Frontier is a new discrete-event simulator for disaggregated LLM serving that incorporates co-location, PDD, AFD, and optimizations, achieving under 4% throughput error and large reductions in latency prediction error versus prior simulators.
Galvatron: Efficient transformer train- ing over multiple gpus using automatic parallelism
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 8verdicts
UNVERDICTED 8roles
background 2polarities
background 2representative citing papers
HexAGenT reduces the SLO scale required for timely agentic LLM workflow completion by an average of 20.1% at 95% attainment and 33.0% at 99% attainment on heterogeneous A100/H100/H200 clusters.
FEPLB reduces token and GEMM stragglers in MoE training by 50-70% using nearly free Copy Engine communication on Hopper architecture.
Autopoiesis uses LLM-driven program synthesis to evolve serving policies online during deployment, delivering up to 53% and average 34% gains over prior LLM serving systems under runtime dynamics.
LiveR enables live reconfiguration for elastic LLM training by asynchronously preparing new parallel worlds and streaming reshaped model state over interconnects, reducing downtime to seconds and achieving 14-23x faster reconfiguration than checkpoint/restart.
HexiSeq optimizes sequence and head partitioning across mixed GPUs to improve long-context LLM training throughput by up to 1.72x in simulations.
FlowTrain decouples encoder and backbone training via flow-based scheduling and heterogeneous allocation to reach over 50% MFU and 1.7x throughput versus prior VLM approaches.
Piper decouples user-defined distributed training strategies from runtime execution using transformations on a unified global training DAG IR, achieving parity on ZeRO and gains on composed strategies like DualPipe.
citing papers explorer
-
Frontier: Towards Comprehensive and Accurate LLM Inference Simulation
Frontier is a new discrete-event simulator for disaggregated LLM serving that incorporates co-location, PDD, AFD, and optimizations, achieving under 4% throughput error and large reductions in latency prediction error versus prior simulators.
-
HexAGenT: Efficient Agentic LLM Serving via Workflow- and Heterogeneity-Aware Scheduling
HexAGenT reduces the SLO scale required for timely agentic LLM workflow completion by an average of 20.1% at 95% attainment and 33.0% at 99% attainment on heterogeneous A100/H100/H200 clusters.
-
FEPLB: Exploiting Copy Engines for Nearly Free MoE Load Balancing in Distributed Training
FEPLB reduces token and GEMM stragglers in MoE training by 50-70% using nearly free Copy Engine communication on Hopper architecture.
-
Autopoiesis: A Self-Evolving System Paradigm for LLM Serving Under Runtime Dynamics
Autopoiesis uses LLM-driven program synthesis to evolve serving policies online during deployment, delivering up to 53% and average 34% gains over prior LLM serving systems under runtime dynamics.
-
LiveR: Fine-Grained Elasticity via Live Reconfiguration for Model Training
LiveR enables live reconfiguration for elastic LLM training by asynchronously preparing new parallel worlds and streaming reshaped model state over interconnects, reducing downtime to seconds and achieving 14-23x faster reconfiguration than checkpoint/restart.
-
HexiSeq: Accommodating Long Context Training of LLMs over Heterogeneous Hardware
HexiSeq optimizes sequence and head partitioning across mixed GPUs to improve long-context LLM training throughput by up to 1.72x in simulations.
-
FlowTrain: Flow-Based Decoupled Training for Industrial-Grade Vision-Language Models
FlowTrain decouples encoder and backbone training via flow-based scheduling and heterogeneous allocation to reach over 50% MFU and 1.7x throughput versus prior VLM approaches.
-
Piper: A Programmable Distributed Training System
Piper decouples user-defined distributed training strategies from runtime execution using transformations on a unified global training DAG IR, achieving parity on ZeRO and gains on composed strategies like DualPipe.