Column-Oriented Datalog on the GPU

2025-01-22

AI Summary

Existing Datalog engines (e.g., Soufflé, VLog, Nemo) fail to fully exploit the massive parallelism and high-bandwidth memory of modern GPUs (e.g., NVIDIA H100). Method: We propose VFLog, the first GPU-optimized columnar Datalog engine. Its core innovations include: (i) deep architectural co-design of columnar storage for GPU memory hierarchies; (ii) GPU-native columnar data structures and a lock-free parallel execution model; and (iii) CUDA-accelerated implementations of all relational algebra operators (join, union, semi-join) with zero-copy host-device data transfer. Contribution/Results: Experiments show VFLog achieves over 200× speedup over state-of-the-art CPU-based columnar engines and delivers 2.5× average speedup over existing GPU-accelerated Datalog systems. It significantly improves throughput and reduces latency in knowledge graph reasoning and program analysis workloads.

Technology Category

• Knowledge Representation and Reasoning: Description Logics
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
• Machine Learning: Hardware-aware ML

Application Category

• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

Datalog is a logic programming language widely used in knowledge representation and reasoning (KRR), program analysis, and social media mining due to its expressiveness and high performance. Traditionally, Datalog engines use either row-oriented or column-oriented storage. Engines like VLog and Nemo favor column-oriented storage for efficiency on limited-resource machines, while row-oriented engines like Souffle use advanced data structures with locking to perform better on multi-core CPUs. The advent of modern datacenter GPUs, such as the NVIDIA H100 with its ability to run over 16k threads simultaneously and high memory bandwidth, has reopened the debate on which storage layout is more effective. This paper presents the first column-oriented Datalog engines tailored to the strengths of modern GPUs. We present VFLog, a CUDA-based Datalog runtime library with a column-oriented GPU datastructure that supports all necessary relational algebra operations. Our results demonstrate over 200x performance gains over SOTA CPU-based column-oriented Datalog engines and a 2.5x speedup over GPU Datalog engines in various workloads, including KRR.

Problem

• GPU-optimized Datalog processing
• Columnar storage
• High-performance computing

Innovation

• CUDA Optimization
• Columnar Datalog Processing
• NVIDIA H100 GPU