Congratulations to Kangbo and Hongou (senior-year undergrad). We will present our chiplet-level architecture modeling and explorations for agentic LLM and optical interconnects.

Congratulations to Siyuan and Wenjie. We will present both near-memory (3D-based and LPDDR-based) and in-memory (SRAM-based) acceleration solutions with co-optimization considerations with algorithm, thermal, and testing.

Congratulations to Kangbo, Yiyang and Wentao. We will present three accelerator designs for Multimodal LLM, 3D Gaussian Splatting, and LLM inference with 3D stacked memory.