Profiling

Proton: Towards Multi-level, Adaptive Profiling for Triton
Cite Project URL
DeepContext: A Context-aware, Cross-platform, and Cross-framework Tool for Performance Profiling and Analysis of Deep Learning Workloads
Performance profiling toolkit that unifies deep learning workload analysis across platforms and frameworks.
Cite DOI arXiv
HPCToolkit
Our tool provides a profile view and a trace view for GPU-accelerated applications. The profile view identifies where GPU APIs are invoked in CPU calling context, approximates calling context for GPU execution, and analyzes instruction mix for GPU kernels. The tool traces CPU and GPU activities for a large number of processes and threads with minimal overhead.
Code DOC
HPCToolkit
Low Overhead and Context Sensitive Profiling of GPU-Accelerated Applications
As we near the end of Moore’s law scaling, the next-generation computing platforms are increasingly exploring heterogeneous …
Cite Project DOI URL
ValueExpert: Exploring Value Patterns in GPU-Accelerated Applications
General-purpose GPUs have become common in modern computing systems to accelerate applications in many domains, including machine …
Cite DOI URL
GPA: A GPU Performance Advisor Based on Instruction Sampling
Presented our work on GPU performance advisor at CGO21
Mar 1, 2021 9:34 PM — 9:34 PM Virtual
Slides
Tools for Top-down Performance Analysis of GPU-Accelerated Applications
Presented the beta version of our comprehensive GPU profiling tool at ICS20
Jul 1, 2020 9:41 PM — 9:41 PM Virtual
Project Slides
A Tool for Top-down Performance Analysis of GPU-accelerated Applications
Presented a poster and a short talk about HPCToolkit’s GPU support
Apr 24, 2020 9:50 PM — 9:50 PM San Diego, CA, USA
Project Poster Slides
A Tool for Top-down Performance Analysis of GPU-Accelerated Applications
To support performance measurement and analysis of GPU-accelerated applications, we extended the HPCToolkit performance tools with …
Cite Project DOI URL
Tools for Top-down Performance Analysis of GPU-Accelerated Applications
This paper describes extensions to Rice University’s HPCToolkit performance tools to support measurement and analysis of …
Cite Project DOI URL