NVIDIA TensorRT 11.1 全局性能调优器：多构建路线自动搜索
原标题：Scaling NVIDIA TensorRT Performance Optimization Beyond a Single Build Route with the Global Performance Tuner, Part 2
AI 摘要
NVIDIA 在 TensorRT 11.1 中推出 Global Performance Tuner，将编译调度策略表示为可搜索的 build route（编译器 knob 组合），通过构建多个候选引擎并实测端到端性能与精度来选出最优方案。TensorRT 11.1 提供 200 多个可调 knob，trtexec 中给出 C++ 示例实现，Python 版本将在 Torch-TensorRT 2.15 发布。在 10 个评测负载上，调优相对默认构建路线带来 1.14x 至 1.63x 的性能提升。
正文节选
The Global Performance Tuner is a feature for automatic performance tuning released in NVIDIA TensorRT 11.1. It provides automatic, accuracy-aware and end-to-end performance tuning within TensorRT by customizing different optimizations for different models. For more details, see Part 1 of this series, which introduces the challenges associated with scheduling (a critical stage in deep learning compiler optimization), the motivation for developing the Global Performance Tuner feature in TensorRT,