Search

To meet the exascale I/O requirements for the High-Performance Computing (HPC), a new I/O subsystem, Burst Buffer, based on solid state drives (SSD), has been developed. However, the diverse HPC workloads and the bursty I/O pattern cause severe data fragmentation that requires costly garbage collection (GC) and increases the number of bytes written to the SSD. To address this data fragmentation challenge, a new multi-stream feature has been developed for SSDs. In this work, we develop an I/O Separation scheme called BIOS to leverage this multi-stream feature to group the I/O streams based on the user IDs. We propose a stream-aware scheduling policy based on burst buffer pools in the workload manager, and integrate the BIOS with the workload manager to optimize the I/O separation scheme in burst buffer. We evaluate the proposed framework with a burst buffer I/O traces from Cori Supercomputer including a diverse set of applications. Experimental results show that the BIOS could improve the performance by 1.44x on average and reduce the Write Amplification Factor (WAF) by up to 1.20x. These demonstrate the potential benefits of the I/O separation scheme for solid state storage systems.

Purpose Silicone oil (SO) is commonly used for tamponade purposes in retinal detachment (RD) surgery, but the long-term visual prognosis after removal of the oil, and in particular, what is known about the recurrence of RD after SO removal, remains unclear. The purpose of this study is to evaluate the long-term vision prognosis after SO removal, and to understand the frequency and characteristics of RD recurrence. Methods We retrospectively reviewed the medical charts of 1017 eyes of patients with a diagnosis of RD who had a pars plana vitrectomy with SO tamponade between January 2009 and December 2018. Best-corrected visual acuity (BCVA) was obatained before and after vitrectomy and also at the last visit. After SO removal, the group who showed improvement in visual acuity and the group who did not were compared. The anatomical results were compared between the group in which the retina was detached again after SO removal and the group in which the retina was not detached. To determine whether the duration of SO tamponade affects RD recurrence, further analysis was performed by dividing subgroups according to SO tamponade duration. RD recurrence, visual acuity, SO tamponade period were investigated. Results Mean follow-up period was 56.65 ± 72.02 months. An average SO tamponade period was 6.68 ± 11.39 months. The average logMAR BCVA was 1.75 ± 0.91 before SO injection, 1.60 ± 0.75 before SO removal and 1.29 ± 0.96 after the removal. After SO removal, 926 of the 1017 (91.1%) patients had well attached retina without recurrence. There was no significant difference in visual acuity before SO removal in re-detachment group compared to no re-detachment group, but visual acuity of re-detachment group was worse than no re-detachment group after SO removal (p Conclusion The prognosis for final vision is generally good when performing surgery using SO in RD, but considering the complications that arise after surgery, long-term retention of SO is not recommended and the timing of SO removal should be considered.

To evaluate the choriocapillaris (CC) flow in central serous chorioretinopathy (CSC) and determine the relationship between CC flow void with the choroidal thickness (CT) and choroidal vascularity index (CVI).Retrospective analysis of 20 patients with CSC (40 eyes, including unaffected fellow eyes) and 20 age- and sex-matched controls. After compensation with optical coherence tomography (OCT) en-face structural image, the CC flow void (%) was measured using the phansalkar threshold with a window radius of 3 and 15 pixels. The mean CC flow voids of acute CSC, recovered-acute CSC, unaffected fellow, and control eyes were compared by matched data analysis. A regression analysis was performed on the choroidal parameters (CT and CVI) and CC flow voids.The CC flow void had an increasing tendency in the following order: control, fellow, recovered-acute CSC, and acute CSC eyes. Acute/recovered comparison showed a significant P value (0.008) in the foveal lesion. Recovered/fellow and fellow/control presented significant P values regardless of location to fovea (all0.05). There were significant positive correlations between CT and CC flow void (P 0.05) in the acute CSC, recovered-acute CSC eyes.The CC flow on OCT angiography decreased in acute CSC eyes, especially in the foveal lesion, with a published compensation method. The findings suggest that unmodulated choroidal blood flow contributed to partially reversible diminished CC flow.

This study evaluated the effects of cataract surgery combined with pars plana vitrectomy (ppV) on choroidal vascularity index (CVI) in eyes with epiretinal membrane (ERM) and full thickness macular hole (FTMH). Medical records of 132 eyes with ERM or FTMH were retrospectively reviewed and classified into a ppV group and a ppV combined with cataract surgery group (phaco + ppV group). The CVI were measured at baseline, 1, 3 and 6 months after the surgery, using the selected swept-source (SS) optical coherence tomography (OCT) scan passing through the central fovea, which was then segmented into luminal and stromal area by image binarization. The mean CVI of phaco + ppV group were 61.25 ± 1.97%, 61.66 ± 1.81%, and 62.30 ± 1.92% at baseline, 1 and 3 months, respectively (p p p p

Purpose To evaluate the incidence and risk factors of neovascularization in unaffected fellow eyes of patients diagnosed with type 3 neovascularization in Korea. Methods This retrospective study included 93 unaffected fellow eyes of 93 patients diagnosed with type 3 neovascularization. For initial type 3 neovascularization diagnosis, optical coherence tomography and angiography were conducted. These baseline data were compared between patients with and without neovascularization in their fellow eyes during the follow-up period. Results The mean follow-up period was 66.1±31.1 months. Neovascularization developed in 49 (52.8%) fellow eyes after a mean period of 29.5±19.6 months. In the fellow eye neovascularization group, the incidence of soft drusen and reticular pseudodrusen was significantly higher than that in the non-neovascularization group (83.7% vs. 36.5%, p Conclusion Neovascularization developed in 52.8% of unaffected fellow eyes. The presence of soft drusen, reticular pseudodrusen, and lower CVI values can be considered risk factors of neovascularization in unaffected fellow eyes of patients with type 3 neovascularization. The lower CVI values suggest that choroidal ischemic change may affect the development of choroidal neovascularization in these patients.

Purpose: To report a rare case of Vogt-Koyanagi-Harada (VKH) patient with bilateral annular choroidal detachment and describe ultra-widefield angiographic findings.Method: Case report.Results: A 61-year-old male was diagnosed as VKH based on ocular signs and neurologic findings. Annular choroidal detachment anterior to the equator was noted in both eyes. Ultra-wide field (UWF) fluorescein angiography revealed multiple leakages at posterior pole and disc with dye pooling at the margin of choroidal detachment. Late-phase UWF indocyanine green angiography (ICGA) revealed hyperpermeability with dilated veins and many vortex ampullae obscured by detached choroid. Choroidal detachment and other manifestations improved after a high-dose systemic steroid treatment. UWF ICGA after the treatment revealed decreased choroidal vessel caliber with visible vortex ampullae.Conclusion: Vortex obstruction and choroidal vein dilatation were observed in a VKH patient with choroidal detachment, which might be related to its pathophysiology.

Hadoop is becoming widely adopted in scientific and commercial areas as an open-source distributed data processing framework. Recently, for real-time processing and analysis of data, an attempt to apply high-performance computing technologies to Hadoop is being made. In this paper, we have expanded the Hadoop Filesystem library to support Lustre, which is a popular high-performance parallel distributed filesystem, and implemented the Hadoop MapReduce execution environment over the Lustre filesystem. We analysed Hadoop MapReduce over Lustre by using Hadoop standard benchmark tools. We found that Hadoop MapReduce over Lustre execution has a performance 2-13 times better than a typical Hadoop MapReduce execution.

Genome sequence alignment is one of the basic procedure of genome sequencing analysis pipeline and also one of the most time-consuming parts. Including BigBWA, a number of tools were proposed to accelerate genome sequence alignment by parallelizing computation with Hadoop technologies. However, HDFS incurs considerable I/O overhead. In this research, we propose a new sequence alignment tool adopting Hadoop on Lustre. Based on BigBWA, we removed data transfer overhead caused by HDFS and parallelized whole I/O steps. Experimental result shows that our solution is five times faster than original BigBWA in a ten-node Lustre based Hadoop cluster.

AMGA (ARDA Metadata Grid Application) is a grid metadata catalog system, that is part of the gLite middleware, developed for the EGEE (Enabling Grids for E-sciencE) project. Since the end of EGEE, gLite has participated in the European Middleware Initiative (EMI) with other major middleware systems such as ARC, dCache and UNICORE. Through AMGA’s three-year EMI collaboration, it has been actively developed, tested and deployed. Previously, AMGA had made progress by focusing on development for the purposes of function implementation and improvement as a metadata catalogue system. After participating in the EMI, it has designed and implemented middleware-interoperability functions according to EMI policies and strategies.

In this paper, a dynamic model is proposed for multi-axis optical structure of an aircraft. Modal analysis, sine-wave analysis, and random vibration analysis are done using a multi-body dynamic program for the multi-axis optical structure. By applying Al 6061-T6 and two types of CFRP to the camera module, x, y, and z responses are found and compared according to each axis excitation. The results will be used for reducing the weight of the camera module.

This paper presents architectural and algorithmic approaches for achieving high-speed CORDIC processing in both of the two operating modes: vectoring and rotation. For vectoring mode CORDIC processing, a modified architecture is proposed, which aims at reduction of computation time by overlapping the stages for redundant addition and selection of rotation direction. In addition, a novel rotation direction prediction scheme for rotation mode CORDIC is presented. The method is based on approximation of the binary angle input to a number with the arctangent weights (tan^−1 2^−i). The implementation is designed to keep the fast timing characteristics of redundant arithmetic in the x/y path of the CORDIC processing. The characteristics are analyzed with respect to latency time and area, and compared with those obtained by conventional CORDIC implementations. The results show that the proposed techniques reduce not only the block latency but also the overall computation time. Thus, they achieve higher throughput in pipelining.

AMGA (ARDA Metadata Grid Application) is one of main EMI (European Middleware Initiative) middleware software for the EMI project which started from 2010 and lasted 3 years. In this time, AMGA metadata catalog technology had the good opportunity to enlarge the range of relation with other EMI middleware components. Especially, AMGA interoperability with several EMI components paved the way for providing usefulness of management and searching available AMGA resource to AMGA users. Additionally, AMGA GUIs (AMGA Manager and AMGA Web application) have been developed focusing on user’s utilization. In this article, we present not only AMGA development experience of in the EMI project but also other activities related to expand use of AMGA.

This paper describes the recent improvement of the AMGA (ARDA Metadata Grid Application) python client library for the Belle II Experiment. We were drawn to the action items related to library improvement after in-depth discussions with the developer of the Belle II distributed computing system. The improvement includes client-side metadata federation support in python, DIRAC SSL library support as well as API refinement for synchronous operation. Some of the improvements have already been applied to the AMGA python client library as bundled with the Belle II distributed computing software. The recent mass Monte- Carlo (MC) production campaign shows that the AMGA python client library is reliably stable.

AMGA (ARDA Metadata Grid Application) is a grid metadata catalogue system that has been developed as a component of the EU FP7 EMI consortium based on the requirements of the HEP (High-Energy Physics) and the biomedical user communities. Currently, AMGA is exploited to manage the metadata in the gBasf2 framework at the Belle II experiment, one of the largest particle physics experiments in the world. In this paper, we present our efforts to optimize the metadata query performance of AMGA to better support the massive MC Campaign of the Belle II experiment. Although AMGA exhibits very outstanding performance for a relatively small amount of data, as the number of directories and the metadata size increase (e.g. hundreds of thousands of directories) during the MC Campaign, AMGA suffers from severe query processing performance degradation. To address this problem, we modified the query search mechanism and the database scheme of AMGA to provide dramatic improvements of metadata search performance and query response time. Throughout our comparative performance analysis of metadata search operations, we show that AMGA can be an optimal solution for a metadata catalogue in a large-scale scientific experimental framework

Cloud computing is a model for the delivery of IT resources as a service than a product. One of the essential characteristics in the cloud computing is on-demand computing. It can be explained that users can reserve and utilize IT capabilities, such as computing or network resources, at a certain time they want. As a part of HERO project, we developed HERO web portal which has on-demand capabilities of co-allocating computing and network resources. It is on top of TIGRIS middleware platform, which is developed based on Globus Toolkit 4 WSRF-compatible Grid services. With HERO web portal, Users can utilize HPC clusters and high-performance network between them in an on-demand manner.

This practices and experience paper describes the coordination, design, implementation, availability, and performance of the Pacific Rim Applications and Grid Middleware Assembly (PRAGMA) Grid Testbed. Applications in high-energy physics, genome annotation, quantum computational chemistry, wildfire simulation, and protein sequence alignment have driven the middleware requirements, and the testbed provides a mechanism for international users to share software beyond the essential, de facto standard Globus core. In this paper, we describe how human factors, resource availability and performance issues have affected the middleware, applications and the testbed design. We also describe how middleware components in grid monitoring, grid accounting, grid Remote Procedure Calls, grid-aware file systems, and grid-based optimization have dealt with some of the major characteristics of our testbed. We also briefly describe a number of mechanisms that we have employed to make software more easily available to testbed administrators.

This paper presents an architecture for accelerating CORDIC vectoring mode operations. The processing is sped up by overlapping redundant sum formation and selection of rotation direction. We analyze the latency time and area, and compare them with a conventional CORDIC implementation. The results show that the proposed scheme reduces not only the the latency but also the overall computation time. Thus, it achieves higher throughput in pipelining.

Presents a low-cost approach to concurrent error correction in high-performance CORDIC processors by using time-shared triple modular redundancy. Operands are partitioned into three sets of disjoint digits and operations are performed three times on different hardware components to correct possible errors by majority voting. The approach has limited latency increase and throughput reduction. Pipelining can be used to maintain the same throughput as a conventional design.