4 papers
CIERA: Cross-Iteration Exponent Reuse for Lossless Allgather in Sharded MoE Training
Ali Zafar Sadiq, Haiying Shen, Masahiro Tanaka
In training Mixture-of-Experts (MoE) models, sharded data parallelism partitions each expert's parameters across GPUs, requiring an Allgather operation to reconstruct the full weig…
Mitigating KV Cache Competition to Enhance User Experience in LLM Inference
Haiying Shen, Tanmoy Sen, Masahiro Tanaka
In Large Language Model (LLM) serving, the KV-cache (KVC) bottleneck causes high tail Time-to-First-Token (TTFT) and Time-Between-Tokens (TBT), impairing user experience, particula…
AccelGen: Heterogeneous SLO-Guaranteed High-Throughput LLM Inference Serving for Diverse Applications
Haiying Shen, Tanmoy Sen
In this paper, we consider a mixed-prompt scenario for a large language model (LLM) inference serving system that supports diverse applications with both short prompts and long pro…
EconoServe: Maximizing Multi-Resource Utilization with SLO Guarantees in LLM Serving
Haiying Shen, Tanmoy Sen
As Large Language Models (LLMs) continue to grow, reducing costs and alleviating GPU demands has become increasingly critical. However, existing schedulers primarily target either…