A simple deep learning framework that optimizes task scheduling and memory usuage on different CPU/GPU architectures.
branch | build status |
---|---|
master | |
feng |
git submodule update --init
mkdir build
cd build
We use mkl for cpu gemm().
source /opt/intel/bin/compilervars.sh intel64
source /opt/intel/mkl/bin/mklvars.sh intel64
Then build with
cmake -DUSE_MKL=on -DAWNN_USE_FLT32=on ..
in the builddir, run
../scripts/build_stampede2.sh
When mkl is not avaible install openblas and build with -DUSE_OPENBLAS=on
sudo ./install-apt.sh
cmake -DUSE_OPENBLAS=on -DAWNN_USE_FLT32=on ..