docker: add xpu-smi to intel image for device detection

Author: mloubout

.github/workflows/docker-bases.yml

@@ -17,7 +17,7 @@ on:
         description: 'Build compiler bases'
   schedule:
     # Run once a week
-    - cron: "0 13 * * 1"
+    - cron: "0 0 1 * *"
 
 jobs:
 #######################################################

docker/Dockerfile.amd

@@ -35,6 +35,7 @@ RUN rm /etc/apt/sources.list.d/* \
     gfortran \
     flex \
     hwloc \
+    gdb \
   && apt-get clean
 
 # Requires cmake > 3.22 

docker/Dockerfile.cpu

@@ -17,7 +17,7 @@ RUN apt-get update && \
 
 # Install python
 RUN apt-get update && \
-    apt-get install -y dh-autoreconf python3-venv python3-dev python3-pip
+    apt-get install -y dh-autoreconf python3-venv python3-dev python3-pip gdb
 
 # Install for basic base not containing it
 RUN apt-get install -y vim wget git flex libnuma-dev tmux \
@@ -27,14 +27,14 @@ RUN apt-get install -y vim wget git flex libnuma-dev tmux \
 
 # Install compilersx
 RUN if [ -n "$gcc" ]; then \
-        apt-get install libgmp-dev libmpc-dev libmpfr-dev libisl-dev -y && \
+        apt-get install libgmp-dev libmpc-dev libmpfr-dev libisl-dev binutils texinfo -y && \
         cd /tmp && mkdir gcc && \
         git clone git://gcc.gnu.org/git/gcc.git gcc && \
         cd gcc && git checkout releases/gcc-${gcc} && \
         mkdir build && cd build && \
         ../configure --prefix=/opt/gcc-${gcc} --enable-languages=c,c++,fortran --disable-multilib \
                      --enable-shared --enable-lto --enable-threads=posix --enable-libstdcxx-time=yes --enable-libgomp && \
-        make -j $(nproc) && make install && \
+        make -j $(nproc) && make install-strip && \
         # Update alternatives
         update-alternatives --install /usr/bin/gcc gcc /opt/gcc-${gcc}/bin/gcc 60 && \
         update-alternatives --install /usr/bin/g++ g++ /opt/gcc-${gcc}/bin/g++ 60 && \
@@ -59,6 +59,7 @@ RUN cd /tmp && mkdir openmpi && \
     cd openmpi && ./autogen.pl && \
     mkdir build &&  cd build && \
     ../configure --prefix=/opt/openmpi/ \
+                 --disable-mpi-fortran \
                  --enable-mca-no-build=btl-uct --enable-mpi1-compatibility && \
     make -j ${nproc} && \
     make install && \

docker/Dockerfile.intel

@@ -10,7 +10,7 @@ ENV DEBIAN_FRONTEND noninteractive
 
 # Install python
 RUN apt-get update && \
-    apt-get install -y dh-autoreconf python3-venv python3-dev python3-pip
+    apt-get install -y dh-autoreconf python3-venv python3-dev python3-pip gdb
 
 # Install for basic base not containing it
 RUN apt-get install -y vim wget git flex libnuma-dev tmux \
@@ -56,9 +56,16 @@ RUN apt-get update -y && apt-get dist-upgrade -y && \
             libegl-mesa0 libegl1-mesa libegl1-mesa-dev libgbm1 libgl1-mesa-dev libgl1-mesa-dri \
             libglapi-mesa libgles2-mesa-dev libglx-mesa0 libigdgmm12 libxatracker2 mesa-va-drivers \
             mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo \
+            # XPU-smi requirements
+            intel-gsc  libmfx-tools intel-metrics-discovery intel-metrics-library \
             # Development packages
             libigc-dev intel-igc-cm libigdfcl-dev libigfxcmrt-dev level-zero-dev
 
+# Install xpu-smi for device detection
+RUN wget https://github.com/intel/xpumanager/releases/download/V1.2.39/xpu-smi_1.2.39_20240906.085820.11f3c29a+deb10u1_amd64.deb && \
+    apt install -y ./xpu-smi_1.2.39_20240906.085820.11f3c29a+deb10u1_amd64.deb && \
+    rm xpu-smi_1.2.39_20240906.085820.11f3c29a+deb10u1_amd64.deb
+
 ENV MPI4PY_FLAGS='. /opt/intel/oneapi/setvars.sh intel64 && '
 ENV MPI4PY_RC_RECV_MPROBE=0
 

docker/Dockerfile.nvidia

@@ -16,7 +16,7 @@ ENV DEBIAN_FRONTEND noninteractive
 # Install python
 RUN apt-get update && \
     apt-get install -y -q gpg apt-utils curl wget vim libnuma-dev tmux numactl \
-                          dh-autoreconf python3-venv python3-dev python3-pip
+                          dh-autoreconf python3-venv python3-dev python3-pip gdb
 
 # nodesource: nvdashboard requires nodejs>=10
 RUN curl https://developer.download.nvidia.com/hpc-sdk/ubuntu/DEB-GPG-KEY-NVIDIA-HPC-SDK | gpg --yes --dearmor -o /usr/share/keyrings/nvidia-hpcsdk-archive-keyring.gpg