18 citations · 23 across the 10 of their papers we have counts for
8 papers · 1 filter
Purlin: Separating Orchestration from the Datapath of Collectives
Osayamen Jonathan Aimuyo, Swapnil Gandhi, Christos Kozyrakis
Distributed inference depends on GPU collective communication that must keep pace with evolving hardware and specialized workloads. However, existing works often couple the semanti…
BOOST: Concurrent Access to Host Memory and HBM to Accelerate LLM Inference
Anish Saxena, Jae Hyung Ju, Hritvik Taneja +4
GPU memory bandwidth and capacity limit throughput in large language model (LLM) inference. The GPU memory system consists of a primary tier of high-bandwidth memory (HBM) and a se…
HiSparse: Scaling Sparse-Attention Decoding with Hierarchical KV Cache Management
Zhiqiang Xie, Zhangheng Huang, Tingwei Huang +3
Top-k sparse attention makes long-context LLM decoding cheap to compute: each step reads only a few thousand selected KV entries rather than the full context. Serving systems, howe…
FailSafe: High-performance Resilient Serving
Ziyi Xu, Zhiqiang Xie, Swapnil Gandhi +1
Tensor parallelism (TP) enables large language models (LLMs) to scale inference efficiently across multiple GPUs, but its tight coupling makes systems fragile: a single GPU failure…
Strata: Hierarchical Context Caching for Long Context Language Model Serving
Zhiqiang Xie, Ziyi Xu, Mark Zhao +5
Large Language Models (LLMs) with expanding context windows face significant performance hurdles. While caching key-value (KV) states is critical for avoiding redundant computation…
SYMI: Efficient Mixture-of-Experts Training via Model and Optimizer State Decoupling
Athinagoras Skiadopoulos, Mark Zhao, Swapnil Gandhi +3
Mixture-of-Experts (MoE) models have become a widely-adopted solution to continue scaling model sizes without a corresponding linear increase in compute. During MoE model training,…
Sparse Checkpointing for Fast and Reliable MoE Training
Swapnil Gandhi, Christos Kozyrakis
As large language models scale, training them requires thousands of GPUs over extended durations--making frequent failures an inevitable reality. While checkpointing remains the pr…
ReCycle: Resilient Training of Large DNNs using Pipeline Adaptation
Swapnil Gandhi, Mark Zhao, Athinagoras Skiadopoulos +1
Training large Deep Neural Network (DNN) models requires thousands of GPUs over the course of several days or weeks. At this scale, failures are frequent and can have a big impact…