GPUs (PTX and SASS) ISA Latencies Microbenchmarks

Low overhead microbenchmarks for NVIDIA GPGPUs Instructions latencies (PTX and SASS) and the access cycles of the various memory hierarchies found in different NVIDIA GPU architectures

Paper

• [HPEC ’19] Low Overhead Instruction Latency Characterization for NVIDIA GPGPUs

• If you find this code useful in your research, please consider citing as:

@INPROCEEDINGS{Arafa2020HPEC,
  author={Y. {Arafa} and A. A. {Badawy} and G. {Chennupati} and N. {Santhi} and S. {Eidenbenz}},
  booktitle={2019 IEEE High Performance Extreme Computing Conference (HPEC)}, 
  title={Low Overhead Instruction Latency Characterization for NVIDIA GPGPUs}, 
  year={2019},
  pages={1-8},
  doi={10.1109/HPEC.2019.8916466}}

Classification

GPUs-ISA-Latencies is part of the original PPT (https://github.com/lanl/PPT) and is Unclassified and contains no Unclassified Controlled Nuclear Information. It abides with the following computer code from Los Alamos National Laboratory

• Code Name: Performance Prediction Toolkit, C17098
• Export Control Review Information: DOC-U.S. Department of Commerce, EAR99
• B&R Code: YN0100000

License

&copy 2017. Triad National Security, LLC. All rights reserved.

This program was produced under U.S. Government contract 89233218CNA000001 for Los Alamos National Laboratory (LANL), which is operated by Triad National Security, LLC for the U.S. Department of Energy/National Nuclear Security Administration.

All rights in the program are reserved by Triad National Security, LLC, and the U.S. Department of Energy/National Nuclear Security Administration. The Government is granted for itself and others acting on its behalf a nonexclusive, paid-up, irrevocable worldwide license in this material to reproduce, prepare derivative works, distribute copies to the public, perform publicly and display publicly, and to permit others to do so.

Recall that this copyright notice must be accompanied by the appropriate open source license terms and conditions. Additionally, it is prudent to include a statement of which license is being used with the copyright notice. For example, the text below could also be included in the copyright notice file: This is open source software; you can redistribute it and/or modify it under the terms of the Performance Prediction Toolkit (PPT) License. If software is modified to produce derivative works, such modified software should be clearly marked, so as not to confuse it with the version available from LANL. Full text of the Performance Prediction Toolkit (PPT) License can be found in the License file in the main development branch of the repository.