No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

Increasing number of cores in chip multiprocessors (CMP) result in increasing traffic to last-level cache (LLC). Without commensurate increase in LLC bandwidth, such traffic cannot be sustained resulting in loss of performance. Further, as the number of cores increases, it is necessary to scale up the LLC size; otherwise, the LLC miss rate will rise, resulting in a loss of performance. Unfortunately, for a unified LLC with uniform cache access time, access latency increases with cache size, resulting in performance loss. Previously, researchers have proposed partitioning the cache into multiple smaller caches interconnected by a communication network which increases aggregate cache bandwidth but causes non-uniform access latency. Such a cache architecture is called non-uniform cache architecture (NUCA). While NUCA addresses the LLC bandwidth issue, partitioning by itself does not address the access latency problem. Consequently, researchers have previously considered data placement techniques to improve access latency. However, earlier data placement work did not account for the frequency with which specific memory references are accessed. A major reason for that is access frequency for all memory references is difficult to track. In this research, we present a hardware-assisted solution called ACTION (Adaptive Cache Block Migration) to track the access frequency of individual memory references and prioritize their placement closer to the affine core. ACTION mechanism implements cache block migration when there is a detectable change in access frequencies due to a change in the program phase. To keep the hardware overhead low, ACTION counts access references in the LLC stream using a simple and approximate method, and uses simple algorithms for placement and migration. We tested ACTION on a 4-core CMP with a 5x5 mesh LLC network implementing a partitioned D-NUCA against workloads exhibiting distinct asymmetry in cache block access frequency. Our simulation results indicate that ACTION can improve CMP performance by as much as 8% over the state-of-the-art (SOTA) solutions.

The ping-pong lemma was introduced by Klein in the late 1800s to show that certain subgroups of isometries of hyperbolic 3-space are free and remains one of very few tools that certify when a pair of group elements generate a free subgroup or semigroup. Quantitatively applying the ping-pong lemma to more general group actions on metric spaces requires a blend of understanding the large-scale global geometry of the underlying space with local combinatorial and dynamical behavior of the action. In the 1980s, Gromov publish a sequence of seminal works introducing several metric notions of non-positive curvature in group theory where he asked which finitely generated groups have uniform exponential growth. We give an overview of various developments of non-positive curvature in group theory and past results related to building free semigroups in the setting of non-positive curvature. We highlight joint work with Radhika Gupta and Kasia Jankiewicz and with Carolyn Abbott and Davide Spriano that extends these tools and techniques to show several groups with that act on cube complexes and many hierarchically hyperbolic groups have uniform exponential growth. / Mathematics

Non-uniform corrosion of copper and brass in potable water systems poses both economic and environmental problems associated with premature plumbing failures and release of metals. With respect to copper pitting corrosion, it was found that forensic testing (i.e., in pipe-loops) is the only investigative technique that can closely mimic conditions found in real water systems and produce unambiguous results; and, if used in combination with electrochemical techniques, it may also provide some mechanistic insights into the pitting process. Using pipe-loops, it was demonstrated that copper pitting in aggressive water qualities (i.e., chlorinated, high pH and low alkalinity) is deterministic and reproducible. Additionally, the effects of various chemical and physical factors on pitting were investigated. Overall, increased flow velocity and frequency, increased chlorine residual and decreased hardness were found to accelerate pitting; whereas increased phosphate and silica were found to decelerate pitting. Several mitigation strategies for copper pitting in aggressive water were further investigated, and experimental data were interpreted utilizing electrochemical theory to evaluate specific effects on the initiation and propagation phases of pitting. Surprisingly, it was found that decreased chlorine may delay pit initiation, however, even relatively low levels of chlorine may eventually initiate and propagate pits. Increased alkalinity appears to decelerate pit growth, but does not prevent pit initiation. NOM can delay pit initiation and propagation, although the potential for DBP formation in chlorinated waters makes inhibition by NOM an unfavorable alternative. At sufficient dosages, phosphate and silica corrosion inhibitors may completely stop pitting, consistent with the success of several field trials. At very low dosages, phosphate and silica may actually accelerate pinhole failures, so these inhibitors should not be under-dosed. While brass alloys exist that can limit dezincification problems, they are not always utilized in potable water applications due to high costs, and so dezincification is a re-emerging issue in some countries, including the US. Little research has been conducted in the past several decades regarding the effects of water chemistry, and almost no work has addressed the roles of physical factors associated with real plumbing systems. Thus, a comprehensive review of these topics was conducted. To better understand the effects of some factors associated with specific plumbing installations on dezincification and other brass corrosion types, a series of pipe-loop studies was carried out. It was confirmed that increased oxidant delivery rates to cathodic surfaces, either via increased oxidant concentration or increased flow velocity, can increase corrosion rates. Several key differences were observed with respect to corrosion of brass located in copper plumbing tube systems as opposed to plastic. When copper tubes contribute copper ions to water, brass corrosion becomes more selective for zinc; but if galvanic connections are made between the copper tubes and brass, selectivity for zinc is reduced while overall corrosion rates are accelerated. As opposed to copper tubing, plastic maintains oxidant (e.g., free chlorine) levels, and may thereby increase brass corrosion and build-up of corrosion by-products. Finally, it was found that increased temperature can significantly increase lead leaching from brass. Following recent outbreaks of brass dezincification failures, NSF/ANSI Standard 14 has been revised to require that all NSF 14-listed brass is dezincification resistant, as certified by satisfactory results from an accelerated test method (ISO 6509). Various brasses were tested using this method as well as a longer-term jar method utilizing real potable water. Results of the two tests were in good agreement with respect to dezincification, specifically; but some inconsistencies were observed with respect to uniform corrosion and lead leaching. / Ph. D.

Recent investigations of a strut-braced wing (SBW) aircraft show that, at high positive load factors, a large tensile force in the strut leads to a considerable compressive axial force in the inner wing, resulting in a reduced bending stiffness and even buckling of the wing. Studying the influence of this compressive force on the structural response of SBW is thus of paramount importance in the early stage of SBW design. The purpose of the this research is to investigate the effect of compressive force on aeroelastic stability of the SBW using efficient structural finite element and aerodynamic lifting surface methods. A procedure is developed to generate wing stiffness distribution for detailed and simplified wing models and to include the compressive force effect in the SBW aeroelastic analysis. A sensitivity study is performed to generate response surface equations for the wing flutter speed as functions of several design variables. These aeroelastic procedures and response surface equations provide a valuable tool and trend data to study the unconventional nature of SBW. In order to estimate the effect of the compressive force, the inner part of the wing structure is modeled as a beam-column. A structural finite element method is developed based on an analytical stiffness matrix formulation of a non-uniform beam element with arbitrary polynomial variations in the cross section. By using this formulation, the number of elements to model the wing structure can be reduced without degrading the accuracy. The unsteady aerodynamic prediction is based on a discrete element lifting surface method. The present formulation improves the accuracy of existing lifting surface methods by implementing a more rigorous treatment on the aerodynamic kernel integration. The singularity of the kernel function is isolated by implementing an exact expansion series to solve an incomplete cylindrical function problem. A hybrid doublet lattice/doublet point scheme is devised to reduce the computational time. SBW aircraft selected for the present study is the fuselage-mounted engine configuration. The results indicate that the detrimental effect of the compressive force to the wing buckling and flutter speed is significant if the wing-strut junction is placed near the wing tip. / Ph. D.