Related papers: Large-Scale Integrated Vector-Matrix Multiplicatio…
Large language model (LLM) inference has been a prevalent demand in daily life and industries. The large tensor sizes and computing complexities in LLMs have brought challenges to memory, computing, and databus. This paper proposes a…
While Mixture of Experts (MoE) models achieve remarkable efficiency by activating only subsets of parameters, they suffer from high memory access costs during inference. Memory-layer architectures offer an appealing alternative with very…
Two-dimensional layered semiconductors have recently emerged as attractive building blocks for next-generation low-power non-volatile memories. However, challenges remain in the controllable sub-micron fabrication of bipolar resistively…
Monolayer MoS$_2$ is a direct band gap semiconductor with potential applications in optoelectronics and photonics. MoS$_2$ also has a large optical nonlinearity. However, the atomic thickness of the monolayer limits the strength of the…
Light-matter interaction with two-dimensional materials gained significant attention in recent years leading to the reporting of weak and strong coupling regimes, and effective nano-laser operation with various structures. Particularly,…
One-dimensional (1D) nanostructures - nanowires (NWs) - exhibit promising properties for integration in different types of functional devices. Their properties can be enhanced even further or tuned for a specific application by combining…
We have implemented fast Fourier transforms for one, two, and three-dimensional arrays on the Cerebras CS-2, a system whose memory and processing elements reside on a single silicon wafer. The wafer-scale engine (WSE) encompasses a…
Recently emerging large-area single-layer MoS2 grown by chemical vapor deposition has triggered great interest due to its exciting potential for applications in advanced electronic and optoelectronic devices. Unlike gapless graphene, MoS2…
Atomically thin two-dimensional (2D) semiconductors such as molybdenum disulphide (MoS2) hold great promise in electrical, optical, and mechanical devices and display novel physical phenomena such as coupled spin-valley physics and the…
We introduce a high-throughput neural network accelerator that embeds most network layers directly in hardware, minimizing data transfer and memory usage while preserving a degree of flexibility via a small neural processing unit for the…
The electronic structure of an atomic-layer-deposited MoS2 monolayer on SiO2 was investigated using X-ray absorption spectroscopy (XAS) and synchrotron X-ray photoelectron spectroscopy (XPS). The angle-dependent evolution of the XAS spectra…
While new species and properties of two-dimensional (2D) materials are being reported with extraordinary regularity, a significant bottleneck in the field is the ability to controllably process material into working devices. We report a…
Van der Waals (vdW) semiconductors are attractive for highly scaled devices and heterogeneous integration since they can be isolated into self-passivated, two-dimensional (2D) layers that enable superior electrostatic control. These…
A machine-learned interatomic potential (MLIP) for multilayer MoS2 was developed using the ultra-fast force field (UF3) framework. The UF3 MLIP reproduces key properties in strong agreement with DFT including lattice constants, interlayer…
The van der Waals heterostructures are a fertile frontier for discovering emergent phenomena in condensed matter systems. They are constructed by stacking elements of a large library of two-dimensional materials, which couple together…
Neural networks are increasingly used in real-time systems, such as automated driving applications. This requires high-performance hardware with predictable timing behavior. State-of-the-art real-time hardware is limited in memory and…
Starting from graphene, 2D layered materials family has been recently set up more than 100 different materials with variety of different class of materials such as semiconductors, metals, semimetals, superconductors. Among these materials,…
Non-volatile resistive switching, also known as memristor effect in two terminal devices, has emerged as one of the most important components in the ongoing development of high-density information storage, brain-inspired computing, and…
Compared to the first generation of deep neural networks, dominated by regular, compute-intensive kernels such as matrix multiplications (MatMuls) and convolutions, modern decoder-based transformers interleave attention, normalization, and…
Memristor-based neuromorphic computing could overcome the limitations of traditional von Neumann computing architectures -- in which data are shuffled between separate memory and processing units -- and improve the performance of deep…