Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2409.02294

Accelerating Fortran Codes: A Method for Integrating Coarray Fortran with CUDA Fortran and OpenMP

James McKevitt·Eduard I. Vorobyov·Igor Kulikov
View original source ↗Read source document ↗Export metadata ↓

Abstract

Fortran's prominence in scientific computing requires strategies to ensure both that legacy codes are efficient on high-performance computing systems, and that the language remains attractive for the development of new high-performance codes. Coarray Fortran (CAF), part of the Fortran 2008 standard introduced for parallel programming, facilitates distributed memory parallelism with a syntax familiar to Fortran programmers, simplifying the transition from single-processor to multi-processor coding. This research focuses on innovating and refining a parallel programming methodology that fuses the strengths of Intel Coarray Fortran, Nvidia CUDA Fortran, and OpenMP for distributed memory parallelism, high-speed GPU acceleration and shared memory parallelism respectively. We consider the management of pageable and pinned memory, CPU-GPU affinity in NUMA multiprocessors, and robust compiler interfacing with speed optimisation. We demonstrate our method through its application to a parallelised Poisson solver and compare the methodology, implementation, and scaling performance to that of the Message Passing Interface (MPI), finding CAF offers similar speeds with easier implementation. For new codes, this approach offers a faster route to optimised parallel computing. For legacy codes, it eases the transition to parallel computing, allowing their transformation into scalable, high-performance computing applications without the need for extensive re-design or additional syntax.

Explore related subjects

astro-ph.IMastro-ph.SRcs.DCcs.PL

Keep this discovery

BibTeXRIS

James McKevitt, Eduard I. Vorobyov, Igor Kulikov. 2024-09-03. Accelerating Fortran Codes: A Method for Integrating Coarray Fortran with CUDA Fortran and OpenMP. https://doi.org/10.1016/j.jpdc.2024.104977

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

The EDD Radio Astronomy Backend Framework

Modern digital radio astronomy receivers produce increasingly wide-bandwidth, high bit-rate data streams that necessitate the development of flexible, scalable, and maintainable backend processing and recording systems. Historically, such backend instrumentation has been tightly coupled to telescope observing modes, limiting reuse between observatories and science cases. We present the Effelsberg Direct Digitisation (EDD) backend framework, a software-defined architecture for constructing real-time radio astronomy backends on commodity off-the-shelf computing infrastructure. We describe its design, implementation, supported observing modes, and operational deployments. EDD separates a common core framework from plugin-provided observing capabilities. The core provides orchestration, telescope interfaces, pipeline lifecycle management, monitoring, and deployment tooling, while plugins implement processing pipelines for specific observing modes. The framework is designed to support both single-dish and interferometric instruments through site-specific configuration and plugin selection. EDD currently supports spectroscopy and spectropolarimetry, pulsar timing and searching, baseband recording, very long baseline interferometry, correlation, and beamforming. Operational deployments include the Effelsberg 100-m telescope, the SKA-MPI prototype dish, the Thai National Radio Telescope, and the ARGOS interferometric prototype array. By separating common services, observing-mode plugins, and site-specific configuration, it allows backend capabilities to be deployed across heterogeneous telescope environments and provides a community resource for broadband radio astronomy instrumentation.

astro-ph.IM↗
arXiv2026

Bayesian Superiority in On/Off analysis

We present a detailed comparison of Bayesian criteria with three non-informative priors - flat, Jeffreys, and scale-invariant - for testing a signal against an unknown background and compare them with the classical frequentist Li-Ma approach in the On/Off problem. We perform Monte Carlo simulations for various background levels and evaluate the Li-Ma and Bayesian criteria by their Type I error rates. We then simulate a nonzero signal and compare the criteria in terms of Type II error rates. We find that the Bayesian criterion with the Jeffreys prior yields lower Type I and Type II error rates than the Li-Ma criterion. In addition, we show that the Bayesian criteria are more robust than the Li-Ma criterion when the background distribution is overdispersed relative to the Poisson distribution.

astro-ph.IM↗
arXiv2026

An RFSoC-based Backend and Timing System for the Balloon-borne Very Long Baseline Interferometry Experiment

We present the design and performance characterization of the digital backend and precision-timing system for the Balloon-borne Very Long Baseline Interferometry Experiment (BVEX), a pathfinder for high-frequency stratospheric VLBI at 22 GHz. The backend uses one of the four 14-bit analog-to-digital converter inputs on an AMD-Xilinx RFSoC 4x2. Although the converters support sampling rates up to 5 GSPS, the flight configuration digitizes the 2-4 GHz intermediate frequency at 4.096 GSPS. CASPER firmware provides both a high-resolution spectrometer for pointing and receiver verification, and a VLBI acquisition chain with two-bit requantization that records at a rate of about 8.2 Gbps. The timestamped data packets are sent over 100 Gigabit Ethernet (GbE) to a 16 TB NVMe array in a storage computer that draws approximately 70-80 W. The timing chain uses a Rakon oven-controlled crystal oscillator as a timing reference while a time-interval counter measures its drift relative to a GPS reference with approximately 60 ps resolution. This is the first deployment of an RFSoC-based VLBI backend and precision-timing system on a stratospheric balloon. Ground tests validated the backend, spectrometer, and timing chain. The August 2025 CSA STRATOS flight ended before reaching the target float altitude because of a balloon failure, and as a result no science observations were obtained. For the planned 2027 reflight, we are developing a conduction-cooled data storage computer with 24 TB of NVMe capacity and a direct data path from the 100 GbE interface to the NVMe array.

astro-ph.IM↗