Work

RAFI -- A Ray/Work Forwarding Infrastructure for Data Parallel Multi-Node/Multi-GPU Computing

Year: 2026

Type: preprint

Abstract: We present RaFI, a CUDA and MPI based software framework that simplifies the task of building GPU-enabled data-parallel software where rays or similar work items need to migrate between different GPUs. RaFI provides a simple interface for CUDA kernels to forward such work items to other GPUs, while under the hood managing all the CUDA and MPI related work required to make this happen. We describe RaFI's motivation and implementation, and show its potential in several example applications.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2605.30294

Authors: Wald, Ingo, Demirci, Serkan, Sahistan, Alper, Zellmann, Stefan, Paris, Andrea, Moran, Patrick, Jaros, Milan, von Landesberger, Tatiana, Gudukbay, Ugur, Pascucci, Valerio

Topic: Parallel Computing and Optimization Techniques

Subfield: Hardware and Architecture

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7162761978

Cites:

Cited by:

Related to: