Irish Centre For High-End Computing Research Articles

Machine Learning with <i>Kay</i>

Abstract. Computational power is very important when training Deep Learning (DL) models with large amounts of data (Wooldridge, 2021). Hence, High-Performance Computing (HPC) can be leveraged to reduce computational cost, and the Irish Centre for High-End Computing (ICHEC) provides significant infrastructure and services for research and development to both academia and industry. A portion of ICHEC's HPC system has been allocated for institutional access, and this paper presents a case study of how to use Kay (Ireland's national supercomputer) in the remote sensing domain. Specifically, this study uses clusters of Kay Graphics Processing Units (GPUs) for training DL models to extract buildings from satellite imagery using a large number of input data samples.

The air-water interface stabilizes α-helical conformations of the insulin B-chain

Adsorption of proteins onto liquid interfaces, such as the air-water interface, often leads to changes in the protein conformation. This can lead to changes in protein assembly behavior, with aggregation and fibrillation often enhanced. To understand the relationship between protein conformation and aggregation, knowledge of protein structure at interfaces, on the single molecular level, is necessary. Using molecular dynamics simulations, the effect of the air-water interface on conformation of the insulin B-chain is investigated. At the air-water interface, the protein adopts an α-helical conformation, whereas in bulk solution, it adopts disordered structures. The α-helical conformation is templated by the partitioning of hydrophobic side chains into the air, leading to the formation of an amphipathic helix. This structure presents a hydrophobic face which may lead to further aggregation, which helps explain the enhancement of insulin fibrillation at interfaces. This knowledge of the molecular conformation gives new insight into the contribution of protein structural change on the interfacial aggregation of proteins.

Towards a definitive historical high-resolution climate dataset for Ireland – promoting climate research in Ireland

Abstract. There is strong and constant demand from various sectors (research, industry and government) for long-term, high-resolution (both temporal and spatial), gridded climate datasets. To address this demand, the Irish Centre for High-End Computing (ICHEC) has recently performed two high-resolution simulations of the Irish climate, utilising the Regional Climate Models (RCMs) COSMO-CLM5 and WRF v3.7.1. The datasets produced contain hourly outputs for an array of sub-surface, surface and atmospheric fields for the entire 36-year period 1981–2016. In this work, we list the climate variables that have been archived at ICHEC. We present preliminary uncertainty estimates (error, standard deviation, mean absolute error) based on Met Éireann station observations, for several of the more commonly used variables: 2 m temperature, 10 m wind speeds and mean sea level pressure at the hourly time scale; and precipitation at hourly and daily time scales. Additionally, analyses of 10 cm soil temperatures, CAPE 3 km, Showalter index and surface lifted index are presented.

Estimation of Future Water Requirement for Plantain (Musa spp.) in Ondo State, Nigeria using Statistically-Downscaled Climate Variables

This study aims to determine the effects of potential climate change on crop water requirements (CWR) of plantain (Musa spp.) in three agro-ecological zones over Ondo State. Climate change scenario-Representative Concentration Pathways (RCP 4.5) derived from statistical downscaling of six global circulation model (GCM) outputs were applied as inputs to CROPWAT model. Statistically-downscaled climate parameters outputs were used as inputs to the CROPWAT to estimate irrigation water requirements (IWR) and CWR of plantain. Canadian Centre of Climate Model and Analysis (CCCMA) model predicted greatest annual maximum temperature (Tmax) increases of 1.3°C to 1.9°C and 0.8°C to 1.5°C at Ondo North and South Agro-Ecological Zones (ONAEZ and OSAEZ) for periods 2050s and 2080s respectively. The outputs of the simulation for Max Planck’s Institute of Meteorology (MPI) model at OSAEZ indicate highest minimum temperature changes of 2.5°C to 3.2°C for periods 2050s and 2080s. Changes in estimated monthly precipitation indicate reductions for most of the models, with the highest decreases in month of January with -95.8% and -96.8.6% for the future periods of 2050s and 2080s using CCCMA. Lowest decreases of -20.0% and -53.8% were predicted by Irish Centre for High End Computing (ICHEC) in ONAEZ. During the wet season (April to September), all the climate models indicated increase in monthly precipitation except CCCMA and MPI which estimated modest decrease of -18.4% and -19.4% in the months of June and July for period 2050s in OCAEZ. The outputs of CROPWAT model to estimate CWR and IWR showed that CCCMA consistently predicts seasonal highest increases of 26.7% (1085.4 mm) and 28.0% (1102.6 mm) for the periods of 2050s and 2080s in comparison to crop water requirement (CWR) of 879.9 mm in ONAEZ. This trend is similar to the predicted changes in CWR obtained in AEZ and OSAEZ respectively. Again, CCCMA model projected highest irrigation water requirements of 451.6 mm and 469.3 mm; 493.8 mm and 500.8 mm for Ondo Centre Agro-Ecological Zone (OCAEZ) and OSAEZ. These results correspond to IWR increase of 19.6% to 36.1% in ONAEZ, 20.3% to 39.9% in OCAEZ and 25.3% to 42.1% in OSAEZ for periods of 2050s and 2080s relative to the current period of 1975-2005. In conclusion, future CWR and IWR of plantain are likely to increase based on projected climate for the periods 2050s and 2080s in Ondo State.

An empirical study of data decomposition for software parallelization

Multi-core programming is becoming increasingly important.Data decomposition is a key challenge during parallelization for multi-core CPUs.We conduct a multi-method study to better understand data decomposition.We derive a set of 10 key requirements for tools to support parallelization.The state-of-the-art tooling support does not support these requirements. Context: Multi-core architectures are becoming increasingly ubiquitous and software professionals are seeking to leverage the capabilities of distributed-memory architectures. The process of parallelizing software applications can be very tedious and error-prone, in particular the task of data decomposition. Empirical studies investigating the complexity of data decomposition and communication are lacking.Objective: Our objective is threefold: (i) to gain an empirical-based understanding of the task of data decomposition as part of the parallelization of software applications; (ii) to identify key requirements for tools to assist developers in this task, and (iii) assess the current state-of-the-art.Methods: Our empirical investigation employed a multi-method approach, using an interview study, participant-observer case study, focus group study, and a sample survey. The empirical investigation involved collaborations with three industry partners: IBMs High Performance Computing Center, the Irish Centre for High-End Computing (ICHEC), and JBA Consulting.Results: This article presents data decomposition as one of the most prevalent tasks of parallelizing applications for multi-core architectures. Based on our studies, we identify ten key requirements for tool support to help HPC developers in this area. Our evaluation of the state-of-the-art shows that none of the extant tool support implements all 10 requirements.Conclusion: While there is a considerable body of research in the area of HPC, a few empirical studies exist which explicitly focus on the challenges faced by practitioners in this area; this research aims to address this gap. The empirical studies in this article provide insights that may help researchers and tool vendors to better understand the needs of parallel programmers.

Finite element analysis of bioresorbable polymers using a thermo-mechanical co-simulation model

Event Abstract Back to Event Finite element analysis of bioresorbable polymers using a thermo-mechanical co-simulation model Rosa Shine1, Caoimhe A. Sweeney1, Nicola Kelly1 and Peter E. Mchugh1 1 College of Engineering and Informatics, National University of Ireland Galway, Center for Biomechanics Research, Dept of Biomedical Engineering, Ireland Introduction: The recent surge in the use of bioresorbable polymers in vascular stent applications is due to their attractive biodegradable natures. Such materials, i.e. poly-L-lactide (PLLA), yield stents which gradually dissolve in-vivo, once the initial scaffolding period has passed. PLLA breaks down through hydrolytic degradation, when water enters the material and causes scission of the polymer chains into monomer products, which are naturally metabolized, leading to a loss of molecular weight. Our objective is to develop a novel computational model that captures the physical aspects of the degradation and also examines its effects on the mechanical performance of a bioresrobable PLLA scaffold. Methods: The physically-based model developed by Wang et al. (2008)[1] describes the degradation of polyesters through changes in the ester bond and monomer concentrations. In a novel approach, the governing equations of this physical model are implemented into the finite-element package Abaqus/Standard (DS SIMULIA,USA), using a thermal analogy to model the degradation of PLLA. A correlation between key events of degradation, hydrolysis of the ester bonds and diffusion of the monomers , with the transfer of heat and temperature in a physical system is established. Correspondence between degradation and thermal parameters is shown in Table 1[5]. The mechanical behavior of PLLA is described using the strain energy function proposed by Knowles(1977)[2] for a hyperelastic incompressible material; W = μ/2b { [(1+b/n(I1 - 3)]n-1} + 1/D1(J-1)2, where μ is the shear modulus, D1 governs compressibility and b and n are softening parameters. The elasticity tensor for this model has previously been derived and implemented into finite-element code by Suchocki et al. (2011)[3]. User material subroutines (UMATHT,UMAT) were applied to a cylindrical part (3mm OD; 2.8mm ID; Length=12mm) to assign the degradation parameters (Table. 1) and mechanical properties (E=3.5GPa;ν= 0.475;b=n=1)[4]. Displacement boundary conditions were applied and the monomer concentration was set to zero at a free surface, representing immediate diffusion of the monomers (Fig. 1). Coupled temperature-displacement elements were used in the mesh and a pressure load of 2atm was applied to the inner surface. Results: Fig. 2 shows the results for the PLLA cylinder, for a 6 month period. The maximum principal stress reaches its highest value at the inner surface of the cylinder (Fig. 2(a)). The ester bond concentration decreases from its initial value, to 970mol/m3 on the outer surface, whereas the monomer concentration increases on this surface. Discussion: This analysis shows it is possible to simulate the degradation and mechanical loading of a bioresorbable polymeric cylinder using finite-element techniques. Stress is generated in the cylinder at the location of the applied loading pressure. The decrease in the ester bond concentration can be attributed to hydrolysis of the polymer chains, controlled by the equations of the physically-based model, which also govern the diffusion of the monomer products. Since this coupled model provides a novel insight into the degradation of bioresorbable polymers, its further development (for example, enhancements to allow investigation of the interactions between in-vivo mechanical loading and degradation rates) and experimental validation would allow for more accurate examination of bioresorbable polymeric stents. This work was funded in part by an NUI Galway, College of Engineering and Informatics, Postgraduate Research Scholarship and a Travelling Studentship in Mechanical & Biomedical Engineering, from the National University of Ireland; The authors also wish to acknowledge the SFI/HEA Irish Centre for High-End Computing (ICHEC) for the provision of computational facilities and support

An AIPS-based, distributed processing method for large radio interferometric datasets

The data output rates of modern radio interferometric telescopes make the traditional data reduction process impractical in many cases. We report on the implementation of a lightweight infrastructure, named AIPSLite, that enables the deployment of AIPS interferometric processing routines on distributed systems in an autonomous and fault tolerant manner. We discuss how this approach was used to search for sources of 6.7 GHz methanol maser emission in the Cep A region with the European VLBI Network (EVN). The field was searched out to a radius of 1.25 min^(−1) at milli-arcsecond spatial resolution and 1024 frequency channels with 0.088 km s^(−1) velocity resolution. The imaged data was on the order of 30 TB. Processing was performed on 128 processors of the Irish Centre for High End Computing (ICHEC) linux cluster with a run time of 42 h, and a total of 212 CPU days.