Kaushik Roy

Utkarsh Saxena, Gobinda Saha, Sakshi Choudhary et al.

Large language models (LLMs) represent a groundbreaking advancement in the domain of natural language processing due to their impressive reasoning abilities. Recently, there has been considerable interest in increasing the context lengths for these models to enhance their applicability to complex tasks. However, at long context lengths and large batch sizes, the key-value (KV) cache, which stores the attention keys and values, emerges as the new bottleneck in memory usage during inference. To address this, we propose Eigen Attention, which performs the attention operation in a low-rank space, thereby reducing the KV cache memory overhead. Our proposed approach is orthogonal to existing KV cache compression techniques and can be used synergistically with them. Through extensive experiments over OPT, MPT, and Llama model families, we demonstrate that Eigen Attention results in up to 40% reduction in KV cache sizes and up to 60% reduction in attention operation latency with minimal drop in performance. Code is available at https://github.com/UtkarshSaxena1/EigenAttn.

Jeffry Victor, Dong Eun Kim, Chunguang Wang et al.

Deep neural network (DNN) accelerators employing crossbar arrays capable of in-memory computing (IMC) are highly promising for neural computing platforms. However, in deeply scaled technologies, interconnect resistance severely impairs IMC robustness, leading to a drop in the system accuracy. To address this problem, we propose SWANN - a technique based on shuffling weights in crossbar arrays which alleviates the detrimental effect of wire resistance on IMC. For 8T-SRAM-based 128x128 crossbar arrays in 7nm technology, SWANN enhances the accuracy from 47.78% to 83.5% for ResNet-20/CIFAR-10. We also show that SWANN can be used synergistically with Partial-Word-LineActivation, further boosting the accuracy. Moreover, we evaluate the implications of SWANN for compact ferroelectric-transistorbased crossbar arrays. SWANN incurs minimal hardware overhead, with less than a 1% increase in energy consumption. Additionally, the latency and area overheads of SWANN are ~1% and ~16%, respectively when 1 ADC is utilized per crossbar array.