5 papers
Compute-in-Memory Attention: A Time-Domain Analog Softmax Circuit with RC-Tunable Temperature
Ankur Singh, Ashish Gautam, Shruti R. Kulkarni +1
Softmax is a key operation in Transformer attention, but its exponentiation and normalization add significant overhead in compute-in-memory (CIM) accelerators, especially when anal…
OptiLookUp: An Optical ROM-Based Lookup Table Engine for Photonic Accelerators
Ankur Singh, Akhilesh Jaiswal
Read-only memory (ROM) provides deterministic access to predefined data mappings. Extending ROM concepts to the optical domain enables high-bandwidth, low-latency, and parallel mem…
GEM3D CIM General Purpose Matrix Computation Using 3D Integrated SRAM eDRAM Hybrid Compute In Memory on Memory Architecture
Subhradip Chakraborty, Ankur Singh, Akhilesh R. Jaiswal
With the rapid growth of deep neural networks (DNNs), compute-in-memory (CIM) has emerged as a promising energy-efficient paradigm for accelerating multiply-and-accumulate (MAC) op…
NVM-in-Cache: Repurposing Commodity 6T SRAM Cache into NVM Analog Processing-in-Memory Engine using a Novel Compute-on-Powerline Scheme
Subhradip Chakraborty, Ankur Singh, Xuming Chen +2
The rapid growth of deep neural network (DNN) workloads has significantly increased the demand for large-capacity on-chip SRAM in machine learning (ML) applications, with SRAM arra…
A Pathway to Near Tissue Computing through Processing-in-CTIA Pixels for Biomedical Applications
Zihan Yin, Subhradip Chakraborty, Ankur Singh +3
Near-tissue computing requires sensor-level processing of high-resolution images, essential for real-time biomedical diagnostics and surgical guidance. To address this need, we int…