通用图形处理器缓存子系统性能优化方法综述

随着工艺和制程技术的不断发展以及体系架构的日趋完善,通用图形处理器(general purpose graphics processing units, GPGPU)的并行计算能力得到了很大的提升,其在高性能、高吞吐量等通用计算应用场景的使用越来越广泛.GPGPU通过支持大量线程的并发执行,可以较好地隐藏长延时访存操作,从而获得高并行计算能力.然而,GPGPU在处理计算和访存不规则的应用时,其存储子系统的效率受到很大影响,尤其是片上缓存的争用情况尤为突出,难以及时提供计算操作所需的数据,使得GPGPU的高并行计算能力不能得到充分发挥.解决片上缓存的争用问题、优化缓存子系统的性能,是优化GPGPU性能的主要解决方案之一,也是目前研究GPGPU性能优化的主要热点之一.目前,针对GPGPU缓存子系统的性能优化研究主要集中在线程级并行度(thread level parallelism, TLP)调节、访存顺序调节、数据通量增强、最后一级缓存(last level cache, LLC)优化和基于非易失性存储(non-volatile memory, NVM)的GPGPU缓存新架构设计等5个方面.也从这5个方面重点分析讨论了目前主要的GPGPU缓存子系统性能优化方法,并在最后指出了未来GPGPU缓存子系统优化需要进一步探讨的问题,对GPGPU缓存子系统性能优化的研究有重要意义.     

Shrunken Pore Syndrome Is Frequently Occurring in Severe COVID-19

A selective decrease in the renal filtration of larger molecules is attributed to the shrinkage of glomerular pores, a condition termed Shrunken Pore Syndrome (SPS). SPS is associated with poor long-term prognosis. We studied SPS as a risk marker in a cohort of patients with COVID-19 treated in an intensive care unit. SPS was defined as a ratio < 0.7 when the estimated glomerular filtration rate (eGFR), determined by cystatin C, calculated by the Cystatin C Caucasian-Asian-Pediatric-Adult equation (CAPA), was divided by the eGFR determined by creatinine, calculated by the revised Lund–Malmö creatinine equation (LMR). Clinical data were prospectively collected. In total, SPS was present in 86 (24%) of 352 patients with COVID-19 on ICU admission. Patients with SPS had a higher BMI, Simplified Physiology Score (SAPS3), and had diabetes and/or hypertension more frequently than patients without SPS. Ninety-nine patients in the total cohort were women, 50 of whom had SPS. In dexamethasone-naïve patients, C-reactive protein (CRP ), TNF-alpha, and interleukin-6 did not differ between SPS and non-SPS patients. Demographic factors (gender, BMI) and illness severity (SAPS3) were independent predictors of SPS. Age and dexamethasone treatment did not affect the frequency of SPS after adjustments for age, sex, BMI, and acute severity. SPS is frequent in severely ill COVID-19 patients. Female gender was associated with a higher proportion of SPS. Demographic factors and illness severity were independent predictors of SPS.     

Specific S100 Proteins Bind Tumor Necrosis Factor and Inhibit Its Activity

Tumor necrosis factor (TNF) inhibitors (anti-TNFs) represent a cornerstone of the treatment of various immune-mediated inflammatory diseases and are among the most commercially successful therapeutic agents. Knowledge of TNF binding partners is critical for identification of the factors able to affect clinical efficacy of the anti-TNFs. Here, we report that among eighteen representatives of the multifunctional S100 protein family, only S100A11, S100A12 and S100A13 interact with the soluble form of TNF (sTNF) in vitro. The lowest equilibrium dissociation constants (K_d) for the complexes with monomeric sTNF determined using surface plasmon resonance spectroscopy range from 2 nM to 28 nM. The apparent K_d values for the complexes of multimeric sTNF with S100A11/A12 estimated from fluorimetric titrations are 0.1–0.3 µM. S100A12/A13 suppress the cytotoxic activity of sTNF against Huh-7 cells, as evidenced by the MTT assay. Structural modeling indicates that the sTNF-S100 interactions may interfere with the sTNF recognition by the therapeutic anti-TNFs. Bioinformatics analysis reveals dysregulation of TNF and S100A11/A12/A13 in numerous disorders. Overall, we have shown a novel potential regulatory role of the extracellular forms of specific S100 proteins that may affect the efficacy of anti-TNF treatment in various diseases.     

Interfacial interactions and properties of filled polymers. II: Dispersion of filler particles

The specific interaction characteristics and the inherent agglomeration of variously surface coated rutile pigments have been assessed, respectively, by inverse gas chromatographic and powder rheological methods. Standardized methods were used to disperse the pigments in polyethylene and chlorinated polyethylene. Measurements were made of energy requirements for dispersion and of the quality of dispersion attained. It was found that in the non-polar polyethylene matrix, dispersion processes depended on the strength of pigment agglomerates, but not on the specific interaction potential of the solids. Conversely, in the acidic chlorinated polyethylene, acid/base interactions influenced dispersion but the process was independent of inherent pigment agglomeration.     

Time-dependent response of aramid-epoxy-aluminum sheet,ARALL, laminates

Selected results of an ongoing investigation aimed at characterizing the timedependent response of an aramid-epoxy-aluminum sheet laminate and its constitutents at 121°C are outlined in this paper. This laminate is a recently developed hybrid composite developed by the Aluminum Company of America, marketed under the ARALL-4 tradename. The paper addresses the time-dependent response of the above hybrid composite under creep loading. It is illustrated that ARALL-4 laminates may exhibit substantial creep effects at stress levels below the proportional limit. The creep response is a nonlinear function of time and the applied stress level and is primarily due to the creep characteristics of the aluminum layers. An analytical model based on the assumptions of the classical lamination theory developed to model the time-dependent response of these laminates under creep and thermal loading is shown to yield good correlation with the experimental data. It is also illustrated that the residual state of stress can influence the extent of creep. This offers the possibility of minimizing the creep effects by altering the state of residual stress with mechanical prestraining.     

The modelling of networks with frequently changing topology whilst maintaining a constant system matrix

This paper describes a novel technique with which a system with changing topology can be modelled whilst maintaining a constant system matrix. This technique employs a new transmission-line switch model which has a constant characteristic impedance, irrespective of its state. The method is explained, compared with the switched-resistance method and demonstrated by two examples. It has been found that the proposed method offers substantial advantages in the formulation of the problems and in the efficiency of computation.     

Preparation,Characterisation, and Rheological Behaviour of Starch-Sodium Trimetaphosphate Hydrogels

Ionic biopolymer hydrogels were prepared by the cross-linking of starches with sodium trimetaphosphate in alkaline medium at 40°C for 2 hours. The swelling capacity is relatively high — up to 310 g H₂0/g polymer. Salt solutions have a marked influence, and result in shrinkage but not in a total collapse. The effect of both the cross-linker and substrate concentrations on the swelling and rheological properties was investigated. The influence of temperature and NaOH concentration on the rheological behaviour suggests that they are both significant in determining the gel properties because of the readiness of the diester phosphate bonds to undergo hydrolysis. The molecular weight between two entanglement points (M_e) and the effectiveness of cross-linking [n_e(r)/n_e(t)] were estimated from the observed G_p′ values, and those calculated from complete conversion of the cross-linker. The effectiveness of cross-linking lay between 0.2 and 2.74% for 10% wlw gels, and reached a maximum of 48.1% at the higher substrate concentration of 20% wlw. ¹³C-NMR signals from the anhydroglucose units became broader and decreased in intensity with rising cross-linker concentration due to the restricted motion arising from the additional bonding.