Varec-CC

Deep Learning Accelerator Stack

Introduction

1. varecc/ includes varec compiler, runtime driver source files.
2. examples/ includes some end-to-end examples
3. models/ includes pre-built YOLO models
4. hardware/ includes pre-built hardware bitstreams
5. patch/ includes some patches of tvm.

Installation

1. Check TVM for dependencies.
2. Run make to clone tvm and build.
3. Set environment variables in setup.sh and run source setup.sh.

Examples

Simulation

1. Set “TARGET” to “sim” in “varecc/config/varec_config.json”
2. Remake libvarec.so by “cd tvm/build && make varec”
3. In examples directory, run “python3 yolo.py”

On-board (Pynq images)

1. Copy varecc/ and tvm/(without build/) to Pynq (/home/xilinx/)
2. (On Pynq)Set “TARGET” to “pynq” in “varecc/config/varec_config.json”
3. Add varecc/VAREC.cmake into tvm/CMakeLists.txt
4. In tvm directory, “mkdir build && cp cmake/config.cmake build/ && cd build && make runtime varec”
5. Change to varecc directory, run “sudo ./scripts/start_rpc.sh” to start RPC
6. (On the host)Set “TARGET” to “pynq” in “varecc/config/varec_config.json”
7. In examples directory, run “python3 yolo.py”