owensgroup / application_classification

CUDA implementation of application classification via belief propagation

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

application_classification

Implementation of ApplicationClassification workflow.

Produces identical results to PNNL code on georgiy and rmat example datasets.

TODO

  • Optimize performance
  • Fuse kernels
  • Get rid of transposes
  • Segmented reduce instead of reduce by key?
  • Could probably improve memory usage
  • Profiling
  • optional more correctness checking

About

CUDA implementation of application classification via belief propagation


Languages

Language:Cuda 72.0%Language:Shell 14.2%Language:Python 8.8%Language:C++ 4.3%Language:Makefile 0.8%