The increasing demand for high-performance query processing on large in-memory datasets has driven the adoption of GPU-based B+-trees for handling high-concurrency query (HCQ) workloads. Existing approaches, by randomly assigning queries to GPU threads, suffer from inefficiencies related to memory access patterns, cache utilization, and thread divergence. This paper introduces a novel query grouping strategy that assigns queries with similar search keys to the same CUDA block, thereby improving query throughput. We formalize the optimal key assignment (OKA) problem as a variation of the K-means problem, establishing its theoretical foundations and proposing an efficient algorithm with proven optimality. We implement this algorithm using highly optimized CUDA code and extend our approach to support range queries, a common but understudied workload in GPU-based HCQ systems. Experimental evaluations demonstrate that our query grouping strategy significantly outperforms prior work, achieving up to 10.6X lower latency and 32.2X higher throughput, while also improving GPU resource utilization (e.g., cache hit rate and memory throughput).

Counting triangles in large graphs, being a crucial problem in graph computing, has attracted significant attention from research communities. There is a large body of work dedicated to algorithmic design and efficient implementation on parallel platforms such as GPUs. Among them, the intersection-based triangle counting algorithm is found to be the most efficient approach and a few GPU implementations have been proposed following this algorithm. However, there remains a gap in understanding how these algorithms perform when confronted with diverse real-world graph datasets. It is a well-established fact that the performance of GPU code is heavily influenced by data characteristics, including graph size and node degree, often leading to issues such as workload imbalances and inefficient memory access. The goal of this study is to systematically evaluate the performance and analyze the behavior of eight recently published intersection-based triangle counting implementations. For that, we developed a unified testing framework that facilitates fast performance assessment of any triangle counting algorithm. Our experiments show that the TRUST algorithm outperforms competitors in most cases. To our surprise, the Polak algorithm, with a simple design, has displayed commendable performance across the board. Notably, it even surpasses the TRUST algorithm when processing small datasets. We conducted an in-depth analysis on the resource consumption patterns of these implementations in relation to their performance and identified key factors that contributed to their behaviors. Based on insights gained from such analysis, we proposed a novel algorithm named GroupTC that delivers outstanding performance under all types of datasets.

Due to high computational efficiency and low storage cost, cross-modal hashing retrieval attracts much attention. However, as heterogeneous data from different modalities often have distinct physical meanings and underlying structures, learning encoding with equal length for different modalities may result in an insurmountable semantic gap. In addition, there are still some issues, e.g., how to combine label and sample information to learn hash codes effectively, how to reduce the time consumption caused by computing n × n similarity matrix, and how to effectively solve the complex discrete optimization problem. To overcome the above challenges, this study propose a novel model called Scalable Discrete and Asymmetric Unequal Length Hashing (SDAULH). First, SDAULH constructs a novel hash model that utilizes unequal length encoding schemes to narrow the semantic gap between heterogeneous modalities. Second, SDAULH develops a dual semantic embedding learning scheme, which combines pairwise similarity between label and sample data to generate a more discriminative hash code. Third, SDAULH associates with both hash codes and label information by an asymmetric relaxation strategy. Furthermore, SDAULH solves directly the discrete optimization problem by generating discrete hash codes. Experimental results on four benchmark datasets demonstrate the promising performance of SDAULH.

In this paper, we present a novel Movable Gripper (MovGrip) which targets on climbing non-enclosable rectangular trusses such as bridges and space stations. It is designed with a transformation mechanism which provides the features of parallel grippers, rotatory grippers, and active wheels. For truss climbing, MovGrip acts as a parallel gripper which allows the change in gripping width according to different size of trusses. Since MovGrip is equipped with active wheels, fast climbing motion can therefore be realized. Moreover, MovGrip can be transformed into a mobile platform which is suitable for the navigation on ground. It weighs only 500 grams with a climbing speed of 3 cm/s. To maintain the climbing stability, we steer the rotation axis of wheels. By this, a directional pulling force (from MovGrip to truss surface) can be distributed from the drive force of wheels which pulls MovGrip towards the truss while climbing. Experimental results show that tilting of MovGrip can be auto-adjusted based on the proposed design. Also, it is shown that the load carrying capability of MovGrip is approximately 1 kg which is 2 times of its weight.