10.5281/zenodo.1169388
https://zenodo.org/records/1169388
oai:zenodo.org:1169388
Grossman, Samuel
Samuel
Grossman
Stanford University
Litz, Heiner
Heiner
Litz
University of California, Santa Cruz
Kozyrakis, Christos
Christos
Kozyrakis
Stanford University
Grazelle-PPoPP18
Zenodo
2018
Graph processing
Graph analytics
Inner loop parallelization
Scheduler awareness
Scheduler-aware
Vectorization
AVX
2018-02-24
eng
10.1145/3178487.3178506
10.5281/zenodo.1169387
1
BSD 3-Clause "New" or "Revised" License
Grazelle is a high-performance hybrid (push-based and pull-based) graph processing framework targeting a single machine containing or more x86-64-based processors. It is the embodiment of the two optimization strategies described in the PPoPP 2018 paper Making Pull-Based Graph Processing Performant.
The content of this repository is intended to support the results presented in the aforementioned paper. It contains all the source code and documentation required to build and run Grazelle as configured for each experiment presented in the paper as well as pre-converted ready-to-use versions of the datasets used in the evaluation.