arXiv subjects

Masha Sosonkina

Publications and source records attributed to Masha Sosonkina.

13 recordsLinked to original sources

arXiv2026

Validating LLM-Modernized Scientific Software Through Differential Fault Injection

Large language model (LLM) agents are increasingly used to modernize the legacy Fortran underlying production scientific software, but validation of these transformations emphasizes nominal executions and may not test whether a modernization preserves the original code's response to faults, perturbations, and reduced precision. We present a differential fault-injection validation method: a harness instruments the shared self-consistent-field driver of GAMESS at twelve sites and applies identical, deterministic faults to the original and LLM-modernized implementations, isolating the converted integral kernels. Across more than 2,200 runs, transient-fault absorption costs match a contraction-based model (predicted slopes 0.74 and 1.49 iterations per bit; measured 0.82 and 1.50), persistent perturbations halve final-energy error per additional bit, and the campaigns expose phase-dependent parallel deadlocks and false convergence under reduced precision. The original and modernized kernels agree in all 200 paired injections, and a measurement-guided synchronization change composes with the modernization, matching in all 40 pairs.

cs.DC↗

arXiv2026

An Agentic Workflow for Legacy HPC Modernization: Converting the Two-Electron-Integral Core of GAMESS

Modernizing legacy Fortran is a problem of volume: the transformations are individually routine, but the codebases can be enormous, and across much of computational science the work simply goes undone. We propose an agentic workflow that takes this work on at production scale, and we set out to measure how far such delegation can reach. In this work, three prompt-specialized agent roles operate under a version-controlled specification that the agents themselves authored and revised, while humans hold a small number of gates. The arrangement is kept safe by an exact verification oracle inherited from the domain, and the boundary of safe delegation lies exactly where that oracle stops seeing. We apply the proposed workflow in a case study, converting the two-electron-integral routines of GAMESS (General Atomic and Molecular Electronic Structure System), a mature quantum-chemistry package with a 48-year development history, from fixed-form Fortran 77 to free-form Fortran 2008. The scope of this work was twelve source files, 56,448 lines, and 225 subroutines for computing electron repulsion integrals. The agents ran as three Claude Code roles in isolated worktrees, and the work spanned four Claude model generations. Because the GAMESS group ships a standard test suite whose printed energies its user community treats as canonical, we could adopt bit-for-bit reproduction of those energies as the merge criterion, where a deviation in the twelfth decimal place counts as a failure rather than drift. All twelve source files pass a 51-test validation battery comprising the 49 standard GAMESS tests and two additional calculations, and across 612 test runs the number of chemistry-relevant differences is zero, and every file also passes the Jenkins tests that are used for continuous integration.

cs.AI↗

arXiv2026

Fault Tolerance of Accelerated Asynchronous Fixed-Point Iterations on Flexible Computing Infrastructure

Asynchronous iterative methods tolerate straggling processors by allowing workers to proceed with stale data, but at a cost: the iterates become inconsistent, potentially degrading convergence. We investigate whether convergence accelerators such as Anderson acceleration compensate for this degradation. We experimentally study three fixed-point iterations: the Jacobi method for sparse linear systems, value iteration for the Bellman equation, and the Hartree--Fock self-consistent field (SCF) iteration. The experiments are conducted using a high-performance execution framework Ray, which abstracts the complexity of distributed systems and enables code parallelization and fault injection with minimal changes. We establish two main results. First, straggler tolerance is universal: asynchronous execution provides wall-clock speedups of $2.9\times$ (Jacobi), $7.7\times$ (VI), and $16.9\times$ (SCF) over synchronous execution with a 100\,ms-delayed worker, independent of whether acceleration is used. Second, Anderson acceleration's effectiveness under asynchrony depends on where staleness enters the computation. We identify two staleness mechanisms: iterate-level corruption, where stale worker returns directly overwrite portions of the accelerated iterate (as in block Jacobi), and evaluation-level perturbation, where staleness acts as a bounded perturbation to the fixed-point map evaluation (as in VI and SCF). Anderson acceleration fails categorically under the first mechanism but retains its benefits under the second, consistent with the perturbation analysis of Toth et al.\ (2017). This distinction, rather than the contraction norm or smoothness of the map, is the primary determinant of whether acceleration survives asynchronous execution.

cs.DC↗

arXiv2025

Heterogeneous Graph Backdoor Attack

Heterogeneous Graph Neural Networks (HGNNs) excel in modeling complex, multi-typed relationships across diverse domains, yet their vulnerability to backdoor attacks remains unexplored. To address this gap, we conduct the first investigation into the susceptibility of HGNNs to existing graph backdoor attacks, revealing three critical issues: (1) high attack budget required for effective backdoor injection, (2) inefficient and unreliable backdoor activation, and (3) inaccurate attack effectiveness evaluation. To tackle these issues, we propose the Heterogeneous Graph Backdoor Attack (HGBA), the first backdoor attack specifically designed for HGNNs, introducing a novel relation-based trigger mechanism that establishes specific connections between a strategically selected trigger node and poisoned nodes via the backdoor metapath. HGBA achieves efficient and stealthy backdoor injection with minimal structural modifications and supports easy backdoor activation through two flexible strategies: Self-Node Attack and Indiscriminate Attack. Additionally, we improve the ASR measurement protocol, enabling a more accurate assessment of attack effectiveness. Extensive experiments demonstrate that HGBA far surpasses multiple state-of-the-art graph backdoor attacks in black-box settings, efficiently attacking HGNNs with low attack budgets. Ablation studies show that the strength of HBGA benefits from our trigger node selection method and backdoor metapath selection strategy. In addition, HGBA shows superior robustness against node feature perturbations and multiple types of existing graph backdoor defense mechanisms. Finally, extension experiments demonstrate that the relation-based trigger mechanism can effectively extend to tasks in homogeneous graph scenarios, thereby posing severe threats to broader security-critical domains.

cs.CR↗

arXiv2020

Parallel solution of saddle point systems with nested iterative solvers based on the Golub-Kahan Bidiagonalization

We present a scalability study of Golub-Kahan bidiagonalization for the parallel iterative solution of symmetric indefinite linear systems with a 2x2 block structure. The algorithms have been implemented within the parallel numerical library PETSc. Since a nested inner-outer iteration strategy may be necessary, we investigate different choices for the inner solvers, including parallel sparse direct and multigrid accelerated iterative methods. We show the strong and weak scalability of the Golub-Kahan bidiagonalization based iterative method when applied to a two-dimensional Poiseuille flow and to two- and three-dimensional Stokes test problems.

math.NA↗

arXiv2015

Experimentation Procedure for Offloaded Mini-Apps Executed on Cluster Architectures with Xeon Phi Accelerators

A heterogeneous cluster architecture is complex. It contains hundreds, or thousands of devices connected by a tiered communication system in order to solve a problem. As a heterogeneous system, these devices will have varying performance capabilities. To better understand the interactions which occur between the various devices during execution, an experimentation procedure has been devised to capture, store, and analyze important and meaningful data. The procedure consists of various tools, techniques, and methods for capturing relevant timing, power, and performance data for a typical execution. This procedure currently applies to architectures with Intel Xeon processors and Intel Xeon Phi accelerators. It has been applied to the Co-Design Molecular Dynamics mini-app, courtesy of the ExMatEx team. This work aims to provide end-users with a strategy for investigating codes executed on heterogeneous cluster architectures with Xeon Phi accelerators.

cs.DC↗

arXiv2015

Using the VBARMS method in parallel computing

The paper describes an improved parallel MPI-based implementation of VBARMS, a variable block variant of the pARMS preconditioner proposed by Li,~Saad and Sosonkina [NLAA, 2003] for solving general nonsymmetric linear systems. The parallel VBARMS solver can detect automatically exact or approximate dense structures in the linear system, and exploits this information to achieve improved reliability and increased throughput during the factorization. A novel graph compression algorithm is discussed that finds these approximate dense blocks structures and requires only one simple to use parameter. A complete study of the numerical and parallel performance of parallel VBARMS is presented for the analysis of large turbulent Navier-Stokes equations on a suite of three-dimensional test cases.

math.NA↗

arXiv2015

Ab Initio No Core Shell Model - Recent Results and Further Prospects

There has been significant recent progress in solving the long-standing problems of how nuclear shell structure and collective motion emerge from underlying microscopic inter-nucleon interactions. We review a selection of recent significant results within the ab initio No Core Shell Model (NCSM) closely tied to three major factors enabling this progress: (1) improved nuclear interactions that accurately describe the experimental two-nucleon and three-nucleon interaction data; (2) advances in algorithms to simulate the quantum many-body problem with strong interactions; and (3) continued rapid development of high-performance computers now capable of performing $20 \times 10^{15}$ floating point operations per second. We also comment on prospects for further developments.

nucl-th↗

arXiv2015

Towards Modeling Energy Consumption of Xeon Phi

In the push for exascale computing, energy efficiency is of utmost concern. System architectures often adopt accelerators to hasten application execution at the cost of power. The Intel Xeon Phi co-processor is unique accelerator that offers application designers high degrees of parallelism, energy-efficient cores, and various execution modes. To explore the vast number of available configurations, a model must be developed to predict execution time, power, and energy for the CPU and Xeon Phi. An experimentation method has been developed which measures power for the CPU and Xeon Phi separately, as well as total system power. Execution time and performance are also captured for two experiments conducted in this work. The experiments, frequency scaling and strong scaling, will help validate the adopted model and assist in the development of a model which defines the host and Xeon Phi. The proxy applications investigated, representative of large-scale real-world applications, are Co-Design Molecular Dynamics (CoMD) and Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics (LULESH). The frequency experiment discussed in this work is used to determine the time on-chip and off-chip to measure the compute- or latencyboundedness of the application. Energy savings were not obtained in symmetric mode for either application.

cs.DC↗

arXiv2014

Accelerating Ab Initio Nuclear Physics Calculations with GPUs

This paper describes some applications of GPU acceleration in ab initio nuclear structure calculations. Specifically, we discuss GPU acceleration of the software package MFDn, a parallel nuclear structure eigensolver. We modify the matrix construction stage to run partly on the GPU. On the Titan supercomputer at the Oak Ridge Leadership Computing Facility, this produces a speedup of approximately 2.2x - 2.7x for the matrix construction stage and 1.2x - 1.4x for the entire run.

physics.comp-ph↗

arXiv2013

Graph partitioning using matrix values for preconditioning symmetric positive definite systems

Prior to the parallel solution of a large linear system, it is required to perform a partitioning of its equations/unknowns. Standard partitioning algorithms are designed using the considerations of the efficiency of the parallel matrix-vector multiplication, and typically disregard the information on the coefficients of the matrix. This information, however, may have a significant impact on the quality of the preconditioning procedure used within the chosen iterative scheme. In the present paper, we suggest a spectral partitioning algorithm, which takes into account the information on the matrix coefficients and constructs partitions with respect to the objective of enhancing the quality of the nonoverlapping additive Schwarz (block Jacobi) preconditioning for symmetric positive definite linear systems. For a set of test problems with large variations in magnitudes of matrix coefficients, our numerical experiments demonstrate a noticeable improvement in the convergence of the resulting solution scheme when using the new partitioning approach.

math.NA↗

arXiv2013

Computational Nuclear Quantum Many-Body Problem: The UNEDF Project

The UNEDF project was a large-scale collaborative effort that applied high-performance computing to the nuclear quantum many-body problem. UNEDF demonstrated that close associations among nuclear physicists, mathematicians, and computer scientists can lead to novel physics outcomes built on algorithmic innovations and computational developments. This review showcases a wide range of UNEDF science results to illustrate this interplay.

nucl-th↗

arXiv2009

{\it Ab initio} nuclear structure - the large sparse matrix eigenvalue problem

The structure and reactions of light nuclei represent fundamental and formidable challenges for microscopic theory based on realistic strong interaction potentials. Several {\it ab initio} methods have now emerged that provide nearly exact solutions for some nuclear properties. The {\it ab initio} no core shell model (NCSM) and the no core full configuration (NCFC) method, frame this quantum many-particle problem as a large sparse matrix eigenvalue problem where one evaluates the Hamiltonian matrix in a basis space consisting of many-fermion Slater determinants and then solves for a set of the lowest eigenvalues and their associated eigenvectors. The resulting eigenvectors are employed to evaluate a set of experimental quantities to test the underlying potential. For fundamental problems of interest, the matrix dimension often exceeds $10^{10}$ and the number of nonzero matrix elements may saturate available storage on present-day leadership class facilities. We survey recent results and advances in solving this large sparse matrix eigenvalue problem. W also outline the challenges that lie ahead for achieving further breakthroughs in fundamental nuclear theory using these {\it ab initio} approaches.

nucl-th↗