DrGPU: A Top-Down Profiler for GPU Applications
Open Access
- 15 April 2023
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 43-53
Abstract
GPUs have become common in HPC systems to accelerate scientific computing and machine learning applications. Efficiently mapping these applications to rapid evolutions of GPU architectures for high performance is a well-known challenge. Various performance inefficiencies exist in GPU kernels that impede applications from obtaining bare-metal performance. While existing tools are able to measure these inefficiencies, they mostly focus on data collection and presentation, requiring significant manual efforts to understand the root causes for actionable optimization. Thus, we develop DrGPU, a novel profiler that performs top-down analysis to guide GPU code optimization. As its salient feature, DrGPU leverages hardware performance counters available in commodity GPUs to quantify stall cycles, decompose them into various stall reasons, pinpoint root causes, and provide intuitive optimization guidance. With the help of DrGPU, we are able to analyze important GPU benchmarks and applications and obtain nontrivial speedups --- up to 1.77X on V100 and 2.03X on GTX 1650.Keywords
Funding Information
- NSF (National Science Foundation) (2125813, 2050007)
This publication has 12 references indexed in Scilit:
- Tools for top-down performance analysis of GPU-accelerated applications2020
- A tool for top-down performance analysis of GPU-accelerated applications2020
- NVBit2019
- CUDAAdvisor: LLVM-based runtime profiling for modern GPUs2018
- gpucc: an open-source GPGPU compiler2016
- OpenACC — First Experiences with Real-World Applications2012
- Ocelot2010
- CASTRO: A NEW COMPRESSIBLE ASTROPHYSICAL SOLVER. I. HYDRODYNAMICS AND SELF-GRAVITYThe Astrophysical Journal, 2010
- The Tau Parallel Performance SystemThe International Journal of High Performance Computing Applications, 2006
- Simultaneous multithreading1995