Ansor: An Auto-scheduler for TVM (AutoTVM v2.0)

@Vengineerの戯言 : Twitter
SystemVerilogの世界へようこそ、すべては、SystemC v0.9公開から始まった 

「Auto-scheduler for TVM」のLianmin Zhengさんから Ansor(AutoTVM v2.0)の提案

discuss.tvm.ai

現在の AutoTVM の問題点

1. The templates are hard to write. It takes a great number of engineering efforts to develop these templates.
2. The number of required templates is large and continues to grow as new operators come.
3. The templates are far from optimal. There is still huge opportunity for performance improvements by enlarging the search space in templates, but manually enumerating these optimizations choices is prohibitive.

を解決するために、Ansor では下記のことをやっているようです。

1. Ansor automatically generates a much larger search space from compute declaration only
2. Ansor searches more efficiently than AutoTVM by utilizing improved cost model, evolutionary search and task scheduler

論文は、Ansor : Generating High-Performance Tensor Programs for Deep Learning

論文のベンチマークを見る限り、かなり速くなりそう。