# Automatic Task Parallelization of Dataflow Graphs in ML/DL models

arXiv: https://arxiv.org/abs/2308.11192 · HTML: https://mlsanity.com/paper/2308.11192/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 25 references that are in the map.
- First posted on arXiv 22 Aug 2023.

## Builds on (most cited first, 7 of 7)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Architectures for Scalable Image Recognition](https://mlsanity.com/paper/1707.07012/) (2017)
- [Device Placement Optimization with Reinforcement Learning](https://mlsanity.com/paper/1706.04972/) (2017)
- [IOS: Inter-Operator Scheduler for CNN Acceleration](https://mlsanity.com/paper/2011.01302/) (2020)
- [Optimus-CC: Efficient Large NLP Model Training with 3D Parallelism Aware Communication Compression](https://mlsanity.com/paper/2301.09830/) (2023)
- [Exploiting Parallelism Opportunities with Deep Learning Frameworks](https://mlsanity.com/paper/1908.04705/) (2019)
- [TensorOpt: Exploring the Tradeoffs in Distributed DNN Training with Auto-Parallelism](https://mlsanity.com/paper/2004.10856/) (2020)

## Abstract (arXiv)

Several methods exist today to accelerate Machine Learning(ML) or Deep-Learning(DL) model performance for training and inference. However, modern techniques that rely on various graph and operator parallelism methodologies rely on search space optimizations which are costly in terms of power and hardware usage. Especially in the case of inference, when the batch size is 1 and execution is on CPUs or for power-constrained edge devices, current techniques can become costly, complicated or inapplicable. To ameliorate this, we present a Critical-Path-based Linear Clustering approach to exploit inherent parallel paths in ML dataflow graphs. Our task parallelization approach further optimizes the structure of graphs via cloning and prunes them via constant propagation and dead-code elimination. Contrary to other work, we generate readable and executable parallel Pytorch+Python code from input ML models in ONNX format via a new tool that we have built called bf Ramiel. This allows us to benefit from other downstream acceleration techniques like intra-op parallelism and potentially pipeline parallelism. Our preliminary results on several ML graphs demonstrate up to 1.9× speedup over serial execution and outperform some of the current mechanisms in both compile and runtimes. Lastly, our methods are lightweight and fast enough so that they can be used effectively for power and resource-constrained devices, while still enabling downstream optimizations.