README.md

IA3 2022

• Please open IA3_2022 folder that contains code and instructions for IA3 2022 paper.
• BibTeX to cite this work:

@INPROCEEDINGS{10027548,
  author={Shovon, Ahmedur Rahman and Dyken, Landon Richard and Green, Oded and Gilray, Thomas and Kumar, Sidharth},
  booktitle={2022 IEEE/ACM Workshop on Irregular Applications: Architectures and Algorithms (IA3)}, 
  title={Accelerating Datalog applications with cuDF}, 
  year={2022},
  volume={},
  number={},
  pages={41-45},
  doi={10.1109/IA356718.2022.00012}}

Configuration

• Configuration data are collected using nvidia-smi, lscpu, free -h
• Theta GPU (single-gpu node):
  • GPU 0: NVIDIA A100-SXM4-40GB
  • CPU Model name: AMD EPYC 7742 64-Core Processor
  • CPU(s): 256
  • Thread(s) per core: 2
  • Core(s) per socket: 64
  • Socket(s): 2
  • CPU MHz: 3399.106
  • L1d cache: 4 MiB
  • L1i cache: 4 MiB
  • L2 cache: 64 MiB
  • L3 cache: 512 MiB
  • Total memory: 1.0Ti

References

• Short CUDA tutorial
• nVidia CUDA C programming guide
• Theta GPU nodes
• Getting started video
• Getting Started on ThetaGPU
• Submit a Job on ThetaGPU
• Running jobs at Theta
• Chapter 39. Parallel Prefix Sum (Scan) with CUDA
• GPGPU introduction
• CUDA 2d thread block
• CUB
• Documentation on CUDF Drop
• Documentation on CUDF Drop Duplicates
• Documentation on CUDF concatenate
• Open addressing hash table
• Open addressing techniques
• A Simple GPU Hash Table
• CUDA Pro Tip: Occupancy API Simplifies Launch Configuration
• CUDA thrust
• Count in thrust
• Basics of Hash Tables
• MurmurHash3
• Thrust stable_sort()