Lawrence Berkeley Research Articles

Online correction of laser focal position using FPGA-based ML models

Ultrafast lasers play an increasingly critical role in the generation, manipulation, and acceleration of electron beams for High Energy Physics applications. Laser plasma accelerators enable order of magnitude improvements in accelerating gradient and promise compact tunable GeV electron beam sources, while novel photocathode systems permit fundamental advances in electron beam manipulation for accelerator and radiation applications Advances in fast feedback systems are required to stabilize laser performance at kHz repetition rate operation against environmental fluctuations. A field programmable gate array (FPGA) based digital control system, coupled with responsive optics, can provide rapid and precise stabilization of ultrafast lasers. A collaboration between RadiaSoft and the Lawrence Berkeley National Laboratory BELLA Center to develop, test, and deploy these systems across a range of beamlines operating at >1 Hz repetition rate, including 1 kHz systems, was created.

18F‐fluorodeoxyglucose‐positron emission tomography Findings in Patients with Genetic Frontotemporal Dementia

AbstractBackgroundMutations in microtubule‐associated protein tau (MAPT), granulin (GRN), and hexanucleotide expansion repeat in the open reading frame of chromosome 9 (C9orf72) are found in 60% of familial frontotemporal dementia (FTD) cases. The current study aims to delineate brain regions with reduced glucose metabolism in patients with genetic FTD in comparison with cognitively normal controls.Methods(18)F‐ fluorodeoxyglucose (FDG)‐positron emission tomography (PET) images were retrospectively obtained from 17 genetic FTD patients. All patients were symptomatic at the time of imaging, and the autopsy was available for 13. Pittsburgh compound B (PiB) amyloid PET scans were done for 13 patients (Table 1). FDG‐PET was collected at Lawrence Berkeley National Laboratory on a PET or PET‐CT scanner as six 5‐min frames acquired 30 min post tracer injection. Standard Uptake Value Ratio (SUVR) images were created for each patient using pons as a reference region. W‐score (age‐adjusted Z‐score) maps were created for each patient’s FDG SUVR using a group of neurologically unimpaired controls (N=74, Mean age 65+16, 41 female) as reference.ResultOn review of single‐subject w‐maps, all patients with MAPT mutation showed hypometabolism in the anteromedial temporal lobes with 66% showing anterolateral temporal hypometabolism and 33% in the frontal and dorsolateral parietal lobes. Hypometabolism in temporal lobes was seen in all GRN patients and 75% had frontal and parietal hypometabolism in an asymmetric pattern. 80% of theC9orf72 patients had hypometabolism in anteromedial temporal and orbitofrontal regions. 60% showed hypometabolism in dorsolateral prefrontal, medial frontal, and anterolateral temporal lobes. Posterior temporal, posterior cingulate, parietal lobe, cerebellum, insula, thalamus, and caudate were hypometabolic in less than 40% of the C9orf72 patients (Figure 1).On group‐level analysis, compared to controls, patients with C9orf72 showed hypometabolism in the frontotemporal regions, cerebellum and thalamus. Hypometabolism in the frontotemporal and parietal regions was seen with GRN mutation. MAPT mutation carriers showed frontal and anteromedial temporal hypometabolism (Figure 2).ConclusionHypometabolism in the posterior regions and cerebellum is more suggestive of C9orf72 repeat expansion. An asymmetric pattern of hypometabolism was seen in all GRN mutation carriers. MAPT mutation resulted in a greater degree of hypometabolism in the anterior medial temporal lobes.

Impact of off‐target signal on longitudinal FTP quantification

AbstractBackgroundDespite off‐target signal, Flortaucipir (FTP) PET is useful for measuring tau in vivo. We previously showed in a cross‐sectional cohort of cognitively unimpaired (CU) amyloid negative (Aβ‐) individuals that the global cortical SUVR is highly correlated to putamen and thalamus SUVR, suggesting that there is substantial off‐target signal in the cortex. Here we explore the impact on off‐target signal on longitudinal FTP quantification.MethodsA group of CUAβ‐ subjects scanned through ADNI and Lawrence Berkeley National Laboratory (LBNL) were used to define corrections for off‐target signal (derivation cohort) which were then applied to a separate longitudinal cohort of CUAβ‐, CUAβ+, and cognitively impaired (CI) Aβ+ subjects (Table 1). SUVRs were calculated using two possible reference regions (inferior cerebellar gray and supratentorial white matter) with or without partial volume correction (PVC). The relationship of off‐target signal between thalamus and cortex was estimated as the regression between thalamus and Braak IV‐VI in CUAβ‐ (figure 1) and then applied in the longitiduinal cohort to temporal meta and Braak ROIs to regress out the influence of the thalamic signal (figure 2). Hedges’ effect size was calculated between diagnosis groups with and without regressing out thalamic signal in the LBNL longitudinal cohort to examine effects in a smaller cohort.ResultsThere was a significant correlation in cross‐sectional CUAβ‐ subjects between thalamus and Braak IV‐VI regardless of reference region and PVC (Figure 1). This correlation increased when comparing annual change between thalamus and Braak IV‐VI in longitudinal CUAβ‐ subjects (Figure 3). There was no significant difference across diagnosis groups in the longitudinal cohort between either baseline thalamus or annual change in thalamus. Correlation between annual change in thalamus and entorhinal cortex or meta temporal ROI were often significant (12/12 for inferior cerebellar gray reference region, 8/12 for white matter reference region, Figure 4). Effect size between diagnosis groups increased as a result of regressing out thalamic signal for all temporal meta ROI comparisons and for CUAβ+>CUAβ‐ and CIAβ+>CUAβ‐ in the entorhinal cortex (Figure 5).ConclusionWhen taking into account off‐target signal, it is possible to improve effect size of longitudinal FTP by ∼12%.

Lattice correction and commissioning simulation of the Advanced Light Source upgrade storage ring

The ALS-U is the upgrade of the existing Lawrence Berkeley National Laboratory Advanced Light Source to a diffraction-limited soft x-ray light source. Here we present the lattice correction studies and commissioning simulations demonstrating that the proposed machine design can be expected to deliver the intended performance when realistic errors and perturbations are fully accounted for. Critical to this demonstration are the high-fidelity, realistic simulations of the beam-based alignment process (both in turn-by-turn mode during early commissioning and with stored beam) that are now made possible by the Toolkit for Simulated Commissioning. In addition to presenting a statistical performance analysis based on a large number of lattice error realizations, we also study the range of further improvements that can be obtained by fine-tuning the correction chain to individual error seeds, mimicking the approach one would follow once the machine is built.

Creation, evolution, and future challenges of ion beam therapy from a medical physicist's viewpoint (part 1). Introduction and Chapter 1. accelerator and beam delivery system.

Radiation therapy for cancer using the Bragg peak of an ion beam has been making steady progress after being proposed by Robert Wilson in 1946. At the end of 2020, 12 dedicated treatment devices existed in operation worldwide, and approximately 40,000 patients have been treated with ion beams (mostly carbon ions). To date, ion beam therapy is superior to other treatments for rare cancers in the head and neck as well as bone and soft tissues; however, most recently, evidence submitted in Japan for the 2022 revision of public health insurance shows that ion beam therapy outperforms photon therapy for intractable common cancers such as pancreatic cancer and liver cancer. This may greatly expand its indications. Lawrence Berkeley Laboratory in the United States started research of ion beam therapy, National Institute of Radiological Sciences in Japan built the first dedicated device Heavy Ion Accelerator in Chiba and started systematic clinical research, and GSI in Germany developed the scanning irradiation method and rotating gantry for the first time. This paper presents the history and future challenges of ion beam therapy in three fields: accelerator and beam delivery system, physical/biological model and treatment planning system, and clinical research. This study is divided into three parts describing the achievements and roles of the three laboratories. In Part 1, accelerator and beam delivery system are described.

Strong-field QED experiments using the BELLA PW laser dual beamlines

The Petawatt (PW) laser facility of the Berkeley Lab Laser Accelerator (BELLA) Center has recently commissioned its second laser pulse transport line. This new beamline can be operated in parallel with the first beamline and enables strong-field quantum electrodynamics (SF-QED) experiments at BELLA. In this paper, we present an overview of the upgraded BELLA PW facility with a SF-QED experimental layout in which intense laser pulses collide with GeV-class laser-wakefield-accelerated electron beams. We present simulation results showing that experiments will allow the study of laser-particle interactions from the classical to the SF-QED regime with a nonlinear quantum parameter of up to $\chi\sim$2. In addition, we show that experiments will enable the study and production of GeV-class, mrad-divergence positron beams via the Breit-Wheeler process.

What you get is not always what you see—pitfalls in solar array assessment using overhead imagery

Effective integration planning for small, distributed solar photovoltaic (PV) arrays into electric power grids requires access to high quality data: the location and power capacity of individual solar PV arrays. Unfortunately, national databases of small-scale solar PV do not exist; those that do are limited in their spatial resolution, typically aggregated up to state or national levels. While several promising approaches for solar PV detection have been published, strategies for evaluating the performance of these models are often highly heterogeneous from study to study. The resulting comparison of these methods for practical applications for energy assessments becomes challenging and may imply that the reported performance evaluations overly optimistic. The heterogeneity comes in many forms, each of which we explore in this work: the degree of diversity of the locations and sensors (e.g. different satellites, aerial photography) from which the training and validation data originate, the validation of ground truth (manual annotation of imagery vs known solar PV locations), the level of spatial aggregation (e.g. array-level vs regional estimates), and inconsistencies in the training and validation datasets (e.g. different datasets are used for each study and those data are not always made accessible). For each, we discuss emerging practices from the literature to address them or suggest directions of future research. As part of our investigation, we evaluate solar PV identification performance in two large regions: the entire state of Connecticut and the city of San Diego, CA. In Connecticut, we also use 33,114 known parcel-level solar PV installations from Berkeley Lab’s Tracking the Sun dataset to evaluate parcel-level performance and evaluate capacity estimates using 169 municipalities. We also make our code (which we call SolarMapper), pre-trained models, training data, and predictions publicly available and provide a web portal for interactively inspecting each prediction that was made. Our findings suggest that traditional performance evaluation of the automated identification of solar PV from satellite imagery may be optimistic due to common limitations in the validation process. The takeaways from this work are intended to inform and catalyze the large-scale practical application of automated solar PV assessment techniques by energy researchers and professionals.

Automatic Qubit Characterization and Gate Optimization with QubiC

As the size and complexity of a quantum computer increases, quantum bit (qubit) characterization and gate optimization become complex and time-consuming tasks. Current calibration techniques require complicated and verbose measurements to tune up qubits and gates, which cannot easily expand to the large-scale quantum systems. We develop a concise and automatic calibration protocol to characterize qubits and optimize gates using QubiC , which is an open source FPGA (field-programmable gate array)-based control and measurement system for superconducting quantum information processors. We propose multi-dimensional loss-based optimization of single-qubit gates and full XY-plane measurement method for the two-qubit CNOT gate calibration. We demonstrate the QubiC automatic calibration protocols are capable of delivering high-fidelity gates on the state-of-the-art transmon-type processor operating at the Advanced Quantum Testbed at Lawrence Berkeley National Laboratory. The single-qubit and two-qubit Clifford gate infidelities measured by randomized benchmarking are of 4.9(1.1) × 10 -4 and 1.4(3) × 10 -2 , respectively.

The 40Ar(d,p)41Ar cross section between 3–7 MeV

To determine the safety of using argon as a deuteron beam stopping material, the 40Ar(d,p)41Ar cross section was measured at average deuteron energies of 3.6 MeV, 5.5 MeV, and 7.0 MeV using an activation method. A 16-MeV deuteron beam produced by Lawrence Berkeley National Laboratory’s 88-Inch Cyclotron was degraded to each energy by nickel foils and the front wall of an aluminum gas chamber. The reduced-energy deuterons were used to activate a sample of natAr gas. After each irradiation, the gas chamber’s 41Ar activation was measured with a high-purity germanium detector. The cross sections measured were larger than a previous measurement by ∼40%.

Metal-Supported Solid Oxide Electrolysis Cells with Coatings to Suppress Chromium Migration

High-temperature electrolysis (HTE) is an efficient technology for converting steam to hydrogen. The performance, durability, and applications of metal-supported solid oxide electrolysis cell technology developed at Lawrence Berkeley National Laboratory (LBNL) will be discussed. The unique LBNL symmetric cell architecture design, with thin zirconia ceramic backbones and electrolyte sandwiched between porous metal supports, offers a number of advantages over conventional all-ceramic cells, including low-cost structural materials (e.g. stainless steel), mechanical ruggedness, excellent tolerance to redox cycling, and extremely fast start-up capability.An inherent limitation of this design is the intimate contact between the electrode and the porous stainless steel, which is a source of chromium (Cr). Cr migrates from the support to the electrode during cell fabrication and operation, and reacts with the oxygen evolution electrocatalyst (LSCF). This Cr poisoning reduces performance and durability. To overcome this limitation, Cr-blocking coatings have been applied to the porous stainless steel. A variety of coating compositions and deposition techniques have been evaluated. The best coating dramatically reduces the amount of Cr detected in the electrode both after catalyst infiltration at 800°C and after 1000 h continuous operation at 700°C, Fig 1.Other aspects of the MS-SOEC design and operation have been optimized for performance and durability. These include the infiltrated catalyst composition and processing, metal support structure, and operating temperature. Figure 1

Metal Supported Solid Oxide Fuel Cells for Ethanol Fueled Vehicles

Porous metal supported solid oxide fuel cells (MS-SOFCs) have been engineered with micro-pores and nano-sized catalysts for directly converting chemical energy in ethanol to electricity. The MS-SOFCs developed at Lawrence Berkeley National Laboratory offer advantages of rapid start-up, mechanical ruggedness, high performance, and low-cost materials. A symmetric metal/ceramic scaffold with micro-pore range of 1 to 25 μm has been fabricated using tape-casting and lamination. Select catalysts were infiltrated in the porous electrodes and metal supports by fast infiltration. Continuous cell performance improvement has been achieved by optimization of cell structure, processing temperature and catalyst compositions. The stability of MS-SOFCs has been evaluated during continuous operation for >300 hours. The degradation mechanisms revealed by electrochemical impedance spectroscopy and high-resolution scanning electron microscopy techniques will be presented. The effect of impurities such as Cr on the cathode stability has been evaluated at an intermediate operating temperature range of 600 to 700°C. The phase change of praseodymium-oxide oxygen reduction catalysts has been identified at low oxygen pressure (such as nitrogen environment) by in situ high temperature X-ray diffraction.

The effects of high energy deuteron ion beam irradiation on the tensile behavior of HT-9

Ion beam implantations are widely performed to understand the effects of irradiation-induced displacement damage on nuclear structural materials. However, the volume of material that can be investigated is often limited by the maximum energies of accelerator facilitates, leading to a limitation in the thickness of samples whose mechanical properties can be evaluated. The Lawrence Berkeley National Laboratory’s 88-Inch Cyclotron offers a wide range of ions and energies, allowing for material ion implantations at larger scales than typical. Four HT-9 SS-J-geometry tensile specimens were polished and then irradiated with deuterons at the 88-Inch Cyclotron to doses of approximately 0.2 dpa prior to small scale tensile testing. The results from this study show irradiation hardening characterized by the tensile test results and black dot irradiation defects. Additionally, a comprehensive look at low temperature irradiations of high-Cr F/M steels is presented and our results show agreement with the available data.

Picking up the pieces: Instrumental Neutron Activation Analysis (INAA) of early intermediate period and middle horizon pottery from Ayacucho, Peru

This study compares chemical data of 224 sherds from Early Intermediate Period Huarpa and Middle Horizon Wari pottery to investigate settlement dynamics. All sherds were first identified stylistically and chronologically. Data from the Lawrence Berkeley National Laboratory (LBNL) and University of Missouri Research Reactor (MURR) were combined and generated 11 chemically similar compositional groups (CGs). CGs are cultural artifacts of artisan agency. Over time and locations, CGs may determine origins, popularity, and preferences for certain local styles as well as identification of foreign pottery.

Status of the Nb$_{3}$Sn Canted-Cosine-Theta Dipole Magnet Program at Lawrence Berkeley National Laboratory

As part of the US Magnet Development Program, Lawrence Berkeley National Laboratory (LBNL) is working on the development of high field stress-managed Nb <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$_{3}$</tex-math></inline-formula> Sn dipole magnets using canted-cosine-theta (CCT) technology. As part of this program, a series of two layer magnets, CCT3/4/5, with short sample bore field of approximately 10 T and a 90 mm diameter open aperture have been designed, fabricated, and tested. The first magnet in the series, CCT3, was limited to less than 70% of the short sample current, this limitation is believed to be due to conductor damage. The second magnet in the series, CCT4, reached 86% of the short sample current, after changes to the groove geometry were made to accommodate dimensional changes in the cable during heat treatment. The third and final magnet of this series, CCT5, reached 88% of the short sample current with improved training relative to CCT4, after changes were made to the impregnation and assembly methods. While this two layer series was used to successfully demonstrate Nb <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$_{3}$</tex-math></inline-formula> Sn CCT magnet technology, improvements in the training behavior of these magnets is desirable. For this purpose, a subscale program has been devised in order to probe the causes and explore reductions / improvements to training in stress-managed magnet technology. The subscale nature of the magnets allows for faster turnaround in the fabrication and testing process. In this paper, we present the design and test results for the first (baseline) subscale Nb <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$_{3}$</tex-math></inline-formula> Sn CCT magnet and demonstrate that the subscale platform and the larger two-layer magnets produce similar training results.

Design, Construction, and Testing of 0.5-m, 18-mm Period Nb3Sn Superconducting Undulator Magnets

Design and fabrication of a new Nb <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> Sn-based superconducting undulator (SCU) are underway at the Advanced Photon Source (APS) of Argonne National Laboratory in collaboration with Fermilab and Lawrence Berkeley National Laboratory. To develop a robust and reliable fabrication process, the magnet development consists of several steps. First, magnetic and mechanical simulations were performed to optimize the magnet design; then, the design matured further by fabricating and testing a series of very short prototypes, ∼8 cm long with a period length of 18 mm. These short prototype studies were previously reported. Second, the design was scaled to an intermediate length of ∼0.5 m. These two steps led to the final design of ∼1.1–m-long magnets, which are currently being fabricated. The quench behavior of each 0.5-m-long undulator magnet, as well as undulator assemblies from these magnets, was studied. The first SCU magnet assembly did not meet the design specifications due to the breakdown of the insulation. With an improved design and fabrication process based on lessons learned, the second SCU assembly achieved the design undulator field of 1.2 T. The design was further optimized, and the third set of magnets was fabricated and successfully tested.

News Article

The materials project – Database and toolbox for materials science So far, numerous databases have been built around the world that summarize a large number of physical and chemical properties, thermodynamical diagrams, crystal structures, and some other useful experimental and calculated data. It is extremely useful for checking and confirming various properties of known materials and, conversely, for discovering promising candidates for functions and properties of interest. Furthermore, one of the remarkable recent trends is employment of deep learning, which is known as the 3rd generation of artificial intelligence, to explore some knowledge and models using already stored big data and some optimized algorithms. The system also includes a platform for quantum mechanical calculation. The Materials Project, established in 2011 by Professor Kristin Persson of Lawrence Berkeley National Laboratory, is an open-access database. Anyone can register as a user and start using it for free of charge. More than 120,000 users have already registered. From the perspective of X-ray experts, the system should be very interesting, because it includes many experimentally collected X-ray absorption spectra, the crystal structure data, and the related quantum mechanical calculation data. The total number of the registered data exceeds 500,000; X-ray spectra and crystal structure data are nearly 20% and 35%, respectively, and others are more theoretical data, such as the density of states, the bond structures and so forth. The Materials Project organizes seminars and annual workshop as well. All of the archives are available on its Youtube site (https://www.youtube.com/c/MaterialsProject). For further information, visit the following page, https://materialsproject.org/

Shell-Based Support Structure for the 45 GHz ECR Ion Source MARS-D

Superconducting electron cyclotron resonance ion sources (ECRISs) using NbTi coils and optimized for 28 GHz resonant heating have been successfully operated for almost two decades. Moving to higher heating frequencies requires increased magnetic fields, but traditional racetrack-and-solenoid ECRIS structures are at their limit using NbTi. Rather than moving to a superconductor untested in this field, the Mixed Axial and Radial field System (MARS) being developed at Lawrence Berkeley National Laboratory employs a novel closed-loop-coil design that more efficiently utilizes conductor fields and will allow the use of NbTi in a next-generation, 45 GHz ECRIS. This article presents the design of the shell-based support structure central to the MARS-D magnet design, as well as structural analysis of its components and optimization of pre-load parameters that will guarantee its successful operation.

Magnetotelluric Investigations of the Kīlauea Volcano, Hawaii

AbstractIn 2002 and 2003 a collaborative effort was undertaken between Lawrence Berkeley National Laboratory, Sandia National Laboratories, the U.S. Geological Survey (USGS) Menlo Park, the USGS Hawaiian Volcano Observatory, and Electromagnetic Instruments Inc. to study the Kīlauea volcano in Hawaii using the magnetotelluric (MT) technique. The work was motivated by a desire to improve understanding of the magma reservoirs and conduits within Kīlauea and the East and Southwest Rift zones, which has implications for understanding Kīlauea's plumbing system. An improved understanding of the rift zones has implications in understanding large‐scale landslides that are generated in the Hilina Slump, which produce significant impacts on coastal communities. Up to eight stations operated simultaneously, with multiple remote reference sites, and data were processed using multi‐station robust processing techniques. In total, data were acquired at 70 sites over the Southwest and East rift zones. Good to excellent quality data were obtained even in the harshest conditions, such as those encountered on the fresh lava flows of the East Rift Zone, where electrical contact resistances are on the order of 100 kΩ. A three‐dimensional (3D) MT model study was done to guide interpretation of the observed MT measurements. Synthetic modeling demonstrates that conductive bodies in the upper 3 km can be spatially resolved where MT station sampling is good. Resistivity anomalies in the 3D inversions have a high degree of spatial correlation with previously published seismic velocity anomalies beneath Kīlauea. Melt fractions between 0.096 and 0.117 are calculated for the Kīlauea and Puʻuʻōʻō low resistivity anomalies, respectively.

Challenges in temperature measurements in gas-phase photothermal catalysis

Luca Mascaretti completed his PhD in 2018 at Politecnico di Milano, Italy. Afterwards, he has worked as a postdoc researcher at the Regional Centre of Advanced Technologies and Materials, Palacký University in Olomouc (Czech Republic). His research activity is related to titanium oxide/nitride-based nanostructured materials for solar energy conversion processes. Andrea Schirato is currently a physics PhD candidate across Politecnico di Milano and the Italian Institute of Technology (Genoa) under the supervision of Profs. G. Della Valle and R. Proietti Zaccaria. His research activities focus on the theoretical study and numerical modeling of ultrafast nonlinear phenomena driven by hot carriers, including non-equilibrium electronic and phononic energy transfer in nanostructured materials and metasurfaces. Tiziano Montini is associate professor at the University of Trieste in Italy. He received his PhD in chemistry from the University of Trieste in 2006, followed by a postdoc fellowship. His research activity concerns on nanomaterials as catalysts for renewable energy production and environmental protection. Alessandro Alabastri is currently an assistant professor in the Electrical and Computer Engineering Department at Rice University. He received his PhD in nanosciences in 2014 from the Italian Institute of Technology and the University of Genoa. In 2015, he was visiting researcher in the Molecular Foundry at the Lawrence Berkeley National Laboratory. He worked on several aspects of light-to-heat conversion, photothermal effects in nanostructures, and heat recovery in light-powered thermofluidic devices. Alberto Naldoni is currently an associate professor at University of Turin, Italy. Since 2017, he is the co-leader of the photoelectrochemistry group at the Regional Centre of Advanced Technologies and Materials, Palacký University Olomouc (Czech Republic). He obtained his PhD in chemical sciences from University of Milan (2010). From 2014 to 2017, he was visiting research faculty at the Birck Nanotechnology Center of Purdue University. His group focuses on nanomaterials for plasmonics, photocatalysis, and photoelectrochemistry. Paolo Fornasiero is a professor at the University of Trieste in Italy. He received his PhD in chemistry from the University of Trieste in 1997, followed by a postdoc fellowship at Reading University (UK). His group currently focuses on research into nanomaterials for environmental and energy-related applications.

Design and implementation of dynamic I/O control scheme for large scale distributed file systems

In this work, we have analyzed the input/output (I/O) activities of Cori, which is a high-performance computing system at the National Energy Research Scientific Computing Center at Lawrence Berkeley National Laboratory. Our analysis results indicate that most users do not adjust storage configurations but rather use the default settings. In addition, owing to the interference from many applications running simultaneously, the performance varies based on the system status. To configure file systems autonomously in complex environments, we developed DCA-IO, a dynamic distributed file system configuration adjustment algorithm that utilizes the system log information to adjust storage configurations automatically. Our scheme aims to improve the application performance and avoid interference from other applications without user intervention. Moreover, DCA-IO uses the existing system logs and does not require code modifications, an additional library, or user intervention. To demonstrate the effectiveness of DCA-IO, we performed experiments using I/O kernels of real applications in both an isolated small-sized Lustre environment and Cori. Our experimental results shows that our scheme can improve the performance of HPC applications by up to 263% with the default Lustre configuration.