Building a High-Performance Graph Storage on Top of Tree-Structured Key-Value Stores

Graph databases have gained widespread adoption in various industries and have been utilized in a range of applications,including financial risk assessment,commodity recommendation,and data lineage tracking.While the principles and design of these databases have been the subject of some investigation,there remains a lack of comprehensive examination of aspects such as storage layout,query language,and deployment.The present study focuses on the design and implementation of graph storage layout,with a particular emphasis on tree-structured key-value stores.We also examine different design choices in the graph storage layer and present our findings through the development of TuGraph,a highly efficient single-machine graph database that significantly outperforms well-known Graph DataBase Management System(GDBMS).Additionally,TuGraph demonstrates superior performance in the Linked Data Benchmark Council(LDBC)Social Network Benchmark(SNB)interactive benchmark.

Towards High-Performance Graph Processing: From a Hardware/Software Co-Design Perspective

Graph processing has been widely used in many scenarios,from scientific computing to artificial intelligence.Graph processing exhibits irregular computational parallelism and random memory accesses,unlike traditional workloads.Therefore,running graph processing workloads on conventional architectures(e.g.,CPUs and GPUs)often shows a significantly low compute-memory ratio with few performance benefits,which can be,in many cases,even slower than a specialized single-thread graph algorithm.While domain-specific hardware designs are essential for graph processing,it is still challenging to transform the hardware capability to performance boost without coupled software codesigns.This article presents a graph processing ecosystem from hardware to software.We start by introducing a series of hardware accelerators as the foundation of this ecosystem.Subsequently,the codesigned parallel graph systems and their distributed techniques are presented to support graph applications.Finally,we introduce our efforts on novel graph applications and hardware architectures.Extensive results show that various graph applications can be efficiently accelerated in this graph processing ecosystem.

Accelerated design of high-performance Mg-Mn-based magnesium alloys based on novel bayesian optimization

Magnesium(Mg),being the lightest structural metal,holds immense potential for widespread applications in various fields.The development of high-performance and cost-effective Mg alloys is crucial to further advancing their commercial utilization.With the rapid advancement of machine learning(ML)technology in recent years,the“data-driven''approach for alloy design has provided new perspectives and opportunities for enhancing the performance of Mg alloys.This paper introduces a novel regression-based Bayesian optimization active learning model(RBOALM)for the development of high-performance Mg-Mn-based wrought alloys.RBOALM employs active learning to automatically explore optimal alloy compositions and process parameters within predefined ranges,facilitating the discovery of superior alloy combinations.This model further integrates pre-established regression models as surrogate functions in Bayesian optimization,significantly enhancing the precision of the design process.Leveraging RBOALM,several new high-performance alloys have been successfully designed and prepared.Notably,after mechanical property testing of the designed alloys,the Mg-2.1Zn-2.0Mn-0.5Sn-0.1Ca alloy demonstrates exceptional mechanical properties,including an ultimate tensile strength of 406 MPa,a yield strength of 287 MPa,and a 23%fracture elongation.Furthermore,the Mg-2.7Mn-0.5Al-0.1Ca alloy exhibits an ultimate tensile strength of 211 MPa,coupled with a remarkable 41%fracture elongation.

Optimization Techniques for GPU-Based Parallel Programming Models in High-Performance Computing

This study embarks on a comprehensive examination of optimization techniques within GPU-based parallel programming models,pivotal for advancing high-performance computing(HPC).Emphasizing the transition of GPUs from graphic-centric processors to versatile computing units,it delves into the nuanced optimization of memory access,thread management,algorithmic design,and data structures.These optimizations are critical for exploiting the parallel processing capabilities of GPUs,addressingboth the theoretical frameworks and practical implementations.By integrating advanced strategies such as memory coalescing,dynamic scheduling,and parallel algorithmic transformations,this research aims to significantly elevate computational efficiency and throughput.The findings underscore the potential of optimized GPU programming to revolutionize computational tasks across various domains,highlighting a pathway towards achieving unparalleled processing power and efficiency in HPC environments.The paper not only contributes to the academic discourse on GPU optimization but also provides actionable insights for developers,fostering advancements in computational sciences and technology.

Measurement and analysis of defects in high-performance concrete with three-dimensional micro-computer tomography

In order to investigate the effects of two mineral admixtures （i. e., fly ash and ground slag）on initial defects existing in concrete microstructures, a high-resolution X-ray micro-CT（ micro-focus computer tomography）is employed to quantitatively analyze the initial defects in four series of highperformance concrete （HPC）specimens with additions of different mineral admixtures. The nigh-resolution 3D images of microstructures and filtered defects are reconstructed by micro- CT software. The size distribution and volume fractions of initial defects are analyzed based on 3D and 2D micro-CT images. The analysis results are verified by experimental results of watersuction tests. The results show that the additions of mineral admixtures in concrete as cementitious materials greatly change the geometrical properties of the microstructures and the spatial features of defects by physical-chemistry actions of these mineral admixtures. This is the major cause of the differences between the mechanical behaviors of HPC with and without mineral admixtures when the water-to-binder ratio and the size distribution of aggregates are constant.

Preliminary Studies on High-performance Liquid Chro-matography Chemiluminescence Determination of theSaturated Fatty Acids(C_(16) C_(18))in Human Serum

The peroxyoxalate chemiluminescence(CL)detection of fatty acids in human se- rum combined with high-performance liquid chromatography (HPLC)is described.Some fatty acids in serum were extracted with a 1 :1(v/v)mixture of chloroform-n-heptane.2-(4-Hydrazinocarbonyl- phenyl)-4,5-diphenylimidazole (HCPI) was used as a fluorescent labelling reagent of the fatty acids. The labelling reaction was carried out at 30℃ for 1 h at pH 6.5 and the resulting reaction mixture was sudjected to HPLC. The labelled fatty acid C_(17)(P-C_(17))was used as the internal standard. The la- belled fatty acids C_(16) and C_(18) were separated within 18 min on an ODS-8OTM column (150 mm× 6 mm ID,5μm,Tosoh Japan).The calibrlation curves of fatty acids from the spiked control serum were Y_1=-0.003 7 + 0.0028X_1,r=0.994 for FA C_( 16) and Y_2=0.00 1 2 + 0.00098X_2,r=0.999 for FA C_( 18),respectively.The average recoveries of facids from the spiked contrl serum were 107.2%(n=8,RSD=4.3%)for FA C_(16) and 97.35%(n=8, RSD=4.0%)for FA C_(18),respectively.The lower detection limits of fatty acids after reaction were 12μmol per 20μl injection for FA C_(16) and 18 μmol per 20μl injection for FA C_(18),respectively(signal to noise ratio, S/N=2).The HPLC/CL method was applied to the determination of FA C_(16) and FA C_(18) in normal human serum and the results showed that the concentrations of fatty acids in normal human serum were 0.134 ± 0.009 μ mol/ml serum(n=5) for FA C_(16) and 0.052±0.028 μmol/ml serum(n=5)for FA C_(18),respectively.

Graph Transformers研究进展综述

随着图结构数据在各种实际场景中的广泛应用,对其进行有效建模和处理的需求日益增加。Graph Transformers(GTs)作为一类使用Transformers处理图数据的模型,能够有效缓解传统图神经网络(GNN)中存在的过平滑和过挤压等问题,因此可以学习到更好的特征表示。根据对近年来GTs相关文献的研究,将现有的模型架构分为两类:第一类通过绝对编码和相对编码向Transformers中加入图的位置和结构信息,以增强Transformers对图结构数据的理解和处理能力;第二类根据不同的方式(串行、交替、并行)将GNN与Transformers进行结合,以充分利用两者的优势。介绍了GTs在信息安全、药物发现和知识图谱等领域的应用,对比总结了不同用途的模型及其优缺点。最后,从可扩展性、复杂图、更好的结合方式等方面分析了GTs未来研究面临的挑战。

Simultaneous determination of five nucleosides and nucleobases in Panax notoginseng using high-performance liquid chromatography

Aim To quantitatively determine five nucleosides and nucleobases, including cytidine, uridine, guanosine, adenosine and uracil in different parts of Panax notoginseng. Methods Separation was performed on a Zorbax SB-Aq column using a gradient elution with mobile phase of 8 mmol^L-1 ammonium acetate aqueous solution （A） and methanol （B）. The assay was carried out at a flow rate of 1 mL·min^-1 at 25 ℃ with the diode-array detection at 260 nm. Results Cytidine, uridine, guanosine, adenosine and uracil had good linearity in the ranges of 1.79 - 57.40 μg·mL^-1 （r^2 = 1.0000）, 3.30 - 105.60 μg·mL^-1 （r^2 = 1.0000）, 3.09 - 98.80 μg·mL^ -1（r^2 = 0.9999）, 2.77 - 88.60 μg·mL^-1 （r^2 = 1.0000） and 0.38 - 12.30 μg·mL ^-1 （r^2 = 1.0000） with average recoveries of 93.9%, 96.5%, 92.7%, 93.2% and 98.8%, respectively. The content of cytidine, uridine, guanosine, adenosine and uracil in different parts of P. notogingeng were significantly different. Conclusion This is the first report on quantitative determination of nucleosides and nucleobases in P notoginseng.

High-Performance Liquid Chromatographical Analysis of Ginsenosides in Panax ginseng,P.quinquef(?)lium and P.notoginseng

The compositions and contents of ginsenbsides in Panax ginseng,P.quinquefolium and P.notoginseng were determined and compared by reversed-phase High-Performance Liquid Chro- matography(HPLC).The method was performed on an Alltech Adsorbosphere HS C_(18) column,using 5×10^(-3)M NaH_2PO_4-H_3PO_4 buffer solution(pH 3.0)and acetonitrile-water(50:50)as gradient eluents. The baseline separation of ginsenosides Rb_1,Rb_2,Rb_1,Rc,Rd,Rf,Ro,and Re+Rg_1 was obtained in one analytical run.The ginsenosides are directly detected at 203 nm.The detection limit is 40μg at a signal to noise ratio of 3:1.The improved sample preparation and clean-up prior to injection with SEP-PAK C_(18)cartridge strongly reduced the front peaks caused by the impurities in the methanolic extracts of samples to afford a smooth baseline and clear background.The HPLC patterns of methanolic extracts mainly including the ginsenosides were found capable of serving as chemical fingerprints to differentiate the three species from each other.It was also found that there are no significant diffe- rences of the HPLC patterns between the wild Panax ginseng and the cultivated,the white and the red ginsengs,Chinese and Korean red ginsengs,and the tap roots of Panax ginseng collected in four consecutive months,only certain differences in contents of ginsenosides do exist.The contents of the nine major ginsenosides present in the rhizome,tap root and rootlet as well as the leaf of Panax quinquefolium were also determined and compared.

基于Graph Transformer的半监督异配图表示学习模型

现有的图卷积网络(GCN)模型基于同配性假设,无法直接应用于异配图的表示学习,且许多异配图表示学习的研究工作受消息传递机制的限制,导致节点特征混淆和特征过度挤压而出现过平滑问题。针对这些问题,提出一种基于Graph Transformer的半监督异配图表示学习模型HPGT(HeteroPhilic Graph Transformer)。首先,使用度连接概率矩阵采样节点的路径邻域,再通过自注意力机制自适应地聚合路径上的节点异配连接模式,编码得到节点的结构信息,用节点的原始属性信息和结构信息构建Transformer层的自注意力模块;其次,将每个节点自身的隐层表示与它的邻域节点的隐层表示分离更新以避免节点通过自注意力模块聚合过量的自身信息,再把每个节点表示与它的邻域表示连接,得到单个Transformer层的输出,另外,将所有的Transformer层的输出跳连到最终的节点隐层表示以防止中间层信息丢失;最后,使用线性层和Softmax层将节点的隐层表示映射到节点的预测标签。实验结果表明,与无结构编码(SE)的模型相比,基于度连接概率的SE能为Transformer层的自注意力模块提供有效的偏差信息,HPGT平均准确率提升0.99%~11.98%;与对比模型相比,在异配数据集(Texas、Cornell、Wisconsin和Actor)上,模型节点分类准确率提升0.21%~1.69%,在同配数据集(Cora、CiteSeer和PubMed)上,节点分类准确率分别达到了0.8379、0.7467和0.8862。以上结果验证了HPGT具有较强的异配图表示学习能力,尤其适用于强异配图节点分类任务。

Plasma free amino acid profiling of esophageal cancer using high-performance liquid chromatography spectroscopy

AIM: To perform plasma free amino acid (PFAA) profiling of esophageal squamous cell carcinoma (ESCC) patients at different pathological stages and healthy subjects.

Advancements in the preparation of high-performance liquid chromatographic organic polymer monoliths for the separation of small-molecule drugs

The various advantages of organic polymer monoliths, including relatively simple preparation processes,abundant monomer availability, and a wide application range of pH, have attracted the attention of chromatographers. Organic polymer monoliths prepared by traditional methods only have macropores and mesopores, and micropores of less than 50 nm are not commonly available. These typical monoliths are suitable for the separation of biological macromolecules such as proteins and nucleic acids, but their ability to separate small molecular compounds is poor. In recent years, researchers have successfully modified polymer monoliths to achieve uniform compact pore structures. In particular, microporous materials with pores of 50 nm or less that can provide a large enough surface area are the key to the separation of small molecules. In this review, preparation methods of polymer monoliths for high-performance liquid chromatography, including ultra-high cross-linking technology, post-surface modification, and the addition of nanomaterials, are discussed. Modified monolithic columns have been used successfully to separate small molecules with obvious improvements in column efficiency.

Expansion Performance and Microstructure of High-performance Concrete using Differently Scaled MgO Agents and Mineral Powder

To investigate the assumptions proposed in this paper,the evolution law governing the strength and expansion performance of MgO and nano-MgO micro-expansive concrete in the environment of mineral powder was firstly observed in this study.Secondly,SEM,XRD,and TG-DSC microscopic tests were conducted to reveal the effects of the active mineral-powder admixture on the hydration degree and expansion performance of MgO and nano-MgO in HPC.Our experimental results successfully verified our hypothesis,which indicated that the expansion performance of macro-MgO and nano-MgO was indeed depressed by the addition of active mineral power admixtures,even though the mechanical property of concrete composites was effectively improved.Furthermore,the hydration test also demonstrated the negative interference on the mineral powders,which was induced by the expansion agents.It is found the amounts of hydrates tend to decrease because the mineral powder ratio reaches and exceeds 40%.Moreover,it is also concluded the effect of expansion agents is governed by the alkalinity cement paste,especially for the nano-MgO.In other words,the expansion performance of nano-MgO will vary more obviously with the hydration process,than MgO.The results of this study provide that effective experimental and theoretical data support the hydration-inhibition mechanism of magnesium expansive agents.

Determination of gouty arthritis' biomarkers in human urine using reversed-phase high-performance liquid chromatography

Creatinine, uric acid, hypoxanthine and xanthine are important diagnostic biomarkers in human urine for gouty arthritis or renal disease diacrisis. A simple method for simultaneous determination of these biomarkers in urine based on reversed-phase high-performance liquid chromatography （RP-HPLC） with ultraviolet （UV） detector was proposed. After pretreatment by dilution, centrifugation and filtration, the biomarkers in urine samples were separated by ODS-BP column by elution with methanol/50 mM NaH2PO4 buffer solution at pH 5.26 （5：95）. Good linearity between peak areas and concentrations of standards was obtained for the biomarkers with correlation coefficients in the range of 0.9957-0.9993. The proposed analytical method has satisfactory repeatability （the recovery of data in a range of creatinine, uric acid, hypoxanthine and xanthine was 93.49-97.90%, 95.38-96.45%, 112.46-115.78%and 90.82-97.13%with standard deviation of o5%, respectively） and the limits of detection （LODs, S/N Z 3） for creatinine, uric acid, hypoxanthine, and xanthine were 0.010, 0.025, 0.050 and 0.025 mg/L, respectively. The established method was proved to be simple, accurate, sensitive and reliable for the quantitation of gouty arthritis＇ biomarkers in human urine samples. The ratio of creatinine to uric acid was found to be a possible factor for assessment of gouty arthritis.

High-Performance Liquid Chromatographic Method for Determination of Germacrone in Rat Plasma

Aim A reliable high-performance liquid chromatographic (HPLC) method wasdeveloped for determination of germacrone in rat plasma. Methods The plasma samples were treatedwith acetonitrile and analyzed by HPLC with UV detection at 244 nm. Results The limit of detectionwas 100 ng·mL^(-1) for germacrone in plasma and the linear range was 0.1004-15.06 μg·mL^(-1) inplasma. The RSD of intra-day and inter-day assay was 1.87% - 4.29% and 1.29% -5.15%, respectively.The recoveries of germacrone were over 95%. The endogenous substances in plasma did not show anyinterference in the analysis. Conclusion The method is accurate and convenient, and suitable forpharmacokinetic studies of germacrone in rats.

Dispersive Liquid-liquid Microextraction Combined with High-performance Liquid Chromatography for the Determination of Clozapine and Chlorpromazine in Urine

A simple method has been proposed for the determination of clozapine （CLZ） and chlorpromazine （CPZ） in human urine by dispersive liquid-liquid microextraction （DLLME） in combination with high-performance liquid chromatography-ultraviolet detector （HPLC-UV）. All important variables influencing the extraction efficiency, such as pH, types of the extraction solvent and the disperser solvent and their volume, ionic strength and centrifugation time were investigated and optimized. Under the optimal conditions, the limit of detection （LODs） and quantification （LOQs） of the method were 13 and 39 ng/mL for CLZ, and 2 and 6 ng/mL for CPZ, respectively. The relative standard deviations （RSDs） of the targets were less than 5.1% （C=0.100 μg/mL, n=9）. Good linear behaviors over the tested concentration ranges were obtained with the values of R20.999 for the targets. The absolute extraction efficiencies of CLZ and CPZ from the spiked blank urine samples were 98.3% and 97.8%, respectively. The applicability of the technique was validated by analyzing urine samples and the mean recoveries for spiked urine samples ranged from 93.3% to 105.0%. The method was successfully applied for the determination of CLZ and CPZ in real human urine.

Studies on Chromatography Fingerprint of Hongqi by High-performance Liquid Chromatography

Chromatography fingerprint (CFP) of 10 samples of hongqi were studied. 23 common peaks were analyzed, their average similarity was 97.29%. CFP were positioned with main index composition such as formononetin, calycosin and then the contents of index composition were determined. The character and exclusive of CFP of 10 samples of hongqi were clear. CFP and content determination of index composition of hongqi could be used to evaluate the quality of hongqi comprehensively.

Constructing a multifunctional mesoporous composite of metallic cobalt nanoparticles and nitrogen-doped reduced graphene oxides for high-performance lithium-sulfur batteries

Electrochemical properties of lithium-sulfur(Li-S)batteries are mainly hindered by both the insulating nature of elemental sulfur(i.e.,molecular S8)and the shuttling effect or sluggish redox kinetics of lithium polysulfide intermediates(Li_(2)S_(n),3≤n≤8).In this paper,a three-dimensional mesoporous reduced graphene oxide-based nanocomposite,with the embedding of metallic Co nanoparticles and the doping of elemental N(Co/NrGO),and its simply ground mixture with powdered S at a mass ratio of 1:6(Co/NrGO/S)are prepared and used as cathode-/separator-coated interlayers and working electrodes in assembled Li-S cells,respectively.One of the effective cell configurations is to paste composite Co/NrGO onto both the S-loading cathode and separator,showing good cycling stability(1070mAh g^(−1) in the 100th cycle at 0.2 C),highrate capability(835mAh g^(−1),2.0 C),and excellent durability(905mAh g^(−1) in the 250th cycle at 0.5 or 0.2 C).Compared with the experimental results of Co-absent NrGO,electrochemical properties of various Co/NrGO-based cell configurations clearly show multiple functions of Co/NrGO,indicating that the absence of Co/NrGO coatings and/or Co nanoparticles may be inadequate to achieve superior S availability of assembled Li-S batteries.

Establishing a protein expression profile database for the normal human pituitary gland using two-dimensional high-performance liquid chromatography combined with LTQ-Orbitrap mass spectrometry

In this study, we selected adult normal pituitary gland tissues from six patients during operations for pituitary microadenomas via the transsphenoidal approach for extended normal pituitary tissue resection around the tumor, and analyzed the protein expression of human normal pituitary using two-dimensional high-performance liquid chromatography combined with LTQ-Orbitrap mass spectrometry proteomics technology. The ten most highly expressed proteins in normal human pituitary were： alpha 3 type VI collagen isoform 5 precursor （abundance among tall pituitary proteins 1.30%）, fibrinogen beta chain preproprotein （0.99%）, vimentin （0.73%）, prolactin （0.69%）, ATP synthase, H~ transporting and mitochondrial F1 complex beta subunit precursor （0.52%）, keratin I （0.49%）, growth hormone （0.45%）, carbonic anhydrase I （0.40%）, heat shock protein 90 kDa I （0.31%）, and annexin V （0.30%）. Based on the biological function classifications of these proteins, the top three categories by content were neuroendocrine proteins （abundance among all pituitary proteins, 40.1%）, catalytic and metabolic proteins （28.3%）, and cell signal transduction proteins （9.8%）. Based on cell positioning classification, the top three categories were cell organelle （24.5%） membrane （20.8%）, and cytoplasm （13.0%）. Based on biological process classification, the top three categories of proteins are involved in physiological processes （42.9%）, cellular processes （40.4%）, and regulation of biological processes （9.1%）. Our experimental findings indicate that a protein expression profile database of normal human pituitary can be precisely and efficiently established by proteomics technology.

GraphMLP-Mixer:基于图-多层感知机架构的高效多行为序列推荐方法

在多行为序列推荐领域,图神经网络(GNNs)虽被广泛应用,但存在局限性,如对序列间协同信号建模不足和处理长距离依赖性等问题.针对这些问题,提出了一种新的解决框架GraphMLP-Mixer.该框架首先构造全局物品图来增强模型对序列间协同信号的建模,然后将感知机-混合器架构与图神经网络结合,得到图-感知机混合器模型对用户兴趣进行充分挖掘.GraphMLP-Mixer具有2个显著优势:一是能够有效捕捉用户行为的全局依赖性,同时减轻信息过压缩问题;二是其时间与空间效率显著提高,其复杂度与用户交互行为的数量成线性关系,优于现有基于GNN多行为序列推荐模型.在3个真实的公开数据集上进行实验,大量的实验结果验证了GraphMLP-Mixer在处理多行为序列推荐问题时的有效性和高效性.