Code Monkey home page Code Monkey logo

shuhai's Introduction

Shuhai

Shuhai is a benchmarking-memory tool that allows FPGA programmers to demystify all the underlying details of memories, e.g., HBM and DDR4, on a Xilinx FPGA.

Two Reasons Why We Need Shuhai?

First, in terms of benchmarking memory, it can be done better on FPGA, rather than on CPU/GPU. In particular, when benchmarking memory on cpu/gpu, we cannot get rid of the negative effect of cache/TLB. So there is a lot of work about benchmarking cache in CPUs/GPUs, rather than directly benchmarking memory. In contrast, when benchmarking memory on an FPGA, the benchmarking hardware engine can directly attach to the memory such that there is no noise between the targeted memory and the benchmarking engine.

Second, we do not need to reinvent a wheel again for each memory/FPGA. With Shuhai, before implementing the concrete application that contains a particular memory access pattern on the FPGA, we are able to benchmark the corresponding memory access pattern to make sure that the memory side will not be the bottleneck. In case it is, the authors need to tune the implementation to have a more efficient memory access pattern such that your application will not be bound by memory.

1. Getting Started

$ git clone https://github.com/RC4ML/Shuhai.git
$ git submodule update --init --recursive

2. Build FPGA Project

$ cd hw/

According to hw/README.md, build vivado project and program the FPGA with the generated bitstream.

3. Build Software Project

$ cd sw/

According to sw/README.md, build the software project and run the application

Frequently Asked Questions

1, Q. the machine failed to detect PCIe on the FPGA when loading the kernel module.

A. Connect the JTAG to another machine that will not crash when downloading the FPGA image. It means that you cannot use the same machine to load the bitstream.

Cite this work

If you use it in your paper, please cite our work (full version).

@inproceedings{wang_fccm20,
  title={Shuhai: Benchmarking High Bandwidth Memory On FPGAs},
  author={Zeke Wang and Hongjing Huang and Jie Zhang and Gustavo Alonso},
  year={2020},
  booktitle={IEEE 28th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM)},
}

@inproceedings{wang_tc21,
  title={Shuhai: A Tool for Benchmarking High Bandwidth Memory On FPGAs},
  author={Hongjing Huang and Zeke Wang and Jie Zhang, Zhenhao He and Chao Wu and Jun Xiao and Gustavo Alonso},
  year={2021},
  booktitle={IEEE Transactions on Computers (TC)}
}


Related publications

shuhai's People

Contributors

qqwert0 avatar wangzeke avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.