Automatic BLAS Offloading on Unified Memory Architecture: A Study on NVIDIA Grace-Hopper
Open Access
- 17 July 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Abstract
Porting codes to GPU often requires major efforts. While several tools exist for automatically offload numerical libraries such as BLAS and LAPACK, they often prove impractical due to the high cost of mandatory data transfer. The new unified memory architecture in NVIDIA Grace-Hopper allows high bandwidth cache-coherent memory access of all memory from both CPU and GPU, potentially eliminating bottleneck faced in conventional architecture. This breakthrough opens up new avenues for application development and porting strategies. In this study, we introduce a new tool for automatic BLAS offload, the tool leverages the high speed cache coherent NVLink C2C interconnect in Grace-Hopper, and enables performant GPU offload for BLAS heavy applications with no code changes or recompilation. The tool was tested on two quantum chemistry or physics codes, great performance benefits were observed.Keywords
All Related Versions
Funding Information
- NSF (OAC-2402542, OAC-1854828, OAC-2139536)
This publication has 5 references indexed in Scilit:
- PEAK: a Light-Weight Profiler for HPC Systems2023
- GPU acceleration of the Locally Selfconsistent Multiple Scattering code for first principles calculation of the ground state and statistical physics of materialsComputer Physics Communications, 2017
- PARSEC – the pseudopotential algorithm for real‐space electronic structure calculations: recent advances and novel applications to nano‐structuresPhysica Status Solidi (b), 2006
- Order-NMultiple Scattering Approach to Electronic Structure CalculationsPhysical Review Letters, 1995
- Finite-difference-pseudopotential method: Electronic structure calculations without a basisPhysical Review Letters, 1994