Build Nvptx-tools on Windows

Question

I tried to install nvptx-tools on Windows in order to enable GCC offload tools, but I couldn't do it. I made a copy of nvptx-tools from here and to build it I followed this tutorial (the part with "build nvptx-tools"). If I put all the commands in a script, it runs and creates only some log files on Windows, while in Linux it build nvptx-tools (I run the script in a vm). I mention that I have installed GCC-2.7.0. Is there any other way to install nvptx-tools on Windows?

My script is:

GCC7ROOT=C:/Users/Marin/Desktop/OpenACCSupport
CUDA=C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v9.1

mkdir -p $GCC7ROOT/build/nvptx-build
cd $GCC7ROOT/build/nvptx-build

$GCC7ROOT/source/nvptx-tools/configure  \
   --prefix=$GCC7ROOT/install \
   --target=nvptx-none \
   --with-cuda-driver-include=$CUDA/include \
   --with-cuda-driver-lib=$CUDA/lib/x64 \
   --with-cuda-runtime-include=$CUDA/include \
   --with-cuda-runtime-lib=$CUDA/lib/x64 \
   CC='gcc -m64' \
   CXX='g++ -m64'

mingw32-make
mingw32-make install

In C:/Users/Marin/Desktop/OpenACCSupport/source I have nvptx-tools sources and I run the script in Windows PowerShell. — Hoza Angelica, Mar 24 '18 at 09:50
Welcome to SO @HozaAngelica. You can use the edit function to add your code to the question and format it properly. — m00am, Mar 24 '18 at 19:38
I was able to build MinGW-w64 GCC 10.2.0 with nvptx offloading. You can find it here: https://github.com/brechtsanders/winlibs_mingw/releases/tag/10.2.0-11.0.0-8.0.0-r4nvptx Now I'm looking for people who can test that it actually works. — Brecht Sanders, Dec 07 '20 at 17:16

Build Nvptx-tools on Windows

0 Answers0