SearcharxivSearch

arXiv subjects

Hetian Chen

Publications and source records attributed to Hetian Chen.

5 recordsLinked to original sources

ATCC: Adaptive Concurrency Control for Unforeseen Agentic Transactions

Data agents, empowered by Large Language Models (LLMs), introduce a new paradigm in transaction processing. Unlike traditional applications with fixed patterns, data agents run online-generated workflows that repeatedly issue SQL statements, reason over intermediate results, and revise subsequent plans. To ensure data consistency, these SQL statements issued by an agent should be integrated into a transaction, referred to as agentic transactions. Agentic transactions exhibit unforeseen characteristics, including long execution times, irregular execution intervals, and non-deterministic access patterns, breaking the assumptions underlying concurrency control (CC) (e.g., short-lived, predefined). Traditional CC schemes, which rely on fixed policies, fail to capture such dynamic behavior, resulting in inadequate performance. This paper introduces ATCC, an adaptive Concurrency Control for Agentic Transactions. ATCC continuously monitors and interprets the runtime behavior of each agentic transaction, evaluates its interactive phases, and dynamically adapts optimistic or pessimistic execution for each transaction. To ensure precise timing for adaptive switches, ATCC employs a reinforcement learning-based policy to balance immediate blocking against future abort costs. Additionally, to mitigate contention-induced tail latency and wasted reasoning cost caused by abort, a cost-aware priority-based lock scheduling is integrated to prioritize expensive or latency-sensitive transactions. Experimental results under agentic-like YCSB and TPC-C workloads demonstrate that ATCC improves the throughput of agentic transactions by up to four orders of magnitude and reduces tail latency by up to 90% compared to state-of-the-art CC schemes.

cs.DB

Electrical Modulation and Probing of Antiferromagnetism in Hybrid Multiferroic Heterostructures

The unique features of ultrafast spin dynamics and the absence of macroscopic magnetization in antiferromagnetic (AFM) materials provide a distinct route towards high-speed magnetic storage devices with low energy consumption and high integration density. However, these advantages also introduce challenges in probing and controlling AFM order, thereby restricting their practical applications. In this study, we demonstrate an all-electric control and probing of the AFM order in heavy metal (HM)/AFM insulator (AFMI) heterostructures on a ferroelectric substrate at room temperature (RT). The AFM order was detected by the anomalous Hall effect (AHE) and manipulated by the ferroelectric field effect as well as the piezoelectric effect in heterostructures of Pt/NiO/0.7Pb(Mg$_{1/3}$Nb$_{2/3}$)O$_{3}$--0.3PbTiO$_{3}$ (PMN--PT). The non-volatile control of AFM order gives rise to a 33\% modulation of AHE, which is further evidenced by synchrotron-based X-ray magnetic linear dichroism (XMLD). Combined with the $in$-$situ$ piezoelectric response of AHE, we demonstrate that ferroelectric polarization contributes mainly to the control of the AFM order. Our results are expected to have broader implications for efficient spintronic devices.

cond-mat.mtrl-sci

Substrate-induced spin-torque-like signal in spin-torque ferromagnetic resonance measurement

Oxide thin films and interfaces with strong spin-orbit coupling have recently shown exceptionally high charge-to-spin conversion, making them potential spin-source materials for spintronics. Epitaxial strain engineering using oxide substrates with different lattice constants and symmetries has emerged as a mean to further enhance charge-to-spin conversion. However, high relative permittivity and dielectric loss of commonly used oxide substrates, such as SrTiO3, can cause significant current shunting in substrates at high frequency, which may strongly affect spin-torque measurement and potentially result in an inaccurate estimation of charge-to-spin conversion efficiency. In this study, we systematically evaluate the influence of various oxide substrates for the widely-used spin-torque ferromagnetic resonance (ST-FMR) measurement. Surprisingly, we observed substantial spin-torque signals in samples comprising only ferromagnetic metal on oxide substrates with high relative permittivity (e.g., SrTiO3 and KTaO3), where negligible signal should be initially expected. Notably, this unexpected signal shows a strong correlation with the capacitive reactance of oxide substrates and the leakage radio frequency (RF) current within the substrate. By revising the conventional ST-FMR analysis model, we attribute this phenomenon to a 90-degree phase difference between the RF current flowing in the metal layer and in the substrate. We suggest that extra attention should be paid during the ST-FMR measurements, as this artifact could dominate over the real spin-orbit torque signal from high-resistivity spin-source materials grown on substrate with high relative permittivity.

cond-mat.mtrl-sci

Multiferroic Magnon Spin-Torque Based Reconfigurable Logic-In-Memory

Magnons, bosonic quasiparticles carrying angular momentum, can flow through insulators for information transmission with minimal power dissipation. However, it remains challenging to develop a magnon-based logic due to the lack of efficient electrical manipulation of magnon transport. Here we present a magnon logic-in-memory device in a spin-source/multiferroic/ferromagnet structure, where multiferroic magnon modes can be electrically excited and controlled. In this device, magnon information is encoded to ferromagnetic bits by the magnon-mediated spin torque. We show that the ferroelectric polarization can electrically modulate the magnon spin-torque by controlling the non-collinear antiferromagnetic structure in multiferroic bismuth ferrite thin films with coupled antiferromagnetic and ferroelectric orders. By manipulating the two coupled non-volatile state variables (ferroelectric polarization and magnetization), we further demonstrate reconfigurable logic-in-memory operations in a single device. Our findings highlight the potential of multiferroics for controlling magnon information transport and offer a pathway towards room-temperature voltage-controlled, low-power, scalable magnonics for in-memory computing.

physics.app-ph

Significant Unconventional Anomalous Hall Effect in Heavy Metal/Antiferromagnetic Insulator Heterostructures

The anomalous Hall effect (AHE) is a quantum coherent transport phenomenon that conventionally vanishes at elevated temperatures because of thermal dephasing. Therefore, it is puzzling that the AHE can survive in heavy metal (HM)/antiferromagnetic (AFM) insulator (AFMI) heterostructures at high temperatures yet disappears at low temperatures. In this paper, we report that an unconventional high-temperature AHE in HM/AFMI is observed only around the Néel temperature of AFM, with large anomalous Hall resistivity up to 40 n$Ω$ cm. This mechanism is attributed to the emergence of a noncollinear AFM spin texture with a non-zero net topological charge. Atomistic spin dynamics simulation shows that such a unique spin texture can be stabilized by the subtle interplay among the collinear AFM exchange coupling, interfacial Dyzaloshinski-Moriya interaction, thermal fluctuation, and bias magnetic field.

cond-mat.mes-hall