diff --git a/README.md b/README.md index 2e73dc27600a27e843171099d2506af0114f9958..8211bbea11a6600af683a9429358b21906c70d53 100644 --- a/README.md +++ b/README.md @@ -22,7 +22,7 @@ HPVM is currently at version 0.5. For more about what HPVM is, see [our website] The following components are required to be installed on your machine to build HPVM. * GCC (>=5.1.0) -* CMake (>=3.4.3) +* CMake (>=3.17.0) * Python (>=2.7) * GNU Make (>=3.79.1) * OpenCL (>=1.0.0) or CUDA (>=9.1, only required for GPU support) @@ -49,6 +49,17 @@ git clone https://gitlab.engr.illinois.edu/llvm/hpvm-release.git/ cd hpvm-release/hpvm ``` +Before installing HPVM, some paths must be set for installation to succeed. The following variables in set_paths.sh must be set: + +* CUDA_TOOLKIT_PATH --- Path to the CUDA toolkit +* CUDA_INCLUDE_PATH --- Path to the CUDA headers +* CUDA_LIB_PATH -- Path to CUDA libraries + +Once the aforementioned variables in set_paths.sh have been specified, run the script. +```shell +source set_paths.sh +``` + HPVM installer script can be used to download, configure and build HPVM along with LLVM and Clang. ```shell bash install.sh diff --git a/hpvm/set_paths.sh b/hpvm/set_paths.sh new file mode 100644 index 0000000000000000000000000000000000000000..cbbada8e291f081114de29a860842d242983ccb2 --- /dev/null +++ b/hpvm/set_paths.sh @@ -0,0 +1,17 @@ +#!/bin/bash + +# These paths can be modified by the HPVM user +CUDA_TOOLKIT_PATH=cuda-toolkit/9.1 +CUDA_INCLUDE_PATH=/software/cuda-9.1/include +CUDA_LIB_PATH=/software/cuda-9.1/lib64/ + +echo "Setting environment paths..." + +# Setting CUDA paths here +module load $CUDA_TOOLKIT_PATH +export CUDA_INCLUDE_PATH=$CUDA_INCLUDE_PATH +export CUDNN_PATH=$CUDA_LIB_PATH +export LIBRARY_PATH=$CUDA_LIB_PATH:$LIBRARY_PATH +export LD_LIBRARY_PATH=$CUDA_LIB_PATH:$LD_LIBRARY_PATH + +echo "Finished setting environment paths!"