Analytical tasks in various domains increasingly encode complex information as dense vector data (e.g., embeddings), often requiring filtered vector search (i.e., vector search with attribute filtering). This search is challenging due to the volume and dimensionality of the data, the number and variety of filters, and the difference in distribution and/or update frequency between vectors and filters. Besides, many real applications require answers in a few milliseconds with high recall on large collections. Graph-based methods are considered the best choice for such applications, despite a lack of theoretical guarantees on query accuracy. Existing solutions for filtered vector search are either: 1) ad-hoc, using existing techniques with no or minor modifications; or 2) hybrid, providing specialized indexing and/or search algorithms. We show that neither is satisfactory and propose RWalks, an index-agnostic graph-based filtered vector search method that efficiently supports both filtered and unfiltered vector search. We demonstrate its scalability and robustness against the state-of-the-art with an exhaustive experimental evaluation on four real datasets (up to 100 million vectors), using query workloads with filters of different types (unique/composite), and varied specificity (proportion of points that satisfy a filter). The results show that RWalks can perform filtered search up to 2x faster than the second-best competitor (ACORN), while building the index 76x faster and answering unfiltered search 13x faster.

The JavaScript Object Notation (JSON) format is ubiquitous, and countless applications depend on it to store and exchange high volumes of data. Despite its great popularity, JSON is nevertheless a very inefficient data format: decoding and querying JSON data is often a major bottleneck for many data-intensive applications. In this paper, we explore how Graphics Processing Units (GPUs) can be used to parallelize both JSON de-serialization and querying. We show how JSON parsing can be implemented on GPUs by means of parallel structural index construction, and we describe how JSON data can then be queried in situ using a lightweight query engine designed to run on GPUs. We present the design and implementation of GpJSON, a GPU-based JSON data processing library. The library can be used from high-level languages such as JavaScript or Python, and features bindings for the GraalVM language runtime. Our evaluation on real-world datasets shows that, on a single NVIDIA Ampere A100, GpJSON achieves at least 2.9x speedup on end-to-end performance (de-serialization plus querying) over state-of-the-art parallel JSON parsers and query engines, and 6-8 x over NVIDIA RAPIDS.

GPUs are readily available in cloud computing and personal devices, but their use for data processing acceleration has been slowed down by their limited integration with common programming languages such as Python or Java. Moreover, using GPUs to their full capabilities requires expert knowledge of asynchronous programming. In this work, we present a novel GPU run time scheduler for multi-task GPU computations that transparently provides asynchronous execution, space-sharing, and transfer-computation overlap without requiring in advance any information about the program dependency structure. We leverage the GrCUDA polyglot API to integrate our scheduler with multiple high-level languages and provide a platform for fast prototyping and easy GPU acceleration. We validate our work on 6 benchmarks created to evaluate task-parallelism and show an average of 44% speedup against synchronous execution, with no execution time slowdown compared to hand-optimized host code written using the C++ CUDA Graphs API.