arXiv subjects

Sameh Abdulah

Publications and source records attributed to Sameh Abdulah.

At least 19 recordsLinked to original sources

arXiv2026

LongVU-TTT: Causal Test-Time Training for Visual Resampling in Long Video Understanding

Long-video MLLMs must model temporal change before a limited visual-token budget removes most frame evidence. We introduce LongVU-TTT, which inserts a convolutional Test-Time Training (TTT) resampler with causal fast-weight updates between the vision encoder and the LLM. Its grouped 2D fast weights adapt to each video and contextualize frame features before compression, while a hybrid uniform-and-change-aware selector retains explicit visual evidence for downstream reasoning. Under controlled conditions, TTT-Conv improves over TTT-MLP by up to +2.12 and bidirectional Mamba2 by up to +3.04 on MLVU, and it is stronger than attention- and fixed-state recurrent resamplers across three benchmarks. Analysis shows that the fast weights behave as a temporal aggregation state rather than a reliable long-horizon episodic memory: their benefit attenuates as evidence becomes more distant, motivating explicit frame retention. LongVU-TTT processes up to 512 frames before reducing them to 128 LLM frames and achieves competitive performance across five video understanding benchmarks.

cs.CV↗

arXiv2026

Cross-Layer Energy Analysis of Multimodal Training on Grace Hopper Superchips

Multimodal deep learning models enable joint learning across heterogeneous data sources, including text, images, and video, but their rapid scaling introduces significant memory and communication bottlenecks. As model sizes and sequence lengths increase, training performance becomes increasingly impacted by data movement rather than computation. Frameworks such as DeepSpeed mitigate these challenges through CPU offloading, activation checkpointing, and communication optimizations. However, these techniques introduce additional system activity, which may affect energy efficiency. Meanwhile, tightly integrated heterogeneous architectures, such as the NVIDIA Grace Hopper (GH200) superchip, provide high-bandwidth CPU-GPU interconnects and unified memory, thereby reducing data transfer overhead. In this work, we present a cross-layer analysis of energy and performance trade-offs in multimodal training on GH200 systems, explicitly characterizing the interactions between application, runtime, and hardware layers. Leveraging high-bandwidth CPU-GPU interconnects, our results show that energy efficiency is primarily governed by data movement and overlap rather than raw compute utilization, and that configurations optimized for runtime are not necessarily optimal for energy. Based on these findings, we distill a set of actionable guidelines for practitioners that demonstrate how to balance offloading strategies, sequence parallelism, and hardware-aware scheduling to achieve energy-efficient training. Our results demonstrate that leveraging high-bandwidth CPU-GPU interconnects enables offloading strategies and sequence parallelism, achieving a strong balance among energy efficiency, runtime performance, and computational throughput, providing practical guidelines for efficient multimodal training on modern heterogeneous systems.

cs.DC↗

arXiv2026

Scaled Block Vecchia Approximation for High-Dimensional Gaussian Process Emulation on GPUs

Emulating computationally intensive scientific simulations is crucial for enabling uncertainty quantification, optimization, and informed decision-making at scale. Gaussian Processes (GPs) offer a flexible and data-efficient foundation for statistical emulation, but their poor scalability limits applicability to large datasets. We introduce the Scaled Block Vecchia (SBV) algorithm for distributed GPU-based systems. SBV integrates the Scaled Vecchia approach for anisotropic input scaling with the Block Vecchia (BV) method to reduce computational and memory complexity while leveraging GPU acceleration techniques for efficient linear algebra operations. To the best of our knowledge, this is the first distributed implementation of any Vecchia-based GP variant. Our implementation employs MPI for inter-node parallelism and the MAGMA library for GPU-accelerated batched matrix computations. We demonstrate the scalability and efficiency of the proposed algorithm through experiments on synthetic and real-world workloads, including a 50M point simulation from a respiratory disease model. SBV achieves near-linear scalability on up to 512 A100 and GH200 GPUs, handles 2.56B points, and reduces energy use relative to exact GP solvers, establishing SBV as a scalable and energy-efficient framework for emulating large-scale scientific models on GPU-based distributed systems.

cs.DC↗

arXiv2026

FedPBS: Proximal-Balanced Scaling Federated Learning Model for Robust Personalized Training for Non-IID Data

Federated learning (FL) enables a set of distributed clients to jointly train machine learning models while preserving their local data privacy, making it attractive for applications in healthcare, finance, mobility, and smart-city systems. However, FL faces several challenges, including statistical heterogeneity and uneven client participation, which can degrade convergence and model quality. In this work, we propose FedPBS, an FL algorithm that couples complementary ideas from FedBS and FedProx to address these challenges. FedPBS dynamically adapts batch sizes to client resources to support balanced and scalable participation, and selectively applies a proximal correction to small-batch clients to stabilize local updates and reduce divergence from the global model. Experiments on benchmarking datasets such as CIFAR-10 and UCI-HAR under highly non-IID settings demonstrate that FedPBS consistently outperforms state-of-the-art methods, including FedBS, FedGA, MOON, and FedProx. The results demonstrate robust performance gains under extreme data heterogeneity, with smooth loss curves indicating stable convergence across diverse federated environments. FedPBS consistently outperforms state-of-the-art federated learning baselines on UCI-HAR and CIFAR-10 under severe non-IID conditions while maintaining stable and reliable convergence.

cs.LG↗

arXiv2026

MPCR: Multi-Precision Computations Package in R

In the early days of computing, severe memory constraints made it necessary to use lower floating-point precision. As hardware capabilities have advanced, modern systems, particularly in computational statistics and scientific computing, have widely adopted 64-bit precision to reduce numerical errors and support complex calculations. However, in some applications, double-precision accuracy exceeds practical requirements, prompting interest in lower-precision alternatives that decrease computational complexity while maintaining adequate accuracy. This trend has accelerated with the advent of hardware optimized for low-precision computations, such as leveraging Tensor Cores technology in recent NVIDIA GPUs. Although lower precision can introduce numerical and accuracy challenges, many applications demonstrate robustness under these conditions. Consequently, new multi-precision algorithms have been developed to balance accuracy and computational cost. To facilitate the adoption of these approaches in statistical computing, this article introduces MPCR, a new R package that supports arithmetic operations at 16-, 32-, and 64-bit precision. Written in C++ and integrated with Rcpp, MPCR delivers highly optimized multi-precision computations on both CPU and GPU, enabling seamless low-precision operations. Several examples demonstrate the benefits of MPCR across both performance and accuracy.

stat.CO↗

arXiv2026

Fisher Scoring for Exact Matérn Covariance Estimation through Stable Smoothness Optimization

Gaussian Random Fields (GRFs) with Matérn covariance functions have emerged as a powerful framework for modeling spatial processes due to their flexibility in capturing different features of the spatial field. However, the smoothness parameter is challenging to estimate using maximum likelihood estimation (MLE), which involves evaluating the likelihood based on the full covariance matrix of the GRF, due to numerical instability. Moreover, MLE remains computationally prohibitive for large spatial datasets. To address this challenge, we propose the Fisher-BackTracking (Fisher-BT) method, which integrates the Fisher scoring algorithm with a backtracking line search strategy and adopts a series approximation for the modified Bessel function. This method enables an efficient MLE estimation for spatial datasets using the ExaGeoStat high-performance computing framework. Our proposed method not only reduces the number of iterations and accelerates convergence compared to derivative-free optimization methods but also improves the numerical stability of the smoothness parameter estimation. Through simulations and real-data analysis using a soil moisture dataset covering the Mississippi River Basin, we show that the proposed Fisher-BT method achieves accuracy comparable to existing approaches while significantly outperforming derivative-free algorithms such as BOBYQA and Nelder-Mead in terms of computational efficiency and numerical stability.

stat.CO↗

arXiv2025

High-Performance Statistical Computing (HPSC): Challenges, Opportunities, and Future Directions

We recognize the emergence of a statistical computing community focused on working with large computing platforms and producing software and applications that exemplify high-performance statistical computing (HPSC). The statistical computing (SC) community develops software that is widely used across disciplines. However, it remains largely absent from the high-performance computing (HPC) landscape, particularly on platforms such as those featured on the Top500 or Green500 lists. Many disciplines already participate in HPC, mostly centered around simulation science, although data-focused efforts under the artificial intelligence (AI) label are gaining popularity. Bridging this gap requires both community adaptation and technical innovation to align statistical methods with modern HPC technologies. We can accelerate progress in fast and scalable statistical applications by building strong connections between the SC and HPC communities. We present a brief history of SC, a vision for how its strengths can contribute to statistical science in the HPC environment (such as HPSC), the challenges that remain, and the opportunities currently available, culminating in a possible roadmap toward a thriving HPSC community.

cs.DC↗

arXiv2025

Scalable Asynchronous Federated Modeling for Spatial Data

Spatial data are central to applications such as environmental monitoring and urban planning, but are often distributed across devices where privacy and communication constraints limit direct sharing. Federated modeling offers a practical solution that preserves data privacy while enabling global modeling across distributed data sources. For instance, environmental sensor networks are privacy- and bandwidth-constrained, motivating federated spatial modeling that shares only privacy-preserving summaries to produce timely, high-resolution pollution maps without centralizing raw data. However, existing federated modeling approaches either ignore spatial dependence or rely on synchronous updates that suffer from stragglers in heterogeneous environments. This work proposes an asynchronous federated modeling framework for spatial data based on low-rank Gaussian process approximations. The method employs block-wise optimization and introduces strategies for gradient correction, adaptive aggregation, and stabilized updates. We establish linear convergence with explicit dependence on staleness, a result of standalone theoretical significance. Moreover, numerical experiments demonstrate that the asynchronous algorithm achieves synchronous performance under balanced resource allocation and significantly outperforms it in heterogeneous settings, showcasing superior robustness and scalability.

stat.ME↗

arXiv2025

RCOMPSs: A Scalable Runtime System for R Code Execution on Manycore Systems

R has become a cornerstone of scientific and statistical computing due to its extensive package ecosystem, expressive syntax, and strong support for reproducible analysis. However, as data sizes and computational demands grow, native R parallelism support remains limited. This paper presents RCOMPSs, a scalable runtime system that enables efficient parallel execution of R applications on multicore and manycore systems. RCOMPSs adopts a dynamic, task-based programming model, allowing users to write code in a sequential style, while the runtime automatically handles asynchronous task execution, dependency tracking, and scheduling across available resources. We present RCOMPSs using three representative data analysis algorithms, i.e., K-nearest neighbors (KNN) classification, K-means clustering, and linear regression and evaluate their performance on two modern HPC systems: KAUST Shaheen-III and Barcelona Supercomputing Center (BSC) MareNostrum 5. Experimental results reveal that RCOMPSs demonstrates both strong and weak scalability on up to 128 cores per node and across 32 nodes. For KNN and K-means, parallel efficiency remains above 70% in most settings, while linear regression maintains acceptable performance under shared and distributed memory configurations despite its deeper task dependencies. Overall, RCOMPSs significantly enhances the parallel capabilities of R with minimal, automated, and runtime-aware user intervention, making it a practical solution for large-scale data analytics in high-performance environments.

cs.DC↗

arXiv2025

GPU-Accelerated Modified Bessel Function of the Second Kind for Gaussian Processes

Modified Bessel functions of the second kind are widely used in physics, engineering, spatial statistics, and machine learning. Since contemporary scientific applications, including machine learning, rely on GPUs for acceleration, providing robust GPU-hosted implementations of special functions, such as the modified Bessel function, is crucial for performance. Existing implementations of the modified Bessel function of the second kind rely on CPUs and have limited coverage of the full range of values needed in some applications. In this work, we present a robust implementation of the modified Bessel function of the second kind on GPUs, eliminating the dependence on the CPU host. We cover a range of values commonly used in real applications, providing high accuracy compared to common libraries like the GNU Scientific Library (GSL) when referenced to Mathematica as the authority. Our GPU-accelerated approach also demonstrates a 2.68X performance improvement using a single A100 GPU compared to the GSL on 40-core Intel Cascade Lake CPUs. Our implementation is integrated into ExaGeoStat, the HPC framework for Gaussian process modeling, where the modified Bessel function of the second kind is required by the Matern covariance function in generating covariance matrices. We accelerate the matrix generation process in ExaGeoStat by up to 12.62X with four A100 GPUs while maintaining almost the same accuracy for modeling and prediction operations using synthetic and real datasets.

cs.DC↗

arXiv2025

Decentralized Inference for Spatial Data Using Low-Rank Models

Advancements in information technology have enabled the creation of massive spatial datasets, driving the need for scalable and efficient computational methodologies. While offering viable solutions, centralized frameworks are limited by vulnerabilities such as single-point failures and communication bottlenecks. This paper presents a decentralized framework tailored for parameter inference in spatial low-rank models to address these challenges. A key obstacle arises from the spatial dependence among observations, which prevents the log-likelihood from being expressed as a summation-a critical requirement for decentralized optimization approaches. To overcome this challenge, we propose a novel objective function leveraging the evidence lower bound, which facilitates the use of decentralized optimization techniques. Our approach employs a block descent method integrated with multi-consensus and dynamic consensus averaging for effective parameter optimization. We prove the convexity of the new objective function in the vicinity of the true parameters, ensuring the convergence of the proposed method. Additionally, we present the first theoretical results establishing the consistency and asymptotic normality of the estimator within the context of spatial low-rank models. Extensive simulations and real-world data experiments corroborate these theoretical findings, showcasing the robustness and scalability of the framework.

stat.ML↗

arXiv2025

A Novel Approach to Translate Structural Aggregation Queries to MapReduce Code

Data management applications are growing and require more attention, especially in the "big data" era. Thus, supporting such applications with novel and efficient algorithms that achieve higher performance is critical. Array database management systems are one way to support these applications by dealing with data represented in n-dimensional data structures. For instance, software like SciDB and RasDaMan can be powerful tools to achieve the required performance on large-scale problems with multidimensional data. Like their relational counterparts, these management systems support specific array query languages as the user interface. As a popular programming model, MapReduce allows large-scale data analysis, facilitates query processing, and is used as a DB engine. Nevertheless, one major obstacle is the low productivity of developing MapReduce applications. Unlike high-level declarative languages such as SQL, MapReduce jobs are written in a low-level descriptive language, often requiring massive programming efforts and complicated debugging processes. This work presents a system that supports translating array queries expressed in the Array Query Language (AQL) in SciDB into MapReduce jobs. We focus on translating some unique structural aggregations, including circular, grid, hierarchical, and sliding aggregations. Unlike traditional aggregations in relational DBs, these structural aggregations are designed explicitly for array manipulation. Thus, our work can be considered an array-view counterpart of existing SQL to MapReduce translators like HiveQL and YSmart. Our translator supports structural aggregations over arrays to meet various array manipulations. The translator can also help user-defined aggregation functions with minimal user effort. We show that our translator can generate optimized MapReduce code, which performs better than the short handwritten code by up to 10.84x.

cs.DB↗

arXiv2025

Block Vecchia Approximation for Scalable and Efficient Gaussian Process Computations

Gaussian Processes (GPs) are vital for modeling and predicting irregularly-spaced, large geospatial datasets. However, their computations often pose significant challenges in large-scale applications. One popular method to approximate GPs is the Vecchia approximation, which approximates the full likelihood via a series of conditional probabilities. The classical Vecchia approximation uses univariate conditional distributions, which leads to redundant evaluations and memory burdens. To address this challenge, our study introduces block Vecchia, which evaluates each multivariate conditional distribution of a block of observations, with blocks formed using the K-means algorithm. The proposed GPU framework for the block Vecchia uses varying batched linear algebra operations to compute multivariate conditional distributions concurrently, notably diminishing the frequent likelihood evaluations. Diving into the factor affecting the accuracy of the block Vecchia, the neighbor selection criterion is investigated, where we found that the random ordering markedly enhances the approximated quality as the block count becomes large. To verify the scalability and efficiency of the algorithm, we conduct a series of numerical studies and simulations, demonstrating their practical utility and effectiveness compared to the exact GP. Moreover, we tackle large-scale real datasets using the block Vecchia method, i.e., high-resolution 3D profile wind speed with a million points.

stat.CO↗

arXiv2024

Accelerating Mixed-Precision Out-of-Core Cholesky Factorization with Static Task Scheduling

This paper explores the performance optimization of out-of-core (OOC) Cholesky factorization on shared-memory systems equipped with multiple GPUs. We employ fine-grained computational tasks to expose concurrency while creating opportunities to overlap data movement asynchronously with computations, especially when dealing with matrices that cannot fit on the GPU memory. We leverage the directed acyclic graph of the task-based Cholesky factorization and map it onto a static scheduler that promotes data reuse while supporting strategies for reducing data movement with the CPU host when the GPU memory is exhausted. The CPU-GPU interconnect may become the main performance bottleneck as the gap between the GPU execution rate and the traditional PCIe bandwidth continues to widen. While the surface-to-volume effect of compute-bound kernels partially mitigates the overhead of data motion, deploying mixed-precision (MxP) computations exacerbates the throughput discrepancy. Using static task scheduling, we evaluate the performance capabilities of the new ultra-fast NVIDIA chip interconnect technology, codenamed NVLink-C2C, that constitutes the backbone of the NVIDIA Grace Hopper Superchip (GH200), against a new four-precision (FP64/FP32/FP16/FP8) left-looking Cholesky factorization. We report the performance results of a benchmarking campaign on various NVIDIA GPU generations and interconnects. We highlight 20% performance superiority against cuSOLVER on a single GH200 with FP64 while hiding the cost of OOC task-based Cholesky factorization, and we scale almost linearly on four GH200 superships. With MxP enabled, our statically scheduled four-precision tile-based Cholesky factorization scores a 3X performance speedup against its FP64-only counterpart, delivering application-worthy FP64 accuracy when modeling a large-scale geospatial statistical application.

cs.DC↗

arXiv2024

Boosting Earth System Model Outputs And Saving PetaBytes in their Storage Using Exascale Climate Emulators

We present the design and scalable implementation of an exascale climate emulator for addressing the escalating computational and storage requirements of high-resolution Earth System Model simulations. We utilize the spherical harmonic transform to stochastically model spatio-temporal variations in climate data. This provides tunable spatio-temporal resolution and significantly improves the fidelity and granularity of climate emulation, achieving an ultra-high spatial resolution of 0.034 (approximately 3.5 km) in space. Our emulator, trained on 318 billion hourly temperature data points from a 35-year and 31 billion daily data points from an 83-year global simulation ensemble, generates statistically consistent climate emulations. We extend linear solver software to mixed-precision arithmetic GPUs, applying different precisions within a single solver to adapt to different correlation strengths. The PaRSEC runtime system supports efficient parallel matrix operations by optimizing the dynamic balance between computation, communication, and memory requirements. Our BLAS3-rich code is optimized for systems equipped with four different families and generations of GPUs, scaling well to achieve 0.976 EFlop/s on 9,025 nodes (36,100 AMD MI250X multichip module (MCM) GPUs) of Frontier (nearly full system), 0.739 EFlop/s on 1,936 nodes (7,744 Grace-Hopper Superchips (GH200)) of Alps, 0.243 EFlop/s on 1,024 nodes (4,096 A100 GPUs) of Leonardo, and 0.375 EFlop/s on 3,072 nodes (18,432 V100 GPUs) of Summit.

stat.CO↗

arXiv2024

Parallel Approximations for High-Dimensional Multivariate Normal Probability Computation in Confidence Region Detection Applications

Addressing the statistical challenge of computing the multivariate normal (MVN) probability in high dimensions holds significant potential for enhancing various applications. One common way to compute high-dimensional MVN probabilities is the Separation-of-Variables (SOV) algorithm. This algorithm is known for its high computational complexity of O(n^3) and space complexity of O(n^2), mainly due to a Cholesky factorization operation for an n X n covariance matrix, where $n$ represents the dimensionality of the MVN problem. This work proposes a high-performance computing framework that allows scaling the SOV algorithm and, subsequently, the confidence region detection algorithm. The framework leverages parallel linear algebra algorithms with a task-based programming model to achieve performance scalability in computing process probabilities, especially on large-scale systems. In addition, we enhance our implementation by incorporating Tile Low-Rank (TLR) approximation techniques to reduce algorithmic complexity without compromising the necessary accuracy. To evaluate the performance and accuracy of our framework, we conduct assessments using simulated data and a wind speed dataset. Our proposed implementation effectively handles high-dimensional multivariate normal (MVN) probability computations on shared and distributed-memory systems using finite precision arithmetics and TLR approximation computation. Performance results show a significant speedup of up to 20X in solving the MVN problem using TLR approximation compared to the reference dense solution without sacrificing the application's accuracy. The qualitative results on synthetic and real datasets demonstrate how we maintain high accuracy in detecting confidence regions even when relying on TLR approximation to perform the underlying linear algebra operations.

cs.DC↗

arXiv2024

GPU-Accelerated Vecchia Approximations of Gaussian Processes for Geospatial Data using Batched Matrix Computations

Gaussian processes (GPs) are commonly used for geospatial analysis, but they suffer from high computational complexity when dealing with massive data. For instance, the log-likelihood function required in estimating the statistical model parameters for geospatial data is a computationally intensive procedure that involves computing the inverse of a covariance matrix with size n X n, where n represents the number of geographical locations. As a result, in the literature, studies have shifted towards approximation methods to handle larger values of n effectively while maintaining high accuracy. These methods encompass a range of techniques, including low-rank and sparse approximations. Vecchia approximation is one of the most promising methods to speed up evaluating the log-likelihood function. This study presents a parallel implementation of the Vecchia approximation, utilizing batched matrix computations on contemporary GPUs. The proposed implementation relies on batched linear algebra routines to efficiently execute individual conditional distributions in the Vecchia algorithm. We rely on the KBLAS linear algebra library to perform batched linear algebra operations, reducing the time to solution compared to the state-of-the-art parallel implementation of the likelihood estimation operation in the ExaGeoStat software by up to 700X, 833X, 1380X on 32GB GV100, 80GB A100, and 80GB H100 GPUs, respectively. We also successfully manage larger problem sizes on a single NVIDIA GPU, accommodating up to 1M locations with 80GB A100 and H100 GPUs while maintaining the necessary application accuracy. We further assess the accuracy performance of the implemented algorithm, identifying the optimal settings for the Vecchia approximation algorithm to preserve accuracy on two real geospatial datasets: soil moisture data in the Mississippi Basin area and wind speed data in the Middle East.

stat.CO↗

arXiv2024

On the Impact of Spatial Covariance Matrix Ordering on Tile Low-Rank Estimation of Matérn Parameters

Spatial statistical modeling and prediction involve generating and manipulating an n*n symmetric positive definite covariance matrix, where n denotes the number of spatial locations. However, when n is large, processing this covariance matrix using traditional methods becomes prohibitive. Thus, coupling parallel processing with approximation can be an elegant solution to this challenge by relying on parallel solvers that deal with the matrix as a set of small tiles instead of the full structure. Each processing unit can process a single tile, allowing better performance. The approximation can also be performed at the tile level for better compression and faster execution. The Tile Low-Rank (TLR) approximation, a tile-based approximation algorithm, has recently been used in spatial statistics applications. However, the quality of TLR algorithms mainly relies on ordering the matrix elements. This order can impact the compression quality and, therefore, the efficiency of the underlying linear solvers, which highly depends on the individual ranks of each tile. Thus, herein, we aim to investigate the accuracy and performance of some existing ordering algorithms that are used to order the geospatial locations before generating the spatial covariance matrix. Furthermore, we highlight the pros and cons of each ordering algorithm in the context of spatial statistics applications and give hints to practitioners on how to choose the ordering algorithm carefully. We assess the quality of the compression and the accuracy of the statistical parameter estimates of the Matérn covariance function using TLR approximation under various ordering algorithms and settings of correlations.

stat.CO↗