Super Monotonic Alignment Search

TL;DR

This paper introduces Super Monotonic Alignment Search (Super-MAS), a GPU-accelerated version of MAS that significantly speeds up text-to-speech alignment by parallelizing the algorithm and reducing execution time.

Contribution

The paper presents a novel GPU implementation of MAS using Triton and PyTorch JIT, achieving up to 72x speedup over CPU-based methods.

Findings

Super-MAS is up to 72 times faster on GPU.

Parallelization reduces inter-device copy overhead.

Code is publicly available for reproducibility.

Abstract

Monotonic alignment search (MAS), introduced by Glow-TTS, is one of the most popular algorithm in text-to-speech to estimate unknown alignments between text and speech. Since this algorithm needs to search for the most probable alignment with dynamic programming by caching all possible paths, the time complexity of the algorithm is $O(T \times S)$, where $T$ is the length of text and $S$ is the length of speech representation. The authors of Glow-TTS run this algorithm on CPU, and while they mentioned it is difficult to parallelize, we found that MAS can be parallelized in text length dimension and CPU execution consumes an inordinate amount of time for inter-device copy. Therefore, we implemented a Triton kernel and PyTorch JIT script to accelerate MAS on GPU without inter-device copy. As a result, Super-MAS Triton kernel is up to 72 times faster in the extreme-length case. The code is available at https://github.com/supertone-inc/super-monotonic-align.