Search

Fat‐tree topologies are widely used in interconnect network designs for parallel supercomputers. In the classic fat‐tree, compute nodes are connected to leaf stage switches by links. Given a large number of compute nodes, many switches and links are required, resulting in high hardware costs. To solve this problem, this paper proposes two hybrid topologies, k$$ k $$‐Cube k$$ k $$‐Ary n$$ n $$‐Tree (CAT) and Mirrored k$$ k $$‐Cube k$$ k $$‐Ary n$$ n $$‐Tree (MiCAT), based on fat‐tree and hypercube. Instead of connecting k$$ k $$ compute nodes directly to a leaf switch, we connect a k$$ k $$‐cube to the switch, and each switch in the k$$ k $$‐cube part connects k$$ k $$ compute nodes. That is, this k$$ k $$‐cube consists of 2k−1$$ {2}^k-1 $$ switches and k(2k−1)$$ k\left({2}^k-1\right) $$ compute nodes. We give the shortest path routing algorithms and evaluate the path diversity, cost, performance, and average packet latency of CAT and MiCAT. The results show that CAT and MiCAT can save up to 87%$$ 87\% $$ switches and 80%$$ 80\% $$ links in a large‐scale parallel system, k=n=8$$ k=n=8 $$ for example, compared to fat‐trees, and meanwhile, both CAT and MiCAT have higher path diversities than fat‐trees. [ABSTRACT FROM AUTHOR]

Nanolayered metallic composites have attracted intensive scientific interests due to their ultrahigh strength. However, the deformation incompatibility among the component layers with high mechanical contrast leads to extremely low tensile ductility in the nanolayered composites, which is a great setback for their engineering applications. Here, by molecular dynamics simulations, we show that a heterogeneous nanolayered design by combining 2.5 nm and 24 nm Cu/Ni bilayers in a composite in an appropriate way can promote the dislocation activity of the hard phase, i.e., the Ni layers. In the new heterogeneous structure, each 24 nm Cu or Ni layer is coated on both surfaces by one 2.5 nm Cu/Ni bilayer. The simulations show that the dislocations in the 24 nm Ni layers can nucleate and glide almost synchronously with those in the 24 nm Cu layers. The enhanced dislocation activities are attributed to the presence of the 2.5 nm Cu layer that can promote the dislocation nucleation and motion in the 24 nm Ni layer by forming more nodes in the dislocation network of the interface. [ABSTRACT FROM AUTHOR]