Intel Games Task Scheduler (GTS)

To the documentation.

Introduction

GTS is a C++ task scheduling framework for multi-processor platforms. It is designed to let developers express their applications in task parallelism and let the scheduling framework maximize the physical parallelism on the available processors.

Key Goals

• Simple expression of task parallelism
• Highly efficient scheduling
• Easy to configure and extend

Target Audience

GTS is designed for the needs of current and future game engine developers. Typical engine developers employ a highly customized, platform-scalable task system (or job system) that provides dedicated worker threads executing thousands of concurrent tasks. These threads may also share resources with driver, networking, middleware, and video streaming threads, all working in synchronized concert to deliver glitch-free real-time rendering on 60+ frames per second (FPS) applications. Future game engines will have to cope with more threads, more tasks, and potentially multiple instruction-set-architectures (ISAs), all running on an ever-expanding hardware landscape.

We built GTS to be simple and friendly to game engine task system use cases. We want a framework that allows the game development community to experiment with and learn from different scheduling algorithms easily. We also want a framework that allows us to demonstrate state-of-the-art algorithms on task scheduling. Finally, we want to encourage games to better express parallelism so they can compute more cool stuff and enable richer PC gaming experiences!

Features

• Easily integrate GTS into an existing game engine complete with low level platform overrides. Game engines support a wide variety of operating systems and hardware platforms, with varying degrees of custom code and work-arounds. Since GTS cannot possibly support every work-around and corner case, we have simplified engine integration by allowing the developer to completely replace the GTS platform layer through a configuration file (user_config.h).
• Express high-level program flow with persistent, dynamic task DAGs that can be executed homogeneously or heterogeneously with the Macro-Scheduler
• Jump right into parallelism with predefined Parallel Patterns
• Easily communicate between threads with Parallel Containers
• Carve up CPU resources as you see fit with the WorkerPool
• Express low-level algorithms and highly efficient execution policies with the Micro-scheduler and Task constructs.
• Remove bottlenecks around heap access with gts_malloc.
• Avoid contention and kernel-mode synchronization with GTS's user-mode synchronization primitives and contention reducing constructs.
• OS-header-free and mostly STL-free interface. GTS won't pollute your engine with unnecessary headers.