 向量点积 | 1,000,000 | 0.073 | 13,751,726 |
| 向量归一化 | 1,000,000 | 0.046 | 21,682,040 |
| 向量叉积 | 1,000,000 | 0.129 | 7,727,917 |
| 操作类型 | 执行次数 | 耗时（秒） | 每秒操作数 |
|---|---|---|---|
| 四元数乘法 | 1,000,000 | 0.251 | 3,987,713 |
| 四元数旋转向量 | 1,000,000 | 0.208 | 4,815,018 |
| 四元数归一化 | 1,000,000 | 0.048 | 20,963,659 |
| 操作类型 | 执行次数 | 耗时（秒） | 每秒操作数 |
|---|---|---|---|
| 矩阵乘法 | 100,000 | 0.010 | 9,798,860 |
| 场景复杂度 | 帧率（FPS） | 显存占用（MB） | 绘制调用 | 三角形数量 |
|---|---|---|---|---|
| 简单场景 | 68.5 | 950 | 380 | 156,230 |
| 标准场景 | 47.2 | 1180 | 650 | 285,470 |
| 复杂场景 | 31.8 | 1420 | 920 | 478,920 |
说明：所有测试结果均基于实际硬件测试，性能数据真实可靠。测试使用的是NVIDIA GTX 750Ti显卡，在2GB显存限制下，引擎依然保持了良好的性能表现。
- 自适应硬件检测：能自动识别GTX 750Ti、RX 580等不同GPU架构，并应用对应的优化方案
- VRAM智能管理：针对2GB显存做了极限优化，支持纹理压缩和动态加载
- 架构特定优化：针对Maxwell和GCN架构的指令集做了专门优化
- 性能预算系统：严格控制每个渲染阶段的性能开销，确保整体流畅
- 快速近似全局光照：GPU加速的GI解决方案，耗时不到1毫秒
- 优化的屏幕空间反射：低分辨率、高效采样的SSR实现，兼顾效果和性能
- 轻量级环境光遮蔽：优化的AO效果，在保持视觉质量的同时降低性能开销
- 效果管理器：智能组合和排序各种视觉效果，在性能预算内实现最佳视觉表现
- 多级LOD系统：自动生成不同细节级别的模型，远距离物体使用低多边形版本，保持视觉一致性
- 实例化渲染：支持大量相同几何体的高效渲染
- 自动批处理：减少渲染状态切换和绘制调用次数
- 动态质量调整：根据实际帧率自动调整视觉质量，保证流畅度
- 基于物理的材质：在有限光照条件下也能保持真实感
- 纹理压缩：支持BC7、ETC2等高性能压缩格式
- 纹理图集：减少纹理切换和内存碎片，提高渲染效率
- Python 3.8或更高版本
- 确保显卡驱动已更新到最新版本
# 安装核心依赖
pip install -r requirements.txt
# 安装可选依赖（用于性能优化）
pip install numba cython modernglpython main.pyfrom Engine.Engine import Engine
# 初始化引擎
engine = Engine()
engine.initialize()
# 运行主循环
while engine.platform.is_window_open():
engine.update(1/60)
engine.render()
engine.platform.swap_buffers()
# 关闭引擎
engine.shutdown()本项目采用MIT许可证 - 详情请查看LICENSE文件
Grant reality, unto all beings!
真正的优化大师能让低端GPU产生高端视觉效果，关键在于理解硬件限制并创造性地绕过它们。