# GF-DiT: Scheduling Parallelism for Diffusion Transformer Serving

arXiv: https://arxiv.org/abs/2606.13501 · HTML: https://mlsanity.com/paper/2606.13501/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 21 of its 39 references that are in the map.
- First posted on arXiv 11 Jun 2026.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 21 of 21)

- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Qwen-Image Technical Report](https://mlsanity.com/paper/2508.02324/) (2025)
- [Open-Sora: Democratizing Efficient Video Production for All](https://mlsanity.com/paper/2412.20404/) (2024)
- [Latte: Latent Diffusion Transformer for Video Generation](https://mlsanity.com/paper/2401.03048/) (2024)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [Photorealistic Video Generation with Diffusion Models](https://mlsanity.com/paper/2312.06662/) (2023)
- [Seedance 1.0: Exploring the Boundaries of Video Generation Models](https://mlsanity.com/paper/2506.09113/) (2025)
- [Open-Sora 2.0: Training a Commercial-Level Video Generation Model in 200k](https://mlsanity.com/paper/2503.09642/) (2025)
- [VDT: General-purpose Video Diffusion Transformers via Mask Modeling](https://mlsanity.com/paper/2305.13311/) (2023)
- [Allegro: Open the Black Box of Commercial-Level Video Generation Model](https://mlsanity.com/paper/2410.15458/) (2024)
- [xDiT: an Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism](https://mlsanity.com/paper/2411.01738/) (2024)
- [TACCL: Guiding Collective Algorithm Synthesis using Communication Sketches](https://mlsanity.com/paper/2111.04867/) (2021)
- [Synthesizing Optimal Collective Algorithms](https://mlsanity.com/paper/2008.08708/) (2020)
- [PipeFusion: Patch-level Pipeline Parallelism for Diffusion Transformers Inference](https://mlsanity.com/paper/2405.14430/) (2024)
- [UCCL-EP: Portable Expert-Parallel Communication](https://mlsanity.com/paper/2512.19849/) (2025)
- [MSCCL++: Rethinking GPU Communication Abstractions for AI Inference](https://mlsanity.com/paper/2504.09014/) (2025)
- [DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling](https://mlsanity.com/paper/2411.15381/) (2024)
- [GENSERVE: Efficient Co-Serving of Heterogeneous Diffusion Model Workloads](https://mlsanity.com/paper/2604.04335/) (2026)
- [Comprehensive Deadlock Prevention for GPU Collective Communication](https://mlsanity.com/paper/2303.06324/) (2023)
- [PATCHEDSERVE: A Patch Management Framework for SLO-Optimized Hybrid Resolution Diffusion Serving](https://mlsanity.com/paper/2501.09253/) (2025)
- [TridentServe: A Stage-level Serving System for Diffusion Pipelines](https://mlsanity.com/paper/2510.02838/) (2025)

## Built on by (most cited first, 1 of 1)

- [Efficient Vision-Language-Action Management and Serving for Robot Factories](https://mlsanity.com/paper/2609.12075/) (2026)

## Abstract (arXiv)

Diffusion Transformers (DiTs) have become the dominant architecture for image and video generation, creating growing demand for efficient DiT serving. Existing systems assign each request a fixed parallel configuration throughout its lifetime. However, DiT workloads exhibit substantial heterogeneity across requests, execution stages, and system conditions, making static parallelism inefficient and often leading to poor GPU utilization and degraded service quality. This paper argues that DiT serving should treat GPU parallelism as a first-class schedulable resource. We present GF-DiT, a policy-programmable runtime for elastic DiT serving that dynamically adapts the parallelism of running requests according to workload demands and service objectives. GF-DiT introduces an asynchronous execution abstraction that decomposes requests into independently schedulable trajectory tasks and enables online GPU reallocation. To make elastic parallelism practical, GF-DiT further proposes group-free collectives, a lightweight communication abstraction that supports low-overhead online formation and reconfiguration of arbitrary execution groups. We implement GF-DiT in vLLM-Omni and evaluate it on representative image and video diffusion workloads. Compared with fixed-pipeline execution with static parallelism, GF-DiT improves throughput by up to 6.01×, reduces mean latency by up to 95%, lowers SLO violation rates by up to 90%, and reduces communication-group setup overhead from 778 ms to approximately 60 μs. Our code is available at https://github.com/SJTU-Liquid/GF-DiT.