9 papers
When Does Disaggregation Pay? Simulating Prefill--Decode--Attention--FFN Specialization for Agentic LLM Inference
Przemyslaw Forys, Haoran Wu, Can Xiao +8
Agentic inference now dominates the LLM inference landscape, requiring LLMs to actively engage in multi-turn interactions with tool-calling capabilities. This introduces a more com…
CascadeLUT: Information-Ordered Streaming Inference for Bandwidth-Constrained FPGAs
Oliver Cassidy, Marta Andronic, George A. Constantinides
Mapping neural networks to FPGAs enables low-latency, energy-efficient inference, particularly for lookup table (LUT)-based models that eliminate multipliers and map directly to re…
Direction-Preserving Number Representations
Bardia Zadeh, George A. Constantinides
Low-precision number formats are widely used in modern machine learning systems due to their efficiency. Accurate direction representation is key to the accuracy of vector operatio…
Training with Fewer Bits: Unlocking Edge LLMs Training with Stochastic Rounding
Taowen Liu, Marta Andronic, Deniz Gündüz +1
LLM training is resource-intensive. Quantized training improves computational and memory efficiency but introduces quantization noise, which can hinder convergence and degrade mode…
NeuraLUT-Assemble: Hardware-aware Assembling of Sub-Neural Networks for Efficient LUT Inference
Marta Andronic, George A. Constantinides
Efficient neural networks (NNs) leveraging lookup tables (LUTs) have demonstrated significant potential for emerging AI applications, particularly when deployed on field-programmab…
Banked Memories for Soft SIMT Processors
Martin Langhammer, George A. Constantinides
Recent advances in soft GPGPU architectures have shown that a small (<10K LUT), high performance (770 MHz) processor is possible in modern FPGAs. In this paper we architect and eva…