9 papers
FACT: Compositional Kernel Synthesis with a Three-Stage Agentic Workflow
Sina Heidari, Dimitrios S. Nikolopoulos
Deep learning compilers and vendor libraries deliver strong baseline performance but their performance is bounded by finite, engineer-curated catalogs. When these omit needed optim…
LLM-Guided Runtime Parameter Optimization for Energy-Efficient Model Inference
Katelyn Crumpacker, Dimitrios Nikolopoulos
Large Language Models (LLMs) have become an integral part of many real-world workflows. However, LLMs consume a lot of energy, which becomes a large concern in the scale of the dem…
ConfigSpec: Profiling-Based Configuration Selection for Distributed Edge--Cloud Speculative LLM Serving
Xiangchen Li, Saeid Ghafouri, Jiakun Fan +3
Speculative decoding enables collaborative Large Language Model (LLM) inference across cloud and edge by separating lightweight token drafting from heavyweight verification. While…
WISP: Waste- and Interference-Suppressed Distributed Speculative LLM Serving at the Edge via Dynamic Drafting and SLO-Aware Batching
Xiangchen Li, Jiakun Fan, Qingyuan Wang +7
As Large Language Models (LLMs) become increasingly accessible to end users, an ever-growing number of inference requests are initiated from edge devices and computed on centralize…
Taming the Memory Footprint Crisis: System Design for Production Diffusion LLM Serving
Jiakun Fan, Yanglin Zhang, Xiangchen Li +1
Diffusion Large Language Models (dLLMs) have emerged as a promising alternative to Autoregressive Models (ARMs), utilizing parallel decoding to overcome sequential bottlenecks. How…
APEX: Asynchronous Parallel CPU-GPU Execution for Online LLM Inference on Constrained GPUs
Jiakun Fan, Yanglin Zhang, Xiangchen Li +1
Deploying large language models (LLMs) for online inference is often constrained by limited GPU memory, particularly due to the growing KV cache during auto-regressive decoding. Hy…