List of Accepted Paper

MLSys 2025 List of Accepted Papers

Title		Where
Scaling Deep Learning Training with MPMD Pipeline Parallelism Anxhelo Xhebraj ⋅ Sean Lee ⋅ Hanfeng Chen ⋅ Vinod Grover		In Room: Mission City Ballroom Poster Location: #32 in Session 9: Parallel and Distributed Systems.
TileLink: Generating Efficient Compute-Communication Overlapping Kernels using Tile-Centric Primitives Size Zheng ⋅ Jin Fang ⋅ Xuegui Zheng ⋅ Qi Hou ⋅ Wenlei Bao ⋅ Ningxin Zheng ⋅ Ziheng Jiang ⋅ Dongyang Wang ⋅ Jianxi Ye ⋅ Haibin Lin ⋅ Li-Wen Chang ⋅ Xin Liu		In Room: Mission City Ballroom Poster Location: #41 in Session 9: Parallel and Distributed Systems.
The Hidden Bloat in Machine Learning Systems Huaifeng Zhang ⋅ Ahmed Ali-Eldin Hassan		In Room: Mission City Ballroom Poster Location: #51 in Session 4: Reliable and Scalable Systems.
AdaParse: An Adaptive Parallel PDF Parsing and Resource Scaling Engine Project Page ↗ Carlo Siebenschuh ⋅ Kyle Hippe ⋅ Ozan Gokdemir ⋅ Alexander Brace ⋅ Arham Khan ⋅ Khalid Hossain ⋅ Yadu Babuji ⋅ Nicholas Chia ⋅ Venkatram Vishwanath ⋅ Arvind Ramanathan ⋅ Rick Stevens ⋅ Ian Foster ⋅ Robert Underwood		In Room: Mission City Ballroom Poster Location: #60 in Session 2: Parallel and Distributed Systems.
FlexAttention: A Programming Model for Generating Fused Attention Variants. Project Page ↗ Juechu Dong ⋅ BOYUAN FENG ⋅ Driss Guessous ⋅ Yanbo Liang ⋅ Horace He		In Room: Mission City Ballroom Poster Location: #3 in Poster Session and Reception - Main Conference.
Know Where You’re Uncertain When Planning with Multimodal Foundation Models: A Formal Framework Project Page ↗ Neel P. Bhatt ⋅ Yunhao Yang ⋅ Rohan Siva ⋅ Daniel Milan ⋅ Ufuk Topcu ⋅ Atlas Wang		In Room: Mission City Ballroom Poster Location: #16 in Poster Session and Reception - Main Conference.
Marconi: Prefix Caching for the Era of Hybrid LLMs Project Page ↗ Rui Pan ⋅ Zhuang Wang ⋅ Zhen Jia ⋅ Can Karakus ⋅ Luca Zancato ⋅ Tri Dao ⋅ Yida Wang ⋅ Ravi Netravali		In Room: Mission City Ballroom Poster Location: #29 in Poster Session and Reception - Main Conference.
Balancing Pipeline Parallelism with Vocabulary Parallelism Project Page ↗ Man Tsung Yeung ⋅ Penghui Qi ⋅ Min Lin ⋅ Xinyi Wan		In Room: Mission City Ballroom Poster Location: #52 in Poster Session and Reception - Main Conference.
Balancing Pipeline Parallelism with Vocabulary Parallelism Project Page ↗ Man Tsung Yeung ⋅ Penghui Qi ⋅ Min Lin ⋅ Xinyi Wan		In Room: Mission City Ballroom Poster Location: #52 in Session 9: Parallel and Distributed Systems.
FlexInfer: Flexible LLM Inference with CPU Computations Seonjin Na ⋅ Geonhwa Jeong ⋅ Byung Hoon Ahn ⋅ Aaron Jezghani ⋅ Jeffrey Young ⋅ Christopher Hughes ⋅ Tushar Krishna ⋅ Hyesoon Kim		In Room: Mission City Ballroom Poster Location: #55 in Session 8: LLM and Diffusion Model Serving.
A Bring-Your-Own-Model Approach for ML-Driven Storage Placement in Warehouse-Scale Computers Chenxi Yang ⋅ Yan Li ⋅ Martin Maas ⋅ Mustafa Uysal ⋅ Ubaid Hafeez ⋅ Arif Merchant ⋅ Richard McDougall		In Room: Mission City Ballroom Poster Location: #18 in Session 6: Edge and Cloud Systems.
SampleAttention: Near-Lossless Acceleration of Long Context LLM Inference with Adaptive Structured Sparse Attention Qianchao Zhu ⋅ Jiangfei Duan ⋅ Chang Chen ⋅ Siran Liu ⋅ Xiuhong Li ⋅ Guanyu Feng ⋅ Xin Lv ⋅ Xiao Chuanfu ⋅ Dahua Lin ⋅ Chao Yang		In Room: Mission City Ballroom Poster Location: #31 in Session 7: Quantization and Sparsity.
FlexAttention: A Programming Model for Generating Fused Attention Variants. Project Page ↗ Juechu Dong ⋅ BOYUAN FENG ⋅ Driss Guessous ⋅ Yanbo Liang ⋅ Horace He		In Room: Mission City Ballroom Poster Location: #3 in Session 10: LLM and Diffusion Model Serving.
ThunderServe: High-performance and Cost-efficient LLM Serving in Cloud Environments YOUHE JIANG ⋅ Fangcheng Fu ⋅ Xiaozhe Yao ⋅ Taiyi Wang ⋅ Bin CUI ⋅ Ana Klimovic ⋅ Eiko Yoneki		In Room: Mission City Ballroom Poster Location: #5 in Session 10: LLM and Diffusion Model Serving.
PipeFill: Using GPUs During Bubbles in Pipeline-parallel LLM Training Daiyaan Arfeen ⋅ Zhen Zhang ⋅ Xinwei Fu ⋅ Gregory R. Ganger ⋅ Yida Wang		In Room: Mission City Ballroom Poster Location: #6 in Session 2: Parallel and Distributed Systems.
SparseTransX: Efficient Training of Translation-Based Knowledge Graph Embeddings Using Sparse Matrix Operations Project Page ↗ Md Saidul Hoque Anik ⋅ Ariful Azad		In Room: Mission City Ballroom Poster Location: #7 in Session 7: Quantization and Sparsity.
ProtoRAIL: A Risk-cognizant Imitation Agent for Adaptive vCPU Oversubscription In the Cloud Lu Wang ⋅ Mayukh Das ⋅ Fangkai Yang ⋅ Bo Qiao ⋅ Hang Dong ⋅ Si Qin ⋅ Victor Ruehle ⋅ Chetan Bansal ⋅ Eli Cortez ⋅ Íñigo Goiri ⋅ S R ⋅ Qingwei Lin ⋅ Dongmei Zhang		In Room: Mission City Ballroom Poster Location: #12 in Session 6: Edge and Cloud Systems.
FastTree: Optimizing Attention Kernel and Runtime for Tree-Structured LLM Inference Zaifeng Pan ⋅ Yitong Ding ⋅ Yue Guan ⋅ Zheng Wang ⋅ Zhongkai Yu ⋅ Xulong Tang ⋅ Yida Wang ⋅ Yufei Ding		In Room: Mission City Ballroom Poster Location: #11 in Session 1: LLM and Diffusion Model Serving.
Supply-Chain Attacks in Machine Learning Frameworks Yue Gao ⋅ Ilia Shumailov ⋅ Kassem Fawaz		In Room: Mission City Ballroom Poster Location: #13 in Session 12: Edge and Cloud Systems.
Graph Learning at Scale: Characterizing and Optimizing Pre-Propagation GNNs Zichao Yue ⋅ Chenhui Deng ⋅ Zhiru Zhang		In Room: Mission City Ballroom Poster Location: #15 in Session 12: Edge and Cloud Systems.
LeanAttention: Hardware-Aware Scalable Attention Mechanism for the Decode-Phase of Transformers Rya Sanovar ⋅ Srikant Bharadwaj ⋅ Renée St. Amant ⋅ Victor Ruehle ⋅ Saravan Rajmohan		In Room: Mission City Ballroom Poster Location: #20 in Session 1: LLM and Diffusion Model Serving.
Lightweight Software Kernels and Hardware Extensions for Efficient Sparse Deep Neural Networks on Microcontrollers Francesco Daghero ⋅ Daniele Jahier Pagliari ⋅ Francesco Conti ⋅ Luca Benini ⋅ Massimo Poncino ⋅ Alessio Burrello		In Room: Mission City Ballroom Poster Location: #22 in Session 7: Quantization and Sparsity.
Training Ultra Long Context Language Model with Fully Pipelined Distributed Transformer Jinghan Yao ⋅ Sam Jacobs ⋅ Masahiro Tanaka ⋅ Olatunji Ruwase ⋅ Hari Subramoni ⋅ Dhabaleswar Panda		In Room: Mission City Ballroom Poster Location: #21 in Session 5: LLM training and fine-tuning.
MiLo: Efficient Quantized MoE Inference with Mixture of Low-Rank Compensators Beichen Huang ⋅ Yueming Yuan ⋅ ZELEI SHAO ⋅ Minjia Zhang		In Room: Mission City Ballroom Poster Location: #23 in Session 3: Quantization and Sparsity.
QServe:W4A8KV4 Quantization and System Co-design for Efficient LLM Serving Project Page ↗ Yujun Lin ⋅ Haotian Tang ⋅ Shang Yang ⋅ Zhekai Zhang ⋅ Guangxuan Xiao ⋅ Chuang Gan ⋅ Song Han		In Room: Mission City Ballroom Poster Location: #1 in Session 3: Quantization and Sparsity.
Efficient On-Device Machine Learning with a Biologically-Plausible Forward-Only Algorithm Project Page ↗ Baichuan Huang ⋅ Amir Aminifar		In Room: Mission City Ballroom Poster Location: #25 in Session 6: Edge and Cloud Systems.
FLStore: Efficient Federated Learning Storage for non-training workloads Ahmad Faraz Khan ⋅ Samuel Fountain ⋅ Ahmed Mohamed Abdelmoniem Sayed ⋅ Ali R. Butt ⋅ Ali Anwar		In Room: Mission City Ballroom Poster Location: #26 in Session 11: Federated Learning.
Enabling Unstructured Sparse Acceleration on Structured Sparse Accelerators Geonhwa Jeong ⋅ Po-An Tsai ⋅ Abhimanyu Rajeshkumar Bambhaniya ⋅ Stephen Keckler ⋅ Tushar Krishna		In Room: Mission City Ballroom Poster Location: #27 in Session 3: Quantization and Sparsity.
FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving Zihao Ye ⋅ Lequn Chen ⋅ Ruihang Lai ⋅ Wuwei Lin ⋅ Yineng Zhang ⋅ Stephanie Wang ⋅ Tianqi Chen ⋅ Baris Kasikci ⋅ Vinod Grover ⋅ Arvind Krishnamurthy ⋅ Luis Ceze		In Room: Mission City Ballroom Poster Location: #30 in Session 1: LLM and Diffusion Model Serving.
Radius: Range-based Gradient Sparsity for Large Foundation Model Pre-training Mingkai Zheng ⋅ Zhao Zhang		In Room: Mission City Ballroom Poster Location: #33 in Session 3: Quantization and Sparsity.
Context Parallelism for Scalable Million-Token Inference Amy Yang ⋅ Jingyi Yang ⋅ Aya Ibrahim ⋅ Xinfeng Xie ⋅ Bangsheng Tang ⋅ Grigory Sizov ⋅ Jongsoo Park ⋅ Jianyu Huang		In Room: Mission City Ballroom Poster Location: #34 in Session 2: Parallel and Distributed Systems.
Seesaw: High-throughput LLM Inference via Model Re-sharding Qidong Su ⋅ Wei Zhao ⋅ Xin Li ⋅ Muralidhar Andoorveedu ⋅ Chenhao Jiang ⋅ Zhanda Zhu ⋅ Kevin Song ⋅ Christina Giannoula ⋅ Gennady Pekhimenko		In Room: Mission City Ballroom Poster Location: #36 in Session 8: LLM and Diffusion Model Serving.
ScaleFusion: Scalable Inference of Spatial-Temporal Diffusion Transformers for High-Resolution Long Video Generation Jiacheng Yang ⋅ Jun Wu ⋅ Zhen Zhang ⋅ Xinwei Fu ⋅ Zhiying Xu ⋅ Zhen Jia ⋅ Yida Wang ⋅ Gennady Pekhimenko		In Room: Mission City Ballroom Poster Location: #37 in Session 8: LLM and Diffusion Model Serving.
TurboAttention: Efficient attention approximation for high throughputs llm Hao Kang ⋅ Srikant Bharadwaj ⋅ James Hensman ⋅ Tushar Krishna ⋅ Victor Ruehle ⋅ Saravan Rajmohan		In Room: Mission City Ballroom Poster Location: #39 in Session 8: LLM and Diffusion Model Serving.
GSplit: Scaling Graph Neural Network Training on Large Graphs via Split-Parallelism Sandeep Polisetty ⋅ Juelin Liu ⋅ Yi Fung ⋅ Seung-Hwan Lim ⋅ Hui Guan ⋅ Marco Serafini		In Room: Mission City Ballroom Poster Location: #40 in Session 2: Parallel and Distributed Systems.
Self-Data Distillation for Recovering Quality in Pruned Large Language Models Vithursan Thangarasa ⋅ Ganesh Venkatesh ⋅ Mike Lasby ⋅ Nish Sinnadurai ⋅ Sean Lie		In Room: Mission City Ballroom Poster Location: #42 in Session 3: Quantization and Sparsity.
MAS-ATTENTION: MEMORY-AWARE STREAM PROCESSING FOR ATTENTION ACCELERATION ON RESOURCE-CONSTRAINED EDGE DEVICES Mohammadali Shakerdargah ⋅ Shan Lu ⋅ Chao Gao ⋅ Di Niu		In Room: Mission City Ballroom Poster Location: #44 in Session 11: Federated Learning.
Interference-aware Edge Runtime Prediction with Conformal Matrix Completion Tianshu Huang ⋅ Arjun Ramesh ⋅ Emily Ruppel ⋅ Nuno Pereira ⋅ Anthony Rowe ⋅ Carlee Joe-Wong		In Room: Mission City Ballroom Poster Location: #47 in Session 4: Reliable and Scalable Systems.
APOLLO: SGD-like Memory, AdamW-level Performance Hanqing Zhu ⋅ Zhenyu Zhang ⋅ Wenyan Cong ⋅ Xi Liu ⋅ Sem Park ⋅ Vikas Chandra ⋅ Bo Long ⋅ David Pan ⋅ Atlas Wang ⋅ Jinwon Lee		In Room: Mission City Ballroom Poster Location: #48 in Session 5: LLM training and fine-tuning.
Lumos: Efficient Performance Modeling and Estimation for Large-scale LLM Training Mingyu Liang ⋅ Hiwot Kassa ⋅ Wenyin Fu ⋅ Brian Coutinho ⋅ Louis Feng ⋅ Christina Delimitrou		In Room: Mission City Ballroom Poster Location: #49 in Session 5: LLM training and fine-tuning.
Rethinking Key-Value Cache Compression Techniques for Large Language Model Serving Wei Gao ⋅ Xinyu Zhou ⋅ Peng Sun ⋅ Tianwei Zhang ⋅ Yonggang Wen		In Room: Mission City Ballroom Poster Location: #53 in Session 1: LLM and Diffusion Model Serving.
XGrammar: Flexible and Efficient Structured Generation Engine for Large Language Models Yixin Dong ⋅ Charlie Ruan ⋅ Yaxing Cai ⋅ Ziyi Xu ⋅ Yilong Zhao ⋅ Ruihang Lai ⋅ Tianqi Chen		In Room: Mission City Ballroom Poster Location: #54 in Session 10: LLM and Diffusion Model Serving.
On Distributed Larger-Than-Memory Subset Selection With Pairwise Submodular Functions Maximilian Böther ⋅ Abe Sebastian ⋅ Pranjal Awasthi ⋅ Ana Klimovic ⋅ Srikumar Ramalingam		In Room: Mission City Ballroom Poster Location: #56 in Session 9: Parallel and Distributed Systems.
Rubick: Exploiting Job Reconfigurability for Deep Learning Cluster Scheduling Project Page ↗ Xinyi Zhang ⋅ Hanyu Zhao ⋅ Wencong Xiao ⋅ Xianyan Jia ⋅ Fei Xu ⋅ Yong Li ⋅ Wei Lin ⋅ Fangming Liu		In Room: Mission City Ballroom Poster Location: #57 in Session 2: Parallel and Distributed Systems.
SOLA: Optimizing SLO Attainment for Large Language Model Serving with State-Aware Scheduling Ke Hong ⋅ Xiuhong Li ⋅ Lufang Chen ⋅ Qiuli Mao ⋅ Guohao Dai ⋅ Xuefei Ning ⋅ Shengen Yan ⋅ Yun Liang ⋅ Yu Wang		In Room: Mission City Ballroom Poster Location: #58 in Session 8: LLM and Diffusion Model Serving.
NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference Project Page ↗ Xuanlin Jiang ⋅ Yang Zhou ⋅ Shiyi Cao ⋅ Ion Stoica ⋅ Minlan Yu		In Room: Mission City Ballroom Poster Location: #59 in Session 10: LLM and Diffusion Model Serving.
ReaL: Efficient RLHF Training of Large Language Models with Parameter Reallocation Zhiyu Mei ⋅ WEI FU ⋅ Kaiwei Li ⋅ Guangju Wang ⋅ Huanchen Zhang ⋅ Yi Wu		In Room: Mission City Ballroom Poster Location: #61 in Session 5: LLM training and fine-tuning.
DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling Sohaib Ahmad ⋅ Qizheng Yang ⋅ Haoliang Wang ⋅ Ramesh Sitaraman ⋅ Hui Guan		In Room: Mission City Ballroom Poster Location: #2 in Session 1: LLM and Diffusion Model Serving.
Know Where You’re Uncertain When Planning with Multimodal Foundation Models: A Formal Framework Project Page ↗ Neel P. Bhatt ⋅ Yunhao Yang ⋅ Rohan Siva ⋅ Daniel Milan ⋅ Ufuk Topcu ⋅ Atlas Wang		In Room: Mission City Ballroom Poster Location: #16 in Session 4: Reliable and Scalable Systems.
SampleAttention: Near-Lossless Acceleration of Long Context LLM Inference with Adaptive Structured Sparse Attention Qianchao Zhu ⋅ Jiangfei Duan ⋅ Chang Chen ⋅ Siran Liu ⋅ Xiuhong Li ⋅ Guanyu Feng ⋅ Xin Lv ⋅ Xiao Chuanfu ⋅ Dahua Lin ⋅ Chao Yang		In Room: Mission City Ballroom Poster Location: #31 in Poster Session and Reception - Main Conference.
A Bring-Your-Own-Model Approach for ML-Driven Storage Placement in Warehouse-Scale Computers Chenxi Yang ⋅ Yan Li ⋅ Martin Maas ⋅ Mustafa Uysal ⋅ Ubaid Hafeez ⋅ Arif Merchant ⋅ Richard McDougall		In Room: Mission City Ballroom Poster Location: #18 in Poster Session and Reception - Main Conference.
QServe:W4A8KV4 Quantization and System Co-design for Efficient LLM Serving Project Page ↗ Yujun Lin ⋅ Haotian Tang ⋅ Shang Yang ⋅ Zhekai Zhang ⋅ Guangxuan Xiao ⋅ Chuang Gan ⋅ Song Han		In Room: Mission City Ballroom Poster Location: #1 in Poster Session and Reception - Main Conference.
ThunderServe: High-performance and Cost-efficient LLM Serving in Cloud Environments YOUHE JIANG ⋅ Fangcheng Fu ⋅ Xiaozhe Yao ⋅ Taiyi Wang ⋅ Bin CUI ⋅ Ana Klimovic ⋅ Eiko Yoneki		In Room: Mission City Ballroom Poster Location: #5 in Poster Session and Reception - Main Conference.
SparseTransX: Efficient Training of Translation-Based Knowledge Graph Embeddings Using Sparse Matrix Operations Project Page ↗ Md Saidul Hoque Anik ⋅ Ariful Azad		In Room: Mission City Ballroom Poster Location: #7 in Poster Session and Reception - Main Conference.
PipeFill: Using GPUs During Bubbles in Pipeline-parallel LLM Training Daiyaan Arfeen ⋅ Zhen Zhang ⋅ Xinwei Fu ⋅ Gregory R. Ganger ⋅ Yida Wang		In Room: Mission City Ballroom Poster Location: #6 in Poster Session and Reception - Main Conference.
FastTree: Optimizing Attention Kernel and Runtime for Tree-Structured LLM Inference Zaifeng Pan ⋅ Yitong Ding ⋅ Yue Guan ⋅ Zheng Wang ⋅ Zhongkai Yu ⋅ Xulong Tang ⋅ Yida Wang ⋅ Yufei Ding		In Room: Mission City Ballroom Poster Location: #11 in Poster Session and Reception - Main Conference.
ProtoRAIL: A Risk-cognizant Imitation Agent for Adaptive vCPU Oversubscription In the Cloud Lu Wang ⋅ Mayukh Das ⋅ Fangkai Yang ⋅ Bo Qiao ⋅ Hang Dong ⋅ Si Qin ⋅ Victor Ruehle ⋅ Chetan Bansal ⋅ Eli Cortez ⋅ Íñigo Goiri ⋅ S R ⋅ Qingwei Lin ⋅ Dongmei Zhang		In Room: Mission City Ballroom Poster Location: #12 in Poster Session and Reception - Main Conference.
Supply-Chain Attacks in Machine Learning Frameworks Yue Gao ⋅ Ilia Shumailov ⋅ Kassem Fawaz		In Room: Mission City Ballroom Poster Location: #13 in Poster Session and Reception - Main Conference.
Graph Learning at Scale: Characterizing and Optimizing Pre-Propagation GNNs Zichao Yue ⋅ Chenhui Deng ⋅ Zhiru Zhang		In Room: Mission City Ballroom Poster Location: #15 in Poster Session and Reception - Main Conference.
LeanAttention: Hardware-Aware Scalable Attention Mechanism for the Decode-Phase of Transformers Rya Sanovar ⋅ Srikant Bharadwaj ⋅ Renée St. Amant ⋅ Victor Ruehle ⋅ Saravan Rajmohan		In Room: Mission City Ballroom Poster Location: #20 in Poster Session and Reception - Main Conference.
Training Ultra Long Context Language Model with Fully Pipelined Distributed Transformer Jinghan Yao ⋅ Sam Jacobs ⋅ Masahiro Tanaka ⋅ Olatunji Ruwase ⋅ Hari Subramoni ⋅ Dhabaleswar Panda		In Room: Mission City Ballroom Poster Location: #21 in Poster Session and Reception - Main Conference.
Lightweight Software Kernels and Hardware Extensions for Efficient Sparse Deep Neural Networks on Microcontrollers Francesco Daghero ⋅ Daniele Jahier Pagliari ⋅ Francesco Conti ⋅ Luca Benini ⋅ Massimo Poncino ⋅ Alessio Burrello		In Room: Mission City Ballroom Poster Location: #22 in Poster Session and Reception - Main Conference.
MiLo: Efficient Quantized MoE Inference with Mixture of Low-Rank Compensators Beichen Huang ⋅ Yueming Yuan ⋅ ZELEI SHAO ⋅ Minjia Zhang		In Room: Mission City Ballroom Poster Location: #23 in Poster Session and Reception - Main Conference.
Efficient On-Device Machine Learning with a Biologically-Plausible Forward-Only Algorithm Project Page ↗ Baichuan Huang ⋅ Amir Aminifar		In Room: Mission City Ballroom Poster Location: #25 in Poster Session and Reception - Main Conference.
FLStore: Efficient Federated Learning Storage for non-training workloads Ahmad Faraz Khan ⋅ Samuel Fountain ⋅ Ahmed Mohamed Abdelmoniem Sayed ⋅ Ali R. Butt ⋅ Ali Anwar		In Room: Mission City Ballroom Poster Location: #26 in Poster Session and Reception - Main Conference.
Enabling Unstructured Sparse Acceleration on Structured Sparse Accelerators Geonhwa Jeong ⋅ Po-An Tsai ⋅ Abhimanyu Rajeshkumar Bambhaniya ⋅ Stephen Keckler ⋅ Tushar Krishna		In Room: Mission City Ballroom Poster Location: #27 in Poster Session and Reception - Main Conference.
Scaling Deep Learning Training with MPMD Pipeline Parallelism Anxhelo Xhebraj ⋅ Sean Lee ⋅ Hanfeng Chen ⋅ Vinod Grover		In Room: Mission City Ballroom Poster Location: #32 in Poster Session and Reception - Main Conference.
FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving Zihao Ye ⋅ Lequn Chen ⋅ Ruihang Lai ⋅ Wuwei Lin ⋅ Yineng Zhang ⋅ Stephanie Wang ⋅ Tianqi Chen ⋅ Baris Kasikci ⋅ Vinod Grover ⋅ Arvind Krishnamurthy ⋅ Luis Ceze		In Room: Mission City Ballroom Poster Location: #30 in Poster Session and Reception - Main Conference.
Radius: Range-based Gradient Sparsity for Large Foundation Model Pre-training Mingkai Zheng ⋅ Zhao Zhang		In Room: Mission City Ballroom Poster Location: #33 in Poster Session and Reception - Main Conference.
Context Parallelism for Scalable Million-Token Inference Amy Yang ⋅ Jingyi Yang ⋅ Aya Ibrahim ⋅ Xinfeng Xie ⋅ Bangsheng Tang ⋅ Grigory Sizov ⋅ Jongsoo Park ⋅ Jianyu Huang		In Room: Mission City Ballroom Poster Location: #34 in Poster Session and Reception - Main Conference.
Seesaw: High-throughput LLM Inference via Model Re-sharding Qidong Su ⋅ Wei Zhao ⋅ Xin Li ⋅ Muralidhar Andoorveedu ⋅ Chenhao Jiang ⋅ Zhanda Zhu ⋅ Kevin Song ⋅ Christina Giannoula ⋅ Gennady Pekhimenko		In Room: Mission City Ballroom Poster Location: #36 in Poster Session and Reception - Main Conference.
TurboAttention: Efficient attention approximation for high throughputs llm Hao Kang ⋅ Srikant Bharadwaj ⋅ James Hensman ⋅ Tushar Krishna ⋅ Victor Ruehle ⋅ Saravan Rajmohan		In Room: Mission City Ballroom Poster Location: #39 in Poster Session and Reception - Main Conference.
ScaleFusion: Scalable Inference of Spatial-Temporal Diffusion Transformers for High-Resolution Long Video Generation Jiacheng Yang ⋅ Jun Wu ⋅ Zhen Zhang ⋅ Xinwei Fu ⋅ Zhiying Xu ⋅ Zhen Jia ⋅ Yida Wang ⋅ Gennady Pekhimenko		In Room: Mission City Ballroom Poster Location: #37 in Poster Session and Reception - Main Conference.
GSplit: Scaling Graph Neural Network Training on Large Graphs via Split-Parallelism Sandeep Polisetty ⋅ Juelin Liu ⋅ Yi Fung ⋅ Seung-Hwan Lim ⋅ Hui Guan ⋅ Marco Serafini		In Room: Mission City Ballroom Poster Location: #40 in Poster Session and Reception - Main Conference.
TileLink: Generating Efficient Compute-Communication Overlapping Kernels using Tile-Centric Primitives Size Zheng ⋅ Jin Fang ⋅ Xuegui Zheng ⋅ Qi Hou ⋅ Wenlei Bao ⋅ Ningxin Zheng ⋅ Ziheng Jiang ⋅ Dongyang Wang ⋅ Jianxi Ye ⋅ Haibin Lin ⋅ Li-Wen Chang ⋅ Xin Liu		In Room: Mission City Ballroom Poster Location: #41 in Poster Session and Reception - Main Conference.
MAS-ATTENTION: MEMORY-AWARE STREAM PROCESSING FOR ATTENTION ACCELERATION ON RESOURCE-CONSTRAINED EDGE DEVICES Mohammadali Shakerdargah ⋅ Shan Lu ⋅ Chao Gao ⋅ Di Niu		In Room: Mission City Ballroom Poster Location: #44 in Poster Session and Reception - Main Conference.
Self-Data Distillation for Recovering Quality in Pruned Large Language Models Vithursan Thangarasa ⋅ Ganesh Venkatesh ⋅ Mike Lasby ⋅ Nish Sinnadurai ⋅ Sean Lie		In Room: Mission City Ballroom Poster Location: #42 in Poster Session and Reception - Main Conference.
Interference-aware Edge Runtime Prediction with Conformal Matrix Completion Tianshu Huang ⋅ Arjun Ramesh ⋅ Emily Ruppel ⋅ Nuno Pereira ⋅ Anthony Rowe ⋅ Carlee Joe-Wong		In Room: Mission City Ballroom Poster Location: #47 in Poster Session and Reception - Main Conference.
APOLLO: SGD-like Memory, AdamW-level Performance Hanqing Zhu ⋅ Zhenyu Zhang ⋅ Wenyan Cong ⋅ Xi Liu ⋅ Sem Park ⋅ Vikas Chandra ⋅ Bo Long ⋅ David Pan ⋅ Atlas Wang ⋅ Jinwon Lee		In Room: Mission City Ballroom Poster Location: #48 in Poster Session and Reception - Main Conference.
Lumos: Efficient Performance Modeling and Estimation for Large-scale LLM Training Mingyu Liang ⋅ Hiwot Kassa ⋅ Wenyin Fu ⋅ Brian Coutinho ⋅ Louis Feng ⋅ Christina Delimitrou		In Room: Mission City Ballroom Poster Location: #49 in Poster Session and Reception - Main Conference.
The Hidden Bloat in Machine Learning Systems Huaifeng Zhang ⋅ Ahmed Ali-Eldin Hassan		In Room: Mission City Ballroom Poster Location: #51 in Poster Session and Reception - Main Conference.
Rethinking Key-Value Cache Compression Techniques for Large Language Model Serving Wei Gao ⋅ Xinyu Zhou ⋅ Peng Sun ⋅ Tianwei Zhang ⋅ Yonggang Wen		In Room: Mission City Ballroom Poster Location: #53 in Poster Session and Reception - Main Conference.
On Distributed Larger-Than-Memory Subset Selection With Pairwise Submodular Functions Maximilian Böther ⋅ Abe Sebastian ⋅ Pranjal Awasthi ⋅ Ana Klimovic ⋅ Srikumar Ramalingam		In Room: Mission City Ballroom Poster Location: #56 in Poster Session and Reception - Main Conference.
XGrammar: Flexible and Efficient Structured Generation Engine for Large Language Models Yixin Dong ⋅ Charlie Ruan ⋅ Yaxing Cai ⋅ Ziyi Xu ⋅ Yilong Zhao ⋅ Ruihang Lai ⋅ Tianqi Chen		In Room: Mission City Ballroom Poster Location: #54 in Poster Session and Reception - Main Conference.
Rubick: Exploiting Job Reconfigurability for Deep Learning Cluster Scheduling Project Page ↗ Xinyi Zhang ⋅ Hanyu Zhao ⋅ Wencong Xiao ⋅ Xianyan Jia ⋅ Fei Xu ⋅ Yong Li ⋅ Wei Lin ⋅ Fangming Liu		In Room: Mission City Ballroom Poster Location: #57 in Poster Session and Reception - Main Conference.
NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference Project Page ↗ Xuanlin Jiang ⋅ Yang Zhou ⋅ Shiyi Cao ⋅ Ion Stoica ⋅ Minlan Yu		In Room: Mission City Ballroom Poster Location: #59 in Poster Session and Reception - Main Conference.
SOLA: Optimizing SLO Attainment for Large Language Model Serving with State-Aware Scheduling Ke Hong ⋅ Xiuhong Li ⋅ Lufang Chen ⋅ Qiuli Mao ⋅ Guohao Dai ⋅ Xuefei Ning ⋅ Shengen Yan ⋅ Yun Liang ⋅ Yu Wang		In Room: Mission City Ballroom Poster Location: #58 in Poster Session and Reception - Main Conference.
AdaParse: An Adaptive Parallel PDF Parsing and Resource Scaling Engine Project Page ↗ Carlo Siebenschuh ⋅ Kyle Hippe ⋅ Ozan Gokdemir ⋅ Alexander Brace ⋅ Arham Khan ⋅ Khalid Hossain ⋅ Yadu Babuji ⋅ Nicholas Chia ⋅ Venkatram Vishwanath ⋅ Arvind Ramanathan ⋅ Rick Stevens ⋅ Ian Foster ⋅ Robert Underwood		In Room: Mission City Ballroom Poster Location: #60 in Poster Session and Reception - Main Conference.
ReaL: Efficient RLHF Training of Large Language Models with Parameter Reallocation Zhiyu Mei ⋅ WEI FU ⋅ Kaiwei Li ⋅ Guangju Wang ⋅ Huanchen Zhang ⋅ Yi Wu		In Room: Mission City Ballroom Poster Location: #61 in Poster Session and Reception - Main Conference.
DiffServe: Efficiently Serving Text-to-Image Diffusion Models with Query-Aware Model Scaling Sohaib Ahmad ⋅ Qizheng Yang ⋅ Haoliang Wang ⋅ Ramesh Sitaraman ⋅ Hui Guan		In Room: Mission City Ballroom Poster Location: #2 in Poster Session and Reception - Main Conference.
VoLUT: Efficient Volumetric streaming enhanced by LUT-based super-resolution Chendong Wang ⋅ Anlan Zhang ⋅ Yifan Yang ⋅ Lili Qiu ⋅ Yuqing Yang ⋅ XINYANG JIANG ⋅ Feng Qian ⋅ Suman Banerjee		In Room: Mission City Ballroom Poster Location: #14 in Poster Session and Reception - Main Conference.
VoLUT: Efficient Volumetric streaming enhanced by LUT-based super-resolution Chendong Wang ⋅ Anlan Zhang ⋅ Yifan Yang ⋅ Lili Qiu ⋅ Yuqing Yang ⋅ XINYANG JIANG ⋅ Feng Qian ⋅ Suman Banerjee		In Room: Mission City Ballroom Poster Location: #14 in Session 12: Edge and Cloud Systems.
AIOpsLab: A Holistic Framework to Evaluate AI Agents for Enabling Autonomous Clouds Project Page ↗ Yinfang Chen ⋅ Manish Shetty ⋅ Gagan Somashekar ⋅ Minghua Ma ⋅ Yogesh Simmhan ⋅ Jonathan Mace ⋅ Chetan Bansal ⋅ Rujia Wang ⋅ S R		In Room: Mission City Ballroom Poster Location: #4 in Poster Session and Reception - Main Conference.
LAVA: Lifetime-Aware VM Allocation with Learned Distributions and Adaptation to Mispredictions Jianheng Ling ⋅ Pratik Worah ⋅ Yawen Wang ⋅ Yunchuan Kong ⋅ Chunlei Wang ⋅ Clifford Stein ⋅ Diwakar Gupta ⋅ Jason Behmer ⋅ Logan Bush ⋅ Prakash Ramanan ⋅ Rajesh Kumar ⋅ Thomas Chestna ⋅ Yajing Liu ⋅ Ying Liu ⋅ Ye Zhao ⋅ Kathryn S. McKinley ⋅ Meeyoung Park ⋅ Martin Maas		In Room: Mission City Ballroom Poster Location: #8 in Poster Session and Reception - Main Conference.
Photon: Federated LLM Pre-Training Project Page ↗ Lorenzo Sani ⋅ Alex Iacob ⋅ Zeyu Cao ⋅ Royson Lee ⋅ Bill Marino ⋅ Yan Gao ⋅ Wanru Zhao ⋅ Dongqi Cai ⋅ Zexi Li ⋅ Xinchi Qiu ⋅ Nic Lane		In Room: Mission City Ballroom Poster Location: #9 in Poster Session and Reception - Main Conference.
SwiftVI: Time-Efficient Planning and Learning with MDPs Project Page ↗ Kasper Overgaard Mortensen ⋅ Konstantinos Skitsas ⋅ Emil Morre Christensen ⋅ Mohammad Sadegh Talebi ⋅ Andreas Pavlogiannis ⋅ Davide Mottin ⋅ Panagiotis Karras		In Room: Mission City Ballroom Poster Location: #10 in Poster Session and Reception - Main Conference.
LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention Project Page ↗ Shang Yang ⋅ Junxian Guo ⋅ Haotian Tang ⋅ Qinghao Hu ⋅ Guangxuan Xiao ⋅ Jiaming Tang ⋅ Yujun Lin ⋅ Zhijian Liu ⋅ Yao Lu ⋅ Song Han		In Room: Mission City Ballroom Poster Location: #19 in Poster Session and Reception - Main Conference.
Youmu: Efficient Columnar Data Pipeline for LLM Training Tianle Zhong ⋅ Jiechen Zhao ⋅ Qiang Su ⋅ Geoffrey Fox		In Room: Mission City Ballroom Poster Location: #17 in Poster Session and Reception - Main Conference.
FedProphet: Memory-Efficient Federated Adversarial Training via Robust and Consistent Cascade Learning Minxue Tang ⋅ Yitu Wang ⋅ Jingyang Zhang ⋅ Louis DiValentin ⋅ Aolin Ding ⋅ Amin Hass ⋅ Yiran Chen ⋅ Hai Li		In Room: Mission City Ballroom Poster Location: #24 in Poster Session and Reception - Main Conference.
Optimizing LLM Queries in Relational Data Analytics Workloads Shu Liu ⋅ Asim Biswal ⋅ Audrey Cheng ⋅ Amog Kamsetty ⋅ Luis Gaspar Schroeder ⋅ Liana Patel ⋅ Shiyi Cao ⋅ Xiangxi Mo ⋅ Ion Stoica ⋅ Joseph Gonzalez ⋅ Matei Zaharia		In Room: Mission City Ballroom Poster Location: #28 in Poster Session and Reception - Main Conference.
HyC-LoRA: Memory Efficient LoRA Fine-tuning with Hybrid Activation Compression Project Page ↗ Yujin Wang ⋅ Shunan Dong ⋅ Zongle Huang ⋅ Yichen You ⋅ Liu He ⋅ Huazhong Yang ⋅ Yongpan Liu ⋅ Hongyang Jia		In Room: Mission City Ballroom Poster Location: #35 in Poster Session and Reception - Main Conference.
Efficient LLM Inference using Dynamic Input Pruning and Cache-Aware Masking Marco Federici ⋅ Davide Belli ⋅ Mart van Baalen ⋅ Amir Jalalirad ⋅ Andrii Skliar ⋅ Bence Major ⋅ Markus Nagel ⋅ Paul Whatmough		In Room: Mission City Ballroom Poster Location: #38 in Poster Session and Reception - Main Conference.
COMET: Fine-grained Computation-communication Overlapping for Mixture-of-Experts Shulai Zhang ⋅ Ningxin Zheng ⋅ Haibin Lin ⋅ Ziheng Jiang ⋅ Wenlei Bao ⋅ Chengquan Jiang ⋅ Qi Hou ⋅ Weihao Cui ⋅ Size Zheng ⋅ Li-Wen Chang ⋅ Quan Chen ⋅ Xin Liu		In Room: Mission City Ballroom Poster Location: #43 in Poster Session and Reception - Main Conference.
AIOpsLab: A Holistic Framework to Evaluate AI Agents for Enabling Autonomous Clouds Project Page ↗ Yinfang Chen ⋅ Manish Shetty ⋅ Gagan Somashekar ⋅ Minghua Ma ⋅ Yogesh Simmhan ⋅ Jonathan Mace ⋅ Chetan Bansal ⋅ Rujia Wang ⋅ S R		In Room: Mission City Ballroom Poster Location: #4 in Session 4: Reliable and Scalable Systems.
MEADOW: Memory-efficient Dataflow and Data Packing for Low Power Edge LLMs Abhishek Moitra ⋅ Arkapravo Ghosh ⋅ Shrey Agrawal ⋅ Aporva Amarnath ⋅ Karthik Swaminathan ⋅ Priyadarshini Panda		In Room: Mission City Ballroom Poster Location: #45 in Poster Session and Reception - Main Conference.
AI Metropolis: Scaling Large Language Model-based Multi-Agent Simulation with Out-of-order Execution Zhiqiang Xie ⋅ Hao Kang ⋅ Ying Sheng ⋅ Tushar Krishna ⋅ Kayvon Fatahalian ⋅ Christos Kozyrakis		In Room: Mission City Ballroom Poster Location: #46 in Poster Session and Reception - Main Conference.
Venn: Resource Management For Collaborative Learning Jobs Jiachen Liu ⋅ Fan Lai ⋅ Eric Ding ⋅ Yiwen Zhang ⋅ Mosharaf Chowdhury		In Room: Mission City Ballroom Poster Location: #50 in Poster Session and Reception - Main Conference.
LAVA: Lifetime-Aware VM Allocation with Learned Distributions and Adaptation to Mispredictions Jianheng Ling ⋅ Pratik Worah ⋅ Yawen Wang ⋅ Yunchuan Kong ⋅ Chunlei Wang ⋅ Clifford Stein ⋅ Diwakar Gupta ⋅ Jason Behmer ⋅ Logan Bush ⋅ Prakash Ramanan ⋅ Rajesh Kumar ⋅ Thomas Chestna ⋅ Yajing Liu ⋅ Ying Liu ⋅ Ye Zhao ⋅ Kathryn S. McKinley ⋅ Meeyoung Park ⋅ Martin Maas		In Room: Mission City Ballroom Poster Location: #8 in Session 12: Edge and Cloud Systems.
Photon: Federated LLM Pre-Training Project Page ↗ Lorenzo Sani ⋅ Alex Iacob ⋅ Zeyu Cao ⋅ Royson Lee ⋅ Bill Marino ⋅ Yan Gao ⋅ Wanru Zhao ⋅ Dongqi Cai ⋅ Zexi Li ⋅ Xinchi Qiu ⋅ Nic Lane		In Room: Mission City Ballroom Poster Location: #9 in Session 11: Federated Learning.
FlexInfer: Flexible LLM Inference with CPU Computations Seonjin Na ⋅ Geonhwa Jeong ⋅ Byung Hoon Ahn ⋅ Aaron Jezghani ⋅ Jeffrey Young ⋅ Christopher Hughes ⋅ Tushar Krishna ⋅ Hyesoon Kim		In Room: Mission City Ballroom Poster Location: #55 in Poster Session and Reception - Main Conference.
SwiftVI: Time-Efficient Planning and Learning with MDPs Project Page ↗ Kasper Overgaard Mortensen ⋅ Konstantinos Skitsas ⋅ Emil Morre Christensen ⋅ Mohammad Sadegh Talebi ⋅ Andreas Pavlogiannis ⋅ Davide Mottin ⋅ Panagiotis Karras		In Room: Mission City Ballroom Poster Location: #10 in Session 6: Edge and Cloud Systems.
Youmu: Efficient Columnar Data Pipeline for LLM Training Tianle Zhong ⋅ Jiechen Zhao ⋅ Qiang Su ⋅ Geoffrey Fox		In Room: Mission City Ballroom Poster Location: #17 in Session 5: LLM training and fine-tuning.
LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention Project Page ↗ Shang Yang ⋅ Junxian Guo ⋅ Haotian Tang ⋅ Qinghao Hu ⋅ Guangxuan Xiao ⋅ Jiaming Tang ⋅ Yujun Lin ⋅ Zhijian Liu ⋅ Yao Lu ⋅ Song Han		In Room: Mission City Ballroom Poster Location: #19 in Session 7: Quantization and Sparsity.
FedProphet: Memory-Efficient Federated Adversarial Training via Robust and Consistent Cascade Learning Minxue Tang ⋅ Yitu Wang ⋅ Jingyang Zhang ⋅ Louis DiValentin ⋅ Aolin Ding ⋅ Amin Hass ⋅ Yiran Chen ⋅ Hai Li		In Room: Mission City Ballroom Poster Location: #24 in Session 11: Federated Learning.
Optimizing LLM Queries in Relational Data Analytics Workloads Shu Liu ⋅ Asim Biswal ⋅ Audrey Cheng ⋅ Amog Kamsetty ⋅ Luis Gaspar Schroeder ⋅ Liana Patel ⋅ Shiyi Cao ⋅ Xiangxi Mo ⋅ Ion Stoica ⋅ Joseph Gonzalez ⋅ Matei Zaharia		In Room: Mission City Ballroom Poster Location: #28 in Session 6: Edge and Cloud Systems.
Marconi: Prefix Caching for the Era of Hybrid LLMs Project Page ↗ Rui Pan ⋅ Zhuang Wang ⋅ Zhen Jia ⋅ Can Karakus ⋅ Luca Zancato ⋅ Tri Dao ⋅ Yida Wang ⋅ Ravi Netravali		In Room: Mission City Ballroom Poster Location: #29 in Session 10: LLM and Diffusion Model Serving.
HyC-LoRA: Memory Efficient LoRA Fine-tuning with Hybrid Activation Compression Project Page ↗ Yujin Wang ⋅ Shunan Dong ⋅ Zongle Huang ⋅ Yichen You ⋅ Liu He ⋅ Huazhong Yang ⋅ Yongpan Liu ⋅ Hongyang Jia		In Room: Mission City Ballroom Poster Location: #35 in Session 5: LLM training and fine-tuning.
Efficient LLM Inference using Dynamic Input Pruning and Cache-Aware Masking Marco Federici ⋅ Davide Belli ⋅ Mart van Baalen ⋅ Amir Jalalirad ⋅ Andrii Skliar ⋅ Bence Major ⋅ Markus Nagel ⋅ Paul Whatmough		In Room: Mission City Ballroom Poster Location: #38 in Session 7: Quantization and Sparsity.
COMET: Fine-grained Computation-communication Overlapping for Mixture-of-Experts Shulai Zhang ⋅ Ningxin Zheng ⋅ Haibin Lin ⋅ Ziheng Jiang ⋅ Wenlei Bao ⋅ Chengquan Jiang ⋅ Qi Hou ⋅ Weihao Cui ⋅ Size Zheng ⋅ Li-Wen Chang ⋅ Quan Chen ⋅ Xin Liu		In Room: Mission City Ballroom Poster Location: #43 in Session 9: Parallel and Distributed Systems.
MEADOW: Memory-efficient Dataflow and Data Packing for Low Power Edge LLMs Abhishek Moitra ⋅ Arkapravo Ghosh ⋅ Shrey Agrawal ⋅ Aporva Amarnath ⋅ Karthik Swaminathan ⋅ Priyadarshini Panda		In Room: Mission City Ballroom Poster Location: #45 in Session 12: Edge and Cloud Systems.
AI Metropolis: Scaling Large Language Model-based Multi-Agent Simulation with Out-of-order Execution Zhiqiang Xie ⋅ Hao Kang ⋅ Ying Sheng ⋅ Tushar Krishna ⋅ Kayvon Fatahalian ⋅ Christos Kozyrakis		In Room: Mission City Ballroom Poster Location: #46 in Session 4: Reliable and Scalable Systems.
Venn: Resource Management For Collaborative Learning Jobs Jiachen Liu ⋅ Fan Lai ⋅ Eric Ding ⋅ Yiwen Zhang ⋅ Mosharaf Chowdhury		In Room: Mission City Ballroom Poster Location: #50 in Session 11: Federated Learning.