Ripple Fpga Save

RippleFPGA, A Simultaneous Pack-and-Place Algorithm for UltraScale FPGA

Project README

RippleFPGA

RippleFPGA is a simultaneous pack-and-place algorithm for FPGA developed by the research team supervised by Prof. Evangeline F. Y. Young in The Chinese University of Hong Kong (CUHK). It produces legalized placement solutions by effectively packing and placing the input netlist on the modern Xilinx FPGA architecture. It directly minimizes the routed wirelength generated by the commercial router.

Features

  • Target at modern commercial FPGA architecture (Xilinx UltraScale XU095)
  • Support large-scale circuits
  • Pack and place simultaneously with an analytical global placement engine (Ripple)
  • Optimize both wirelength and routing congestion
  • Results verified by Xilinx Vivado® Design Suite
  • ...

More details are in the following papers:

1. How to Build

Step 1: Download the source codes. For example,

$ git clone https://github.com/cuhk-eda/ripple-fpga

Step 2: Go to the project root and build by

$ cd ripple-fpga/src
$ make mode=release_mt

Note that this will generate a folder bin under the root, which contains binaries and auxiliary files. More details are in Makefile.

Dependencies

  • GCC (version 4.8.0) or other working c++ compliers
  • Boost (version >= 1.58)

2. How to Run

Toy Test

Go to the bin directory and run binary placer with a toy design:

$ cd bin
$ ./placer -aux toy_example/design.aux -out toy_example.out

Batch Test

Our placer uses the bookshelf format and has been tested on two contest benchmark suites ISPD'16 Contest and ISPD'17 Contest, which can be downloaded via Dropbox (ISPD'16, ISPD'17). After downloading the benchmarks and putting them under folder BM_DIR, you can use our script run.sh:

$ cd bin
$ export BENCHMARK_PATH=BM_DIR
$ ./run all

3. Modules

  • src: c++ source codes
    • alg: external algorithm packages
    • cong: congestion estimation
    • db: database
    • dp: detailed placement
    • gp: global placement
    • lg: legalization
    • lgclk: legailzation related to clock constraints
    • pack: packing
    • utils: utilities
  • toy_example: toy example in bookshelf format

4. Results

All the experiments were performed on a 64-bit Linux workstation with Intel Xeon 3.7GHz CPU and 16GB memory. The placer is run in 2 threads and routed wirelength is reported by Xilinx Vivado® Design Suite with patches provided by the ISPD contest orgranizers.

Results on ISPD'16 benchmarks

Results on ISPD'17 benchmarks

5. License

READ THIS LICENSE AGREEMENT CAREFULLY BEFORE USING THIS PRODUCT. BY USING THIS PRODUCT YOU INDICATE YOUR ACCEPTANCE OF THE TERMS OF THE FOLLOWING AGREEMENT. THESE TERMS APPLY TO YOU AND ANY SUBSEQUENT LICENSEE OF THIS PRODUCT.

License Agreement for RippleFPGA

Copyright (c) 2019 by The Chinese University of Hong Kong

All rights reserved

CU-SD LICENSE (adapted from the original BSD license) Redistribution of the any code, with or without modification, are permitted provided that the conditions below are met.

Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

Neither the name nor trademark of the copyright holder or the author may be used to endorse or promote products derived from this software without specific prior written permission.

Users are entirely responsible, to the exclusion of the author, for compliance with (a) regulations set by owners or administrators of employed equipment, (b) licensing terms of any other software, and (c) local, national, and international regulations regarding use, including those regarding import, export, and use of encryption software.

THIS FREE SOFTWARE IS PROVIDED BY THE AUTHOR "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR ANY CONTRIBUTOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, EFFECTS OF UNAUTHORIZED OR MALICIOUS NETWORK ACCESS; PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Open Source Agenda is not affiliated with "Ripple Fpga" Project. README Source: cuhk-eda/ripple-fpga
Stars
82
Open Issues
1
Last Commit
4 years ago

Open Source Agenda Badge

Open Source Agenda Rating