arXiv subjects

Zhenwei Cao

Publications and source records attributed to Zhenwei Cao.

5 recordsLinked to original sources

arXiv2025

PWDFT-SW: Extending the Limit of Plane-Wave DFT Calculations to 16K Atoms on the New Sunway Supercomputer

First-principles density functional theory (DFT) with plane wave (PW) basis set is the most widely used method in quantum mechanical material simulations due to its advantages in accuracy and universality. However, a perceived drawback of PW-based DFT calculations is their substantial computational cost and memory usage, which currently limits their ability to simulate large-scale complex systems containing thousands of atoms. This situation is exacerbated in the new Sunway supercomputer, where each process is limited to a mere 16 GB of memory. Herein, we present a novel parallel implementation of plane wave density functional theory on the new Sunway supercomputer (PWDFT-SW). PWDFT-SW fully extracts the benefits of Sunway supercomputer by extensively refactoring and calibrating our algorithms to align with the system characteristics of the Sunway system. Through extensive numerical experiments, we demonstrate that our methods can substantially decrease both computational costs and memory usage. Our optimizations translate to a speedup of 64.8x for a physical system containing 4,096 silicon atoms, enabling us to push the limit of PW-based DFT calculations to large-scale systems containing 16,384 carbon atoms.

cs.DC↗

arXiv2024

VE: Modeling Multivariate Time Series Correlation with Variate Embedding

Multivariate time series forecasting relies on accurately capturing the correlations among variates. Current channel-independent (CI) models and models with a CI final projection layer are unable to capture these dependencies. In this paper, we present the variate embedding (VE) pipeline, which learns a unique and consistent embedding for each variate and combines it with Mixture of Experts (MoE) and Low-Rank Adaptation (LoRA) techniques to enhance forecasting performance while controlling parameter size. The VE pipeline can be integrated into any model with a CI final projection layer to improve multivariate forecasting. The learned VE effectively groups variates with similar temporal patterns and separates those with low correlations. The effectiveness of the VE pipeline is demonstrated through experiments on four widely-used datasets. The code is available at: https://github.com/swang-song/VE.

cs.LG↗

arXiv2012

The weak localization for the alloy-type Anderson model on a cubic lattice

We consider alloy type random Schrödinger operators on a cubic lattice whose randomness is generated by the sign-indefinite single-site potential. We derive Anderson localization for this class of models in the Lifshitz tails regime, i.e. when the coupling parameter $λ$ is small, for the energies $E \le -C λ^2$.

math-ph↗

arXiv2012

On the efficiency of Hamiltonian-based quantum computation for low-rank matrices

We present an extension of Adiabatic Quantum Computing (AQC) algorithm for the unstructured search to the case when the number of marked items is unknown. The algorithm maintains the optimal Grover speedup and includes a small counting subroutine. Our other results include a lower bound on the amount of time needed to perform a general Hamiltonian-based quantum search, a lower bound on the evolution time needed to perform a search that is valid in the presence of control error and a generic upper bound on the minimum eigenvalue gap for evolutions. In particular, we demonstrate that quantum speedup for the unstructured search using AQC type algorithms may only be achieved under very rigid control precision requirements.

math-ph↗

arXiv2010

Adiabatic quantum computation: Enthusiast and Sceptic's perspectives

Enthusiast's perspective: We analyze the effectiveness of AQC for a small rank problem Hamiltonian $H_F$ with the arbitrary initial Hamiltonian $H_I$. We prove that for the generic $H_I$ the running time cannot be smaller than $O(\sqrt N)$, where $N$ is a dimension of the Hilbert space. We also construct an explicit $H_I$ for which the running time is indeed $O(\sqrt N)$. Our algorithm can be used to solve the unstructured search problem with the unknown number of marked items. Sceptic's perspective: We show that for a robust device, the running time for such $H_F$ cannot be much smaller than $O(N/\ln N)$.

quant-ph↗