Optimal Sparse Matrix Dense Vector Multiplication in the I/O-Model

We study the problem of sparse-matrix dense-vector multiplication (SpMV) in external memory. The task of SpMV is to compute y:=Ax, where A is a sparse N×N matrix and x is a vector. We express sparsity by a parameter k, and for each choice of k consider the class of matrices where the number of nonzero entries is kN, i.e., where the average number of nonzero entries per column is k.     

Study of Low-Temperature Thermocompression Bonding in Ag-In Solder for Packaging Applications

Low-temperature solders have wide applications in integrated circuits and micro-electromechanical systems packaging. In this article, a study on Ag-In solder for chip-to-chip thermocompression bonding was carried out. The resulting joint consists of AgIn₂ and Ag₉In₄ phases, with the latter phase having a melting temperature higher than 400°C. Complete consumption of In solder into a Ag-rich intermetallic compound is achieved by applying a bond pressure of 1.4 MPa at 180°C for 40 min. We also observe that the bonding pressure effect enables a Ag-rich phase to be formed within a shorter bonding duration (10 min) at a higher pressure of 1.6 MPa. Finally, prolonged aging leads to the formation of the final phase of Ag₉In₄ in the bonded joints.     

Semiconductor device reliability vs process quality

This paper describes a method for defining a quantitative model relating “quality” expressed in terms of parameter distributions and “reliability” expressed in terms of failure rates. This model makes it possible to generate a more realistic failure rate estimate for semiconductor devices. With this model, failure rate predictions based on conventional life test data and on process parameter distributions, can be defined.     

Highly flame‐retardant bio‐based polyurethanes using novel reactive polyols

Poor flame retardancy of polyurethanes (PU) is a global issue as it limits their applications particularly in construction, automobile, and household appliances industries. The global challenge of high flammability of PU can be addressed by incorporating flame‐retardant materials. However, additive flame‐retardants are non‐compatible and depreciate the properties of PU. Hence, reactive flame‐retardants (RFR) based on aliphatic (Ali‐1 and Ali‐2) and aromatic (Ar‐1 and Ar‐2) structured bromine compounds were synthesized and used to prepare bio‐based PU using limonene dimercaptan. The aromatic bromine containing foams showed higher close cell content (average 97 and 100%) and compressive strength (230 and 325 kPa) to that of aliphatic bromine containing foams. Similar behavior was observed for a horizontal burning test where with a low concentration of bromine (5 wt %) in the foams for Ar‐1 and Ar‐2 displayed a burning time of 12.5 and 11.8 s while, Ali‐1 and Ali‐2 displayed burning time of 25.7 and 37 s, respectively. Neat foam showed a burning time of 74 s. The percentage weight loss for neat PU foam was 26.5%, while foams containing 5 wt % bromine in Ali‐1, Ali‐2, Ar‐1, and Ar‐2 foams displayed weight loss of 11.3, 14, 7.9, and 14%, respectively. Our results suggest that flame retardant PU foams could be prepared effectively by using RFR materials. © 2017 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2018 , 135, 46027.     

Approximate Shortest Paths Guided by a Small Index

Distance oracles and graph spanners are excerpts of a graph that allow to compute approximate shortest paths. Here, we consider the situation where it is possible to access the original graph in addition to the graph excerpt while computing paths. This allows for asymptotically much smaller excerpts than distance oracles or spanners. The quality of an algorithm in this setting is measured by the size of the excerpt (in bits), by how much of the original graph is accessed (in number of edges), and the stretch of the computed path (as the ratio between the length of the path and the distance between its end points). Because these three objectives are conflicting goals, we are interested in a good trade-off. We measure the number of accesses to the graph relative to the number of edges in the computed path.     

A Note on the Parallel Runtime of Self-Stabilizing Graph Linearization

Topological self-stabilization is an important concept to build robust open distributed systems (such as peer-to-peer systems) where nodes can organize themselves into meaningful network topologies. The goal is to devise distributed algorithms where nodes forward, insert, and delete links to neighboring nodes, and that converge quickly to such a desirable topology, independently of the initial network configuration. This article proposes a new model to study the parallel convergence time. Our model sheds light on the achievable parallelism by avoiding bottlenecks of existing models that can yield a distorted picture. As a case study, we consider local graph linearization—i.e., how to build a sorted list of the nodes of a connected graph in a distributed and self-stabilizing manner. In order to study the main structure and properties of our model, we propose two variants of a most simple local linearization algorithm. For each of these variants, we present analyses of the worst-case and best-case parallel time complexities, as well as the performance under a greedy selection of the actions to be executed. It turns out that the analysis is non-trivial despite the simple setting, and to complement our formal insights we report on our experiments which indicate that the runtimes may be better in the average case.