Anomaly detection in manufacturing remains a significant challenge, particularly due to the unique characteristics of industrial data and the scarcity of abnormal samples. This study investigates the effectiveness of unsupervised learning approaches, leveraging multi-model frameworks that combine pre-trained models on large-scale datasets with models specifically trained to capture industrial data features. This study proposes two key methodologies to maximize multi-model efficiency. Dynamic loss weighting optimizes the contribution of each model during training, enabling networks with diverse expertise to synergize effectively. Soft-hard feature loss, in particular, focuses on precisely capturing subtle anomaly regions that traditional methods might miss. By emphasizing features with high-error values while appropriately utilizing those with lower error values, the proposed approach enables more detailed anomaly detection compared to existing methods, allowing for the detection of even minor defects through refined anomaly region analysis. Quantitative results on the MVTec and VisA datasets demonstrate that the proposed method achieves remarkable performance improvements, with up to +0.6% (AU-PRO) on the MVTec AD dataset and up to +0.4% (AU-ROC) on the VisA dataset compared to baseline methods. In addition to its superior quantitative performance, the proposed method enables more precise anomaly detection than conventional approaches. Furthermore, the proposed method eliminates the need for experimental tuning, enabling its application to diverse multi-model approaches and ensuring adaptability across various datasets.

Streaming is emerging as an important programming model for multicores. Streaming provides an elegant way to express task decomposition and inter-task communication, while hiding laborious orchestration details such as load balancing, assignment (of stream computation to nodes) and computation/communication scheduling from the programmer. This paper develops a novel communication optimization for streaming applications based on the observation that streaming computations typically involve large, systematic data transfers between known communicating pairs of nodes over extended periods of time. From the above observation, we advocate a family of routing algorithms that expend some over overheads to compute disjoint paths for stream communication. Disjoint-path routing is an attractive design point because (a) the overheads of discovering disjoint paths are amortized over large periods of time and (b) the benefits of disjoint path routing are significant for bandwidth-sensitive streaming applications. We develop one instance of disjoint-path routing called tentacle routing-a backtracking, best-effort technique. On a 4 times 4 (6 times 6) system, tentacle routing results in 55% (84%) and 28% (41%) mean throughput improvement for high-network-contention streaming applications, and for all streaming applications, respectively.

Researchers have proposed clustered microarchitectures for performance and energy effciency. Typically, clustered microarchitectures offer fast, local bypassing between instructions within clusters but global bypasses are slower. Traditional clustered microarchitectures (TCM) are implemented by partitioning the register file and associated functional

Crossbar arbitration - which determines the allocation of output ports to packets in the input queues - is a performance-critical stage in the overall performance of routers for input-queued networks. The overall performance of crossbar arbitration depends on two metrics: (a) matching power - the ability of the arbiter to maximize the number of matches between requesting inputs and free outputs and (b) arbitration throughput - the number of such matches per unit time. Ideally, crossbar arbitration should maximize both metrics. Unfortunately, implementing high performance matching schemes compromises arbitration throughput. Similarly, simpler arbitration mechanisms that deliver high arbitration throughput offer lower matching power. The major contribution of this paper is the design of a table-lookup based crossbar arbitration mechanism - TabArb - that delivers superior matching and high arbitration throughput for minimal-routed, two dimensional mesh and torus networks. The two key innovations of TabArb are: (a) it forwards multiple requests from each input port to multiple output ports to expose adequate matching potential and (b) it employs precomputed tables that store maximum cardinality matches for all possible request combinations. Our technique improves the saturation throughput of adaptive routed mesh network by 14.8%. It offers little improvement for the DOR router due to limited opportunity.

Minimizing latency and maximizing throughput are important goals in the design of routing algorithms for interconnection networks. Ideally, we would like a routing algorithm to (a) route packets using the minimal number of hops to reduce latency and preserve communication locality, (b) deliver good worst-case and average-case throughput and (c) enable low-complexity (and hence, low latency) router implementation. In this paper, we focus on routing algorithms for an important class of interconnection networks: two dimensional (2D) mesh networks. Existing routing algorithms for mesh networks fail to satisfy one or more of design goals mentioned above. Variously, the routing algorithms suffer from poor worst case throughput (ROMM by Nesson, and Johnsson (1995), DOR by Sullivan and Bashkow, 1977), poor latency due to increased packet hops (VALIANT by Valiant and Brebner (1981)) or increased latency due to hardware complexity (minimal-adaptive by Duato, (1993), Upadhyay, Varavithya and Mohapatra, (1997)). The major contribution of this paper is the design of an oblivious routing algorithm-O1TURN-with provable near-optimal worst-case throughput, good average-case through-put, low design complexity and minimal number of network hops for 2D-mesh networks, thus satisfying all the stated design goals.