High-level synthesis task-level parallelization
Project description
Task-Level Parallelization for HLS
Feature Synopsis
- software simulation (not cycle-accurate)
- Xilinx HLS backend
- HeteroCL frontend
Application Synopsis
App | Properties | Details | # Streams | # Tasks | # Steps |
---|---|---|---|---|---|
bandwidth |
static | bandwidth test using async_mmap |
4 | 4 | 1 |
cannon |
static,feedback | Cannon's algorithm | 20 | 7 | 1 |
graph |
dynamic,feedback | connected components | 6 | 3 | 1 |
jacobi |
static | 5-point stencil | 23 | 17 | 1 |
vadd |
static | naïve vector addition | 3 | 4 | 1 |
Getting Started
Prerequisites
- Ubuntu 16.04+
Install from Binary
./install.sh
Install from Source
Build Prerequisites
- CMake 3.13+
- A C++11 compiler
- Google glog library
- Clang headers
Build tlpcc
mkdir build
cd build
cmake ..
make
make test
cd ..
sudo ln -s backend/python/tlpc /usr/local/bin/
sudo ln -s build/backend/tlpcc /usr/local/bin/
Known Issues
- Template functions cannot be tasks
- Vivado HLS include paths (e.g.,
/opt/Xilinx/Vivado/2019.2/include
) must not be specified intlpc --cflags
;- Workaround is to
export CPATH=/opt/Xilinx/Vivado/2019.2/include
- Workaround is to
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
tlpc-0.0.20200711.1.tar.gz
(28.9 kB
view hashes)
Built Distributions
tlpc-0.0.20200711.1-py3.6.egg
(70.0 kB
view hashes)
Close
Hashes for tlpc-0.0.20200711.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 23ee0d0386dd6844fbdd2d0af43dceb38235a73993d04bb8132ac0395301af17 |
|
MD5 | a91c3599355a9f209f91c1357c358347 |
|
BLAKE2b-256 | 37567a6f58342d2fd8d8cb78fa9f42ea81bd8d6117bd3abe26de315a45f6640e |