Skip to content

sfu-dis/pibench-online

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

97 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PiBench Online

Build Status

What is PiBench Online?

PiBench Online is an interactive system for benchmarking PM indexes in a fair and reproducible manner, it is based on PiBench, the first unified, highly customizable benchmarking framework for PM indexes.

PiBench Online enables push-button evaluation of PM indexes: the user can upload a shared library that implements the index, set parameters to run customized benchmarks, and analyze results interactively, all through an easy-to-use web interface. We hope it can promote fair comparison and reproducibility in future PM index research

Check out a demo video here and try it out at http://pibench.org. Use the code in this repo to deploy your own.

Deploy your own instance

Pibench Online allows you to deploy a server instance on your own machine and connect it with the pibench online web interface (http://pibench.org).

You can either clone the code and build from source (see below), or simply use our official docker image.

Docker

# 1. Pull the pibench-online image from docker hub
# 2. connect the host's port 8000 with container's port 8000
#    so users can access the web api from host's 8000 port
docker run -p 127.0.0.1:8000:8000 pibench-online:latest

Bare metal

cargo run --release

About Us

PiBench Online is built by: Xiangpeng Hao, Tianzheng Wang (Simon Fraser University), Lucas Lersch (TU Dresden & SAP SE) and Ismail Oukid (Snowflake Computing).

If you use PiBench Online in your work, please cite our VLDB 2020 demo paper:

Xiangpeng Hao, Lucas Lersch, Tianzheng Wang, Ismail Oukid:
PiBench Online: Interactive Benchmarking of Persistent Memory Indexes. PVLDB 13 (2020)

Our own range index evaluation using PiBench is presented in the following VLDB 2020 paper:

Lucas Lersch, Xiangpeng Hao, Ismail Oukid, Tianzheng Wang, Thomas Willhalm:
Evaluating Persistent Memory Range Indexes. PVLDB 13(4): 574-587 (2019)

Request PiBench resources: [email protected]

Contribution

Backend

The backend server requires rust nightly, which you can install from rustup

When rust toolchain is properly setup, building pibench online is as simple as:

cargo build [--release]

Frontend

The PiBench Online frontend requires nodejs and the yarn package manager.

You also need to be familiar with Vue.js and ElementUI as well as Echarts.js.

Project setup

yarn install

Compiles and hot-reloads for development

yarn serve

Compiles and minifies for production

yarn build