1

Complex Tensor Core: Software-Hardware Co-Design for Accelerating Complex-Valued Neural Networks on GPUs
Modern data center solid state drives (SSDs) integrate multiple general-purpose embedded cores to manage flash translation layer, …
PDF Cite Project Poster Slides DOI Slide Show
Complex Tensor Core: Software-Hardware Co-Design for Accelerating Complex-Valued Neural Networks on GPUs
Make Every Batch Count: Fault Entry Merging for Efficient Batching in Unified Virtual Memory
Modern data center solid state drives (SSDs) integrate multiple general-purpose embedded cores to manage flash translation layer, …
PDF Cite Project Poster Slides DOI Slide Show
Make Every Batch Count: Fault Entry Merging for Efficient Batching in Unified Virtual Memory
Performance Characterization and Optimization of LLM Inference on Tenstorrent AI Accelerators
Sparse matrix-vector multiplication (SpMV) is a fundamental operation across diverse domains, including scientific computing, machine …
PDF Cite Project Slides DOI Slide Show
Performance Characterization and Optimization of LLM Inference on Tenstorrent AI Accelerators
SumcheckPIM: An Efficient HBM-Based PIM Architecture for Linear Complexity Zero Knowledge Proofs
Zero-knowledge proofs (ZKPs) are emerging as a core technology for privacy-preserving computation. Despite steady progress in protocol …
PDF Cite Project Slides DOI Slide Show
SumcheckPIM: An Efficient HBM-Based PIM Architecture for Linear Complexity Zero Knowledge Proofs
Three Birds, One Stone: Fast, Accurate-aware and Cost-Efficient Accelerator for Ternary LLM
On-device LLM inference is increasingly important for latency- and privacy-sensitive applications, yet it remains challenging due to …
PDF Cite Project Slides DOI Slide Show
Three Birds, One Stone: Fast, Accurate-aware and Cost-Efficient Accelerator for Ternary LLM
FINEA: An Efficient Neural Network Accelerator Exploiting Factorized Input Features
Modern deep neural network (DNN) models increasingly adopt quantized data formats to alleviate the computational burdens of convolution …
PDF Cite Project Slides DOI Slide Show
FINEA: An Efficient Neural Network Accelerator Exploiting Factorized Input Features
HALO: Hybrid Systolic Arrays via Logical Partitioning for Acceleration of Complex-Valued Neural Networks
Complex-Valued Neural Networks (CVNNs) are an emerging class of deep learning models that process data with both real and imaginary …
PDF Cite Project Slides DOI Slide Show Poster
HALO: Hybrid Systolic Arrays via Logical Partitioning for Acceleration of Complex-Valued Neural Networks
SSFFT: Energy-Efficient Selective Scaling for Fast Fourier Transform in Embedded GPUs
Fast Fourier Transform (FFT) is critical in applications such as signal processing, communications, and AI. Embedded GPUs are often …
PDF Cite Project Slides DOI Slide Show
SSFFT: Energy-Efficient Selective Scaling for Fast Fourier Transform in Embedded GPUs
SparsePIM: An Efficient HBM-Based PIM Architecture for Sparse Matrix-Vector Multiplications
Sparse matrix-vector multiplication (SpMV) is a fundamental operation across diverse domains, including scientific computing, machine …
PDF Cite Project Slides DOI Slide Show
SparsePIM: An Efficient HBM-Based PIM Architecture for Sparse Matrix-Vector Multiplications
Hierarchical Traversal Stack Design Using Shared Memory for GPU Ray Tracing
Ray tracing is widely used to generate photorealistic images by tracing the paths of light rays through a scene and their interactions …
PDF Cite Project Slides DOI Slide Show
Hierarchical Traversal Stack Design Using Shared Memory for GPU Ray Tracing