PandA Bambu HLS Framework
The primary objective of the PandA project is to develop a usable framework that will enable the research of new ideas in the HW-SW Co-Design field.
The PandA framework includes methodologies supporting the research on high-level synthesis of hardware accelerators, on parallelism extraction for embedded systems, on hardware/software partitioning and mapping, on metrics for performance estimation of embedded software applications and on dynamic reconfigurable devices.
PandA is free software, free in the sense that it respects the user’s freedom, released under the GNU General Public License version 3 and being developed at Politecnico di Milano (Italy).
The source files currently distributed mainly cover the high-level synthesis of C/C++ based descriptions. In particular, the tool Bambu provides a research environment to experiment with new ideas across HLS, high-level verification and debugging, FPGA/ASIC design, design flow space exploration, and parallel hardware accelerator design. Bambu accepts as input standard C/C++ specifications and compiler intermediate representations coming from the well-known Clang/LLVM and GCC compilers. The broad spectrum and flexibility of input formats allow the electronic design automation research community to explore and integrate new transformations and optimizations.
If you use Bambu in your research, please cite:
@INPROCEEDINGS{ferrandi2021bambu,
author={Ferrandi, Fabrizio and Castellana, Vito Giovanni
and Curzel, Serena and Fezzardi, Pietro and Fiorito, Michele
and Lattuada, Marco and Minutoli, Marco and Pilato, Christian
and Tumeo, Antonino},
booktitle={2021 58th ACM/IEEE Design Automation Conference (DAC)},
title={Invited: Bambu: an Open-Source Research Framework for the
High-Level Synthesis of Complex Applications},
year={2021},
pages={1327-1330},
abstract = {This paper presents the open-source high-level synthesis (HLS) research
framework Bambu. Bambu provides a research environment to experiment with
new ideas across HLS, high-level verification and debugging, FPGA/ASIC design,
design flow space exploration, and parallel hardware accelerator design. The
tool accepts as input standard C/C++ specifications and compiler intermediate
representations (IRs) coming from the well-known Clang/LLVM and GCC compilers.
The broad spectrum and flexibility of input formats allow the electronic
design automation (EDA) research community to explore and integrate new
transformations and optimizations. The easily extendable modular framework
already includes many optimizations and HLS benchmarks used to evaluate
the QoR of the tool against existing approaches [1]. The integration with
synthesis and verification backends (commercial and open-source) allows
researchers to quickly test any new finding and easily obtain performance
and resource usage metrics for a given application. Different FPGA devices
are supported from several different vendors: AMD/Xilinx, Intel/Altera,
Lattice Semiconductor, and NanoXplore. Finally, integration with the OpenRoad
open-source end-to-end silicon compiler perfectly fits with the recent push
towards open-source EDA.},
publisher={{IEEE}},
doi={10.1109/DAC18074.2021.9586110},
ISSN={0738-100X},
month={Dec},
pdf={https://re.public.polimi.it/retrieve/668507/dac21_bambu.pdf}
}
Installation instructions
Installation instructions for many different operation systems are available at https://panda.dei.polimi.it/?page_id=88 or in the INSTALL file in this repository.
You may also download pre-compiled AppImage distributions and VMs at https://panda.dei.polimi.it/?page_id=81 or you can generate them yourself following the instructions under etc/Appimage and etc/VMs.
Dockerfiles are also available under etc/containers along with a VS Code devcontainer configuration which may be useful for development purposes. Pre-built docker images can be downloaded at Docker Hub.
A Google Colab notebook with many examples to play with Bambu is available.
Contacts
Issues, patches, and pull requests are welcome at https://github.com/ferrandi/PandA-bambu.
For further information send an e-mail to [email protected], visit PandA website or the Google group page.