GPU-Resident Gaussian Process Regression Leveraging Asynchronous Tasks with HPX

2026-02-23

AI Summary

Gaussian Process Regression (GPR) struggles to scale to large datasets due to the cubic time complexity of its exact inference. This work proposes the first fully GPU-resident GPR prediction pipeline, integrating block-wise Cholesky decomposition with efficient GPU memory management and leveraging the HPX asynchronous task runtime alongside multiple CUDA streams for coordinated scheduling. The approach achieves substantial computational speedups while preserving numerical stability: on datasets with more than 128 samples, it accelerates Cholesky decomposition by 4.3× and overall prediction by 4.6×. In large-model scenarios, the method outperforms cuSOLVER by up to 11% in performance.

Technology Category

• Machine Learning: Hardware-aware ML
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
• Reasoning under Uncertainty: Probabilistic Programming

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

Gaussian processes (GPs) are a widely used regression tool, but the cubic complexity of exact solvers limits their scalability. To address this challenge, we extend the GPRat library by incorporating a fully GPU-resident GP prediction pipeline. GPRat is an HPX-based library that combines task-based parallelism with an intuitive Python API. We implement tiled algorithms for the GP prediction using optimized CUDA libraries, thereby exploiting massive parallelism for linear algebra operations. We evaluate the optimal number of CUDA streams and compare the performance of our GPU implementation to the existing CPU-based implementation. Our results show the GPU implementation provides speedups for datasets larger than 128 training samples. We observe speedups of up to 4.3 for the Cholesky decomposition itself and 4.6 for the GP prediction. Furthermore, combining HPX with multiple CUDA streams allows GPRat to match, and for large datasets, surpass cuSOLVER's performance by up to 11 percent.

Problem

• Gaussian Process Regression
• cubic complexity
• scalability
• GPU acceleration
• large-scale datasets

Innovation

• GPU-resident
• Gaussian Process Regression
• HPX
• asynchronous tasks
• tiled algorithms