Matrix-vector multiplication (MVM) and content-based search are two key operations in many machine learning workloads. This article proposes a ferroelectric FET (FeFET) time-domain compute-in-memory (TD-CiM) array tha...
详细信息
Matrix-vector multiplication (MVM) and content-based search are two key operations in many machine learning workloads. This article proposes a ferroelectric FET (FeFET) time-domain compute-in-memory (TD-CiM) array that can accelerate both operations in a homogeneous fabric. We demonstrate that 1) the AND and XOR/XNOR logic functions required by MVM and content-based search can be realized using a single compute-in-memory (CiM) cell composed of 2FeFETs connected in series;2) an inverter chain-based TD-CiM array along with a two-phase time-domain computation principle of the TD-CiM can be employed to implement the MVM and content-based search functions;3) a signal delay-to-digital output conversion can be implemented by associating a loading capacitor with each stage of the inverter chain-based TD-CiM array, ensuring the full digital compatibility;and 4) the proposed 2FeFET cell and inverter chain-based TD-CiM array are robust against FeFET variation according to our comprehensive theoretical and experimental validation. We show how the FeFET TD-CiM can be exploited to accelerate hyperdimensional computing (HDC) and adjusted to process different tasks through dynamic and fine-grained resource allocation. HDC application benchmarking results show that the proposed FeFET-based TD-CiM offers on average 106x/63x energy reduction/speedup compared to GPU-based implementation. With more than 8500 TOPS/W energy-efficiency, the proposed FeFET-based TD-CiM exhibits huge potential as a processing fabric for various memory-intensive applications.
In this work the design and implementation of a fully integrated ISO 11784/5 RFID reader front-end ASIC is presented, for both HDX and FDX transponder types. The ASIC was designed in a 0.18 μm. CMOS-HV technology, an...
详细信息
Finite Impulse Response (FIR) filters are essential components in digital signalprocessing, with their efficiency significantly influenced by the design of multipliers and adders. This work presents a novel Verilog-b...
详细信息
There are many approaches to testbed systems in computer networks, but especially when evaluating or developing new security protocols and testing their behavior when used via satellite networks, many challenges remai...
详细信息
Slow Start increases the congestion window by doubling each round trip until congestion is detected. As a result, reaching optimal speed typically takes several round trip times, especially important in high bandwidth...
详细信息
This paper presents the design and implementation of a low-power MIPS32 processor with an integrated cache system, leveraging clock gating techniques to optimize power efficiency on Field-Programmable Gate Arrays (FPG...
详细信息
In this paper, we investigate reconfigurable intelligent surface (RIS)-assisted integrated sensing and communication (ISAC) systems for robust physical layer security (PLS) schemes. Traditionally, eavesdroppers (Eves)...
详细信息
This simulation study aims to compare the physical layer performance of the 5th Generation (5G) New Radio Non-Terrestrial Networks (NR NTN) air interface with Digital Video Broadcasting (DVB) technologies, specificall...
详细信息
This paper studies integration issues for the achievement of TN-NTN systems providing mission-critical (MCX) services in the EU 5G-GOVSATCOM project. The standardization of NTN as an integral part of 5G terrestrial sy...
详细信息
Space-based radars (SBRs) systems are able to provide an unobstructed field of view for space target detection and tracking. However, the large temperature dynamic range and poor heat dissipation performance of the SB...
详细信息
Space-based radars (SBRs) systems are able to provide an unobstructed field of view for space target detection and tracking. However, the large temperature dynamic range and poor heat dissipation performance of the SBR system cause severe thermal noise, leading to deficiency in distant or dim space target detection tasks. In essence, the challenges above can be categorized as typical low signal-to-noise ratio (SNR) problems, and the track before detect (TBD) processing scheme is applied to solve them in this article. Nevertheless, the typical TBD methods reckon without the following aspects and thus are not well compatible with space target surveillance tasks via the SBR system. First, the typical TBD methods discard the phase information of radar raw data in constructing the likelihood ratio. In addition, most existing work merely considers modeling the amplitude fluctuation as Swerling types, which is not accurate enough for space targets when compared with the log-normal distribution (LND) model. Moreover, orbital space targets follow the orbital dynamic principle while most existing TBD methods neglect this important information, which will cause space targets filtering estimation bias. To address the aforementioned problems, we propose a TBD method based on the complex-amplitude likelihood ratio (CLR) of the LND model and soft orbit-information constraint (OC). In this article, with the aim of acquiring a more accurate likelihood ratio, we first derive the closed mathematical form of the amplitude likelihood ratio (ALR) and the CLR of the LND model. Meanwhile, some approximations are proposed to alleviate the integral computation. Then, the proposed ALR and CLR of the LND model are utilized to be implemented into the TBD scheme. Finally, we design elegant soft OC strategies to modify the associated weights corresponding with birth particles in sequential Monte Carlo (SMC) implementation. Simulation results are provided to validate the effectiveness of the propo
暂无评论