arXiv subjects

Ray Wang

Publications and source records attributed to Ray Wang.

5 recordsLinked to original sources

arXiv2026

LoVoCCS. III. Third Generation Pipeline & The Hercules Supercluster

The Local Volume Complete Cluster Survey (LoVoCCS) is a volume-complete survey of over one-hundred nearby ($0.03 < z < 0.12$), X-ray luminous ($L_{500} > 10^{44} \text{ erg s}^{-1}$) galaxy clusters in the southern sky. Observations for the survey concluded in December 2025, reaching Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) Year 1-2 depth in each field and providing observations with $\lesssim 1"$ seeing for weak lensing science. In this paper, we present the latest pipeline for reducing observations using the third-generation of the LSST Science Pipelines. We use recent observations of the Hercules Supercluster to validate the pipeline's data-products and conduct an extensive multi-plane weak-lensing analysis of a $\sim 16 \text{ deg}^2$ complex covering Abell 2147, 2151, 2152, and several additional structures. We confirm that the dynamical mass of the complex is biased due to the dynamical state of Abell 2147, which is consistent with being $\sim 0.2-0.4 \text{ Gyr}$ out-of periapsis, and estimate that the total mass of the supercluster is $8.9^{+1.7}_{-1.4} \times 10^{14}~M_{\odot}$.

astro-ph.GA↗

arXiv2026

Scalable Training of Mixture-of-Experts Models with Megatron Core

Scaling Mixture-of-Experts (MoE) training introduces systems challenges absent in dense models. Because each token activates only a subset of experts, this sparsity allows total parameters to grow much faster than per-token computation, creating coupled constraints across memory, communication, and computation. Optimizing one dimension often shifts pressure to another, demanding co-design across the full system stack. We address these challenges for MoE training through integrated optimizations spanning memory (fine-grained recomputation, offloading, etc.), communication (optimized dispatchers, overlapping, etc.), and computation (Grouped GEMM, fusions, CUDA Graphs, etc.). The framework also provides Parallel Folding for flexible multi-dimensional parallelism, low-precision training support for FP8 and NVFP4, and efficient long-context training. On NVIDIA GB300 and GB200, it achieves 1,233/1,048 TFLOPS/GPU for DeepSeek-V3-685B and 974/919 TFLOPS/GPU for Qwen3-235B. As a performant, scalable, and production-ready open-source solution, it has been used across academia and industry for training MoE models ranging from billions to trillions of parameters on clusters scaling up to thousands of GPUs. This report explains how these techniques work, their trade-offs, and their interactions at the systems level, providing practical guidance for scaling MoE models with Megatron Core.

cs.DC↗

arXiv2026

CFHT MegaCam Two Deep Fields Imaging Survey (2DFIS) II: Decoding the Lensing Profile of a "Rotating" Cluster with Deep CFHT Imaging

We present a multi-wavelength analysis of the galaxy cluster RXCJ0110.0+1358 ($z=0.058$), a rotating cluster candidate, combining deep CFHT imaging, SDSS photometry, spectroscopic redshifts, and XMM-Newton X-ray observations. We find a notable discrepancy between the optical and X-ray views: while optical data reveal a pronounced bimodal galaxy distribution with significant kinematic substructure signatures, the X-ray emission exhibits a single, smoothly extended component centered on the BCG. Our weak lensing analysis resolves this discrepancy by revealing that the mass is predominantly concentrated in the southeast ($\log M_{200}/M_\odot = 14.04_{-0.40}^{+0.24}$), while the northwestern substructure has a negligible mass ($\sim 10^{13} M_\odot$). This immense mass disparity rules out the dynamical possibility of a rotating system. We demonstrate that the apparent optical bimodality arises from the projection of a filament, which led optical group-finding algorithms to misclassify these galaxies as cluster members. This contamination creates a spurious substructure that mimics a rotation signal and leads to an overestimation of the luminosity-based halo mass, resolving the observed inconsistencies.

astro-ph.GA↗

arXiv2024

DAXA: Traversing the X-ray desert by Democratising Archival X-ray Astronomy

We introduce a new, open-source, Python module for the acquisition and processing of archival data from many X-ray telescopes - Democratising Archival X-ray Astronomy (hereafter referred to as DAXA). Our software is built to increase access to, and use of, large archives of X-ray astronomy data; providing a unified, easy-to-use, Python interface to the disparate archives and processing tools. We provide this interface for the majority of X-ray telescopes launched within the last 30 years. This module enables much greater access to X-ray data for non-specialists, while preserving low-level control of processing for X-ray experts. It is useful for identifying relevant observations of a single object of interest but it excels at creating multi-mission datasets for serendipitous or targeted studies of large samples of X-ray emitting objects. The management and organization of datasets is also made easier; DAXA archives can be version controlled and updated if new data become available. Once relevant observations are identified, the raw data can be downloaded (and optionally processed) through DAXA, or pre-processed event lists, images, and exposure maps can be downloaded if they are available. X-ray observations are perfectly suited to serendipitous discoveries and archival analyses, and with a decade-long `X-ray desert' potentially on the horizon archival data will take on even greater importance; enhanced access to those archives will be vital to the continuation of X-ray astronomy.

astro-ph.IM↗

arXiv2023

CAGRA: Highly Parallel Graph Construction and Approximate Nearest Neighbor Search for GPUs

Approximate Nearest Neighbor Search (ANNS) plays a critical role in various disciplines spanning data mining and artificial intelligence, from information retrieval and computer vision to natural language processing and recommender systems. Data volumes have soared in recent years and the computational cost of an exhaustive exact nearest neighbor search is often prohibitive, necessitating the adoption of approximate techniques. The balanced performance and recall of graph-based approaches have more recently garnered significant attention in ANNS algorithms, however, only a few studies have explored harnessing the power of GPUs and multi-core processors despite the widespread use of massively parallel and general-purpose computing. To bridge this gap, we introduce a novel parallel computing hardware-based proximity graph and search algorithm. By leveraging the high-performance capabilities of modern hardware, our approach achieves remarkable efficiency gains. In particular, our method surpasses existing CPU and GPU-based methods in constructing the proximity graph, demonstrating higher throughput in both large- and small-batch searches while maintaining compatible accuracy. In graph construction time, our method, CAGRA, is 2.2~27x faster than HNSW, which is one of the CPU SOTA implementations. In large-batch query throughput in the 90% to 95% recall range, our method is 33~77x faster than HNSW, and is 3.8~8.8x faster than the SOTA implementations for GPU. For a single query, our method is 3.4~53x faster than HNSW at 95% recall.

cs.DS↗