conference

When Application-Specific ISA Meets FPGAs: A Multi-LayerVirtualization Framework for Heterogeneous Cloud FPGAs

The paper is the first to propose a new virtualization mechanism that can natively support scale-out acceleration across **heterogeneous** cloud FPGAs and demonstrate its effectiveness on a custom FPGA cluster with heterogeneous FPGA resources. Cloud …

Hyper-AP: Enhancing Associative Processing Through A Full-Stack Optimization

3D-stacking memory technology such as High-Bandwidth Memory (HBM) and Hybrid Memory Cube (HMC) provides orders of magnitude more bandwidth and significantly increased channel-level parallelism (CLP) due to the new parallel memory architecture. …

ViTAL: Virtualizing FPGAs in the Cloud

Field-Programmable Gate Arrays (FPGAs) have been integrated into the cloud infrastructure to enhance its computing performance by supporting on-demand acceleration. However, system support for FPGAs in the context of the cloud environment is still in …

Liquid Silicon: A Nonvolatile Fully Programmable Processing-In-Memory Processor with Monolithically Integrated ReRAM for Big Data/Machine Learning Applications

A nonvolatile fully programmable processing-in-memory (PIM) processor named Liquid Silicon (L-Si) is demonstrated, which combines the superior programmability of general-purpose computing devices (e.g. FPGA) and the high power efficiency of …

Nb(1-x)O2 based Universal Selector with Ultra-high Endurance (>10^12), high speed (10ns) and Excellent Vth Stability

MEG: A RISCV-based system simulation infrastructure for exploring memory optimization using FPGAs and Hybrid Memory Cube (Best Paper Nominee)

Emerging 3D memory technologies, such as the Hybrid Memory Cube (HMC) and High Bandwidth Memory (HBM), provide increased bandwidth and massive memory-level parallelism. Efficiently integrating emerging memories into existing system pose new …

Unleashing the Power of Soft Logic for Convolutional Neural Network Acceleration via Product Quantization (Poster)

Efficient Large-scale Approximate Nearest Neighbor Search on the OpenCL-FPGA

We present a new method for Product Quantization (PQ) based approximated nearest neighbor search (ANN) in high dimensional spaces. Specifically, we first propose a quantization scheme for the codebook of coarse quantizer, product quantizer, and …

PQ-CNN: Accelerating Product Quantized Convolutional Neural Network (Poster)

Adaptive Quantization of Neural Networks

Despite the state-of-the-art accuracy of Deep Neural Networks (DNN) in various classification problems, their deployment onto resource constrained edge computing devices remains challenging due to their large size and complexity. Several recent …