DBScholar

Back to papers

DFLOP: A Data-driven Framework for Multimodal LLM Training Pipeline Optimization

Summary: DFLOP makes multimodal LLM pipeline parallelism data-aware by profiling input-induced cost variance and using predictive stage/microbatch scheduling. It mitigates heterogeneous-modality skew, improving utilization and throughput by up to 3.6× over existing frameworks. (summarized by gpt-5.6-luna on Jul 26 2026)

Paper ID
7410
Venue
SIGMOD
Year
2026
Pagerank
5.093636e-05
Overall Rank
10,219 | 29.89%
DOI
10.1145/3802037

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

BibTeX Citation

@inproceedings{an_sigmod26,
        title = {{DFLOP: A Data-driven Framework for Multimodal LLM Training Pipeline Optimization}},
        author = {An, Hyeonjun and Kim, Sihyun and Lim, Chaerim and Kim, Hyunjoon and Sen, Rathijit and Jung, Sangmin and Lee, Hyeonsoo and Kim, Dongwook and Yu, Takki and Jeong, Jinkyu and Kim, Youngsok and Park, Kwanghyun},
        series = {{SIGMOD} '26},
        booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
        publisher = {Association for Computing Machinery},
        doi = {10.1145/3802037},
        url = {https://dl.acm.org/doi/10.1145/3802037},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank Citing Paper Year Venue Pagerank
Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 8 of 8 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank Cited Paper Year Venue Pagerank
521 PyTorch Distributed: Experiences on Accelerating Data Parallel Training 2020 VLDB 0.0001713368
1,079 Hybrid Parallelization Strategies for Large-Scale Machine Learning in SystemML 2014 VLDB 0.00012258469
2,473 PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel 2023 VLDB 8.5326287e-05
2,823 Query Processing on Tensor Computation Runtimes 2022 VLDB 8.0893814e-05
2,865 End-to-end Optimization of Machine Learning Prediction Queries 2022 SIGMOD 8.0180243e-05
5,003 Galvatron: Efficient Transformer Training over Multiple GPUs Using Automatic Parallelism 2023 VLDB 6.4065691e-05
5,704 Optimizing Data Pipelines for Machine Learning in Feature Stores 2023 VLDB 6.1146371e-05
6,538 UPLIFT: Parallelization Strategies for Feature Transformations in Machine Learning Workloads 2022 VLDB 5.8477764e-05
Previous Page 1 / 1 Next

Semantically Similar Papers

# Overall Rank Paper Year Venue
1 10,769 Malleus: Straggler-Resilient Hybrid Parallel Training of Large-scale Models via Malleable Data and Model Parallelization 2025 SIGMOD
2 9,219 MemFlow: Memory-Aware Distributed Deep Learning 2020 SIGMOD
3 10,614 LLM-AutoDP: Automatic Data Processing via LLM Agents for Model Fine-tuning 2026 VLDB
4 8,883 FlexMoE: Scaling Large-scale Sparse Pre-trained Model Training via Dynamic Device Placement 2023 SIGMOD
5 13,375 Reimagining Deep Learning Systems Through the Lens of Data Systems 2024 VLDB
6 8,321 FusionFlow: Accelerating Data Preprocessing for Machine Learning with CPU-GPU Cooperation 2024 VLDB
7 8,101 SDP_PIPE: A Semi-Decentralized Framework for Heterogeneity-aware Pipeline-parallel Training 2023 VLDB
8 9,880 MEMO: Fine-grained Tensor Management For Ultra-long Context LLM Training 2025 SIGMOD
9 3,764 FastFlow: Accelerating Deep Learning Model Training with Smart Offloading of Input Data Pipeline 2023 VLDB
10 8,725 mLoRA: Fine-Tuning LoRA Adapters via Highly-Efficient Pipeline Parallelism in Multiple GPUs 2025 VLDB