2 papers
cs.AR2025
Tasa: Thermal-aware 3D-Stacked Architecture Design with Bandwidth Sharing for LLM Inference
Siyuan He, Peiran Yan, Yandong He +2
The autoregressive decoding in LLMs is the major inference bottleneck due to the memory-intensive operations and limited hardware bandwidth. 3D-stacked architecture is a promising…
cs.AR2025
LP-Spec: Leveraging LPDDR PIM for Efficient LLM Mobile Speculative Inference with Architecture-Dataflow Co-Optimization
Siyuan He, Zhantong Zhu, Yandong He +1
LLM inference on mobile devices faces extraneous challenges due to limited memory bandwidth and computational resources. To address these issues, speculative inference and processi…