Toolchain 2023.5 (#3196)

* Update install_elpa.sh update elpa version to 2023.05.001 * Update cmake version to 3.27.6 * Update gcc version to 13.2.0 * Update install_openblas.sh - update OpenBLAS version to 0.3.24 - update NUM_THREADS to 128 in compiling * Set default intel compiler to `icpx` * update build.sh scripts * Update README.md * Update README.md * modified toolchain*.sh * Update toolchain_intel-mpich.sh * remove error message in build*.sh when do install * Update README.md * minor update * Update README.md * version 2023.5 tag * Update README for GPU version * update optional message in README
deepmodeling · Nov 11, 2023 · d62ca6f · d62ca6f
1 parent 35bd3a7
commit d62ca6f
Show file tree

Hide file tree

Showing 13 changed files with 133 additions and 67 deletions.
diff --git a/toolchain/README.md b/toolchain/README.md
@@ -1,5 +1,5 @@
 # The ABACUS Toolchain
-Version 2023.4
+Version 2023.5
 
 ## Author
 [QuantumMisaka](https://github.com/QuantumMisaka) 
@@ -31,14 +31,16 @@ and give setup files that you can use to compile ABACUS.
 - [ ] Better compliation method for ABACUS-DEEPMD and ABACUS-DEEPKS.
 - [ ] A better `setup` and toolchain code structure.
 - [ ] Modulefile generation scripts.
-- [ ] Support for `acml` toolchain (scripts are partly in toolchain now) or other AMD compiler and math lib like `AOCL` and `AOCC`
+- [ ] Support for AMD compiler and math lib like `AOCL` and `AOCC`
 
 
 ## Usage Online & Offline
 Main script is `install_abacus_toolchain.sh`, 
 which will use scripts in `scripts` directory 
 to compile install dependencies of ABACUS.
 
+You can just `./install_abacus_toolchain.sh -h` to get more help message.
+
 **Notice: You SHOULD `source` or `module load` related environments before use toolchain method for installation, espacially for `gcc` or `intel-oneAPI` !!!! for example, `module load mkl mpi icc compiler`**
 
 **Notice: You SHOULD keep your environments systematic, for example, you CANNOT load `intel-OneAPI` environments while use gcc toolchain !!!**
@@ -51,7 +53,7 @@ to compile install dependencies of ABACUS.
 
 All packages will be downloaded from [cp2k-static/download](https://www.cp2k.org/static/downloads). by  `wget` , and will be detailedly compiled and installed in `install` directory by toolchain scripts, despite of:
 - `CEREAL` which will be downloaded from [CEREAL](https://github.com/USCiLab/cereal)  
-- `LibNPY` which will be downloaded from [LIBNPY](https://github.com/llohse/libnpy)
+- `Libnpy` which will be downloaded from [LIBNPY](https://github.com/llohse/libnpy)
 - `LibRI` which will be downloaded from [LibRI](https://github.com/abacusmodeling/LibRI)
 - `LibCOMM` which will be downloaded from [LibComm](https://github.com/abacusmodeling/LibComm)
 Notice: These packages will be downloaded by `wget` from `github.com`, which is hard to be done in Chinese Internet. You may need to use offline installation method. 
@@ -78,7 +80,28 @@ just by using this toolchain
 > cp ***.tar.gz build/
 ```
 
-Notice: for `CEREAL`, `LibNPY`, `LibRI` and `LibCOMM`, 
+The needed dependencies version default:
+- `cmake` 3.27.6
+- `gcc` 13.2.0 (which will always NOT be installed, But use system)
+- `OpenMPI` 4.1.5
+- `MPICH` 4.1.2
+- `OpenBLAS` 0.3.24 (Intel toolchain need `get_vars.sh` tool from it)
+- `ScaLAPACK` 2.2.1
+- `FFTW` 3.3.10
+- `LibXC` 6.2.2
+- `ELPA` 2023.05.001
+- `CEREAL` 1.3.2
+And Intel-oneAPI need user or server manager to manually install from Intel.
+[Intel-oneAPI](https://www.intel.cn/content/www/cn/zh/developer/tools/oneapi/toolkits.html)
+
+Dependencies below are optional， which is NOT installed by default:
+- `LibTorch` 2.0.1
+- `Libnpy` 0.1.0
+- `LibRI` 0.1.0
+- `LibComm` 0.1.0
+Users can install them by using `--with-*=install` in toolchain*.sh, which is `no` in default.
+
+Notice: for `CEREAL`, `Libnpy`, `LibRI` and `LibComm`, 
 you need to download them from github.com, 
 rename it as formatted, and put them in `build` directory at the same time
 e.g.:
@@ -118,7 +141,7 @@ If compliation is successful, a message will be shown like this:
 >     ./build_abacus_intel.sh
 > or you can modify the builder scripts to suit your needs.
 ```
-You can run build_abacus_gnu.sh or build_abacus_intel.sh to build ABACUS 
+You can run `build_abacus_gnu.sh` or `build_abacus_intel.sh` to build ABACUS 
 by gnu-toolchain or intel-toolchain respectively, the builder scripts will
 automatically locate the environment and compile ABACUS.
 You can manually change the builder scripts to suit your needs.
@@ -139,13 +162,20 @@ or you can also do it in a more completely way:
 > rm -rf install build/*/* build/OpenBLAS*/ build/setup_*
 ```
 
-Users can get help messages by simply:
+## Common Problem and Solution
+### GPU version of ABACUS
+add following options in build*.sh:
 ```shell
-> ./install_abacus_toolchain.sh -h # or --help
+cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
+        -DCMAKE_CXX_COMPILER=icpx \
+        -DMPI_CXX_COMPILER=mpiicpc \
+        ......
+        -DUSE_CUDA=1 \
+        -DCMAKE_CUDA_COMPILER=${path to cuda toolkit}/bin/nvcc \
+        ......
 ```
 
-
-## Common Problem and Solution
+### shell problem
 If you encounter problem like:
 ```shell
 /bin/bash^M: bad interpreter: No such file or directory
@@ -157,6 +187,36 @@ or   `permission denied` problem, you can simply run:
 And also, you can fix `permission denied` problem via `chmod +x`
 if `pre_set.sh` have no execution permission.
 
+### libtorch and deepks problem
+If deepks feature have problem, you can manually change libtorch version
+from 2.0.1 to 1.12.0 in `toolchain/scripts/stage4/install_libtorch.sh`.
+Also, you can install ABACUS without deepks by removing all the deepks and related options.
+
+NOTICE: if you want deepks feature, your intel-mkl environment should be accessible in building process. you can check it in `build_abacus_gnu.sh`
+
+### deepmd feature problem
+When you encounter problem like `GLIBCXX_3.4.29 not found`, it is sure that your `gcc` version is lower than the requirement of `libdeepmd`.
+
+After my test, you need `gcc`>11.3.1 to enable deepmd feature in ABACUS.
+
+### ELPA problem via Intel-oneAPI toolchain in AMD server
+The default compiler for Intel-oneAPI is `icpx` and `icx`, which will cause problem when compling ELPA in AMD server.
+
+The best way is to change `icpx` to `icpc`, `icx` to `icc`. user can manually change it in toolchain*.sh via `--with-intel-classic=yes`
+
+
+### LibRI and LibComm problem
+(There is some problem sometimes when compling with LibRI and LibComm, detailed information is needed)
+
+
+### Intel-oneAPI problem
+Sometimes Intel-oneAPI have problem to link `mpirun`, 
+which will always show in 2023.2.0 version of MPI in Intel-oneAPI. 
+Try `source /path/to/setvars.sh` or install another version of IntelMPI may help.
+
+More problem and possible solution can be accessed via [#2928](https://github.com/deepmodeling/abacus-develop/issues/2928)
+
+
 
 ## Advanced Installation Usage
 

diff --git a/toolchain/build_abacus_gnu.sh b/toolchain/build_abacus_gnu.sh
@@ -18,13 +18,13 @@ source $INSTALL_DIR/setup
 cd $ABACUS_DIR
 ABACUS_DIR=$(pwd)
 
-BUILD_DIR=build_abacus
+BUILD_DIR=build_abacus_gnu
 rm -rf $BUILD_DIR
 
 PREFIX=$ABACUS_DIR
-LAPACK=$INSTALL_DIR/openblas-0.3.23/lib
+LAPACK=$INSTALL_DIR/openblas-0.3.24/lib
 SCALAPACK=$INSTALL_DIR/scalapalack-2.2.1/lib
-ELPA=$INSTALL_DIR/elpa-2021.11.002/cpu
+ELPA=$INSTALL_DIR/elpa-2023.05.001/cpu
 FFTW3=$INSTALL_DIR/fftw-3.3.10
 CEREAL=$INSTALL_DIR/cereal-1.3.2/include/cereal
 LIBXC=$INSTALL_DIR/libxc-6.2.2
@@ -46,7 +46,6 @@ cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
         -DENABLE_LCAO=ON \
         -DENABLE_LIBXC=ON \
         -DUSE_OPENMP=ON \
-        -DENABLE_ASAN=OFF \
         -DUSE_ELPA=ON \
 #         -DENABLE_DEEPKS=1 \
 #         -DTorch_DIR=$LIBTORCH \
@@ -58,12 +57,14 @@ cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
 # 	      -DTensorFlow_DIR=$DEEPMD \
 
 # # add mkl env for libtorch to link
-# # gnu-toolchain will lack of -lmkl when load libtorch
-# # need to fix -- zhaoqing in 2023-09-02
+# if one want to install libtorch, mkl should be load in build process
+# for -lmkl when load libtorch
 # module load mkl
 
+# if one want's to include deepmd, your gcc version should be >= 11.3.0
+
 cmake --build $BUILD_DIR -j `nproc` 
-cmake --install $BUILD_DIR 
+cmake --install $BUILD_DIR 2>/dev/null
 
 # generate abacus_env.sh
 cat << EOF > "${TOOL}/abacus_env.sh"

diff --git a/toolchain/build_abacus_intel-mpich.sh b/toolchain/build_abacus_intel-mpich.sh
@@ -4,7 +4,7 @@
 #SBATCH -n 16
 #SBATCH -o install.log
 #SBATCH -e install.err
-# install ABACUS with libxc and deepks
+# build and install ABACUS with libxc, also can with deepks and deepmd
 # JamesMisaka in 2023.08.31
 
 # Build ABACUS by intel-toolchain with mpich
@@ -19,21 +19,21 @@ source $INSTALL_DIR/setup
 cd $ABACUS_DIR
 ABACUS_DIR=$(pwd)
 
-BUILD_DIR=build_abacus
+BUILD_DIR=build_abacus_intel-mpich
 rm -rf $BUILD_DIR
 
 PREFIX=$ABACUS_DIR
-ELPA=$INSTALL_DIR/elpa-2021.11.002/cpu
+ELPA=$INSTALL_DIR/elpa-2023.05.001/cpu
 CEREAL=$INSTALL_DIR/cereal-1.3.2/include/cereal
 LIBXC=$INSTALL_DIR/libxc-6.2.2
-LIBTORCH=$INSTALL_DIR/libtorch-2.0.1/share/cmake/Torch
-LIBNPY=$INSTALL_DIR/libnpy-0.1.0/include
+# LIBTORCH=$INSTALL_DIR/libtorch-2.0.1/share/cmake/Torch
+# LIBNPY=$INSTALL_DIR/libnpy-0.1.0/include
 # LIBRI=$INSTALL_DIR/LibRI-0.1.0
 # LIBCOMM=$INSTALL_DIR/LibComm-0.1.0
 # DEEPMD=$HOME/apps/anaconda3/envs/deepmd
 
 cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
-        -DCMAKE_CXX_COMPILER=icpc \
+        -DCMAKE_CXX_COMPILER=icpx \
         -DMPI_CXX_COMPILER=mpicxx \
         -DMKLROOT=$MKLROOT \
         -DELPA_DIR=$ELPA \
@@ -42,19 +42,20 @@ cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
         -DENABLE_LCAO=ON \
         -DENABLE_LIBXC=ON \
         -DUSE_OPENMP=ON \
-        -DENABLE_ASAN=OFF \
         -DUSE_ELPA=ON \
-        -DENABLE_DEEPKS=1 \
-        -DTorch_DIR=$LIBTORCH \
-        -Dlibnpy_INCLUDE_DIR=$LIBNPY \
+        # -DENABLE_DEEPKS=1 \
+        # -DTorch_DIR=$LIBTORCH \
+        # -Dlibnpy_INCLUDE_DIR=$LIBNPY \
         # -DENABLE_LIBRI=ON \
         # -DLIBRI_DIR=$LIBRI \
         # -DLIBCOMM_DIR=$LIBCOMM \
 	    # -DDeePMD_DIR=$DEEPMD \
 	    # -DTensorFlow_DIR=$DEEPMD \
 
+# if one want's to include deepmd, your gcc version should be >= 11.3.0
+
 cmake --build $BUILD_DIR -j `nproc` 
-cmake --install $BUILD_DIR 
+cmake --install $BUILD_DIR 2>/dev/null
 
 # generate abacus_env.sh
 cat << EOF > "${TOOL}/abacus_env.sh"

diff --git a/toolchain/build_abacus_intel.sh b/toolchain/build_abacus_intel.sh
@@ -19,44 +19,44 @@ source $INSTALL_DIR/setup
 cd $ABACUS_DIR
 ABACUS_DIR=$(pwd)
 
-BUILD_DIR=build_abacus
+BUILD_DIR=build_abacus_intel
 rm -rf $BUILD_DIR
 
 PREFIX=$ABACUS_DIR
-ELPA=$INSTALL_DIR/elpa-2021.11.002/cpu
+ELPA=$INSTALL_DIR/elpa-2023.05.001/cpu
 CEREAL=$INSTALL_DIR/cereal-1.3.2/include/cereal
 LIBXC=$INSTALL_DIR/libxc-6.2.2
-LIBTORCH=$INSTALL_DIR/libtorch-2.0.1/share/cmake/Torch
-LIBNPY=$INSTALL_DIR/libnpy-0.1.0/include
+# LIBTORCH=$INSTALL_DIR/libtorch-2.0.1/share/cmake/Torch
+# LIBNPY=$INSTALL_DIR/libnpy-0.1.0/include
 # LIBRI=$INSTALL_DIR/LibRI-0.1.0
 # LIBCOMM=$INSTALL_DIR/LibComm-0.1.0
 # DEEPMD=$HOME/apps/anaconda3/envs/deepmd
 
 # if use deepks and deepmd
 cmake -B $BUILD_DIR -DCMAKE_INSTALL_PREFIX=$PREFIX \
-        -DCMAKE_CXX_COMPILER=icpc \
+        -DCMAKE_CXX_COMPILER=icpx \
         -DMPI_CXX_COMPILER=mpiicpc \
         -DMKLROOT=$MKLROOT \
         -DELPA_DIR=$ELPA \
         -DCEREAL_INCLUDE_DIR=$CEREAL \
         -DLibxc_DIR=$LIBXC \
         -DENABLE_LCAO=ON \
         -DENABLE_LIBXC=ON \
-        -DENABLE_LIBRI=OFF \
         -DUSE_OPENMP=ON \
-        -DENABLE_ASAN=OFF \
         -DUSE_ELPA=ON \
-        -DENABLE_DEEPKS=1 \
-        -DTorch_DIR=$LIBTORCH \
-        -Dlibnpy_INCLUDE_DIR=$LIBNPY \
+#         -DENABLE_DEEPKS=1 \
+#         -DTorch_DIR=$LIBTORCH \
+#         -Dlibnpy_INCLUDE_DIR=$LIBNPY \
 #         -DENABLE_LIBRI=ON \
 #         -DLIBRI_DIR=$LIBRI \
 #         -DLIBCOMM_DIR=$LIBCOMM \
 # 	      -DDeePMD_DIR=$DEEPMD \
 # 	      -DTensorFlow_DIR=$DEEPMD \
 
 cmake --build $BUILD_DIR -j `nproc` 
-cmake --install $BUILD_DIR 
+cmake --install $BUILD_DIR 2>/dev/null
+
+# if one want's to include deepmd, your gcc version should be >= 11.3.0
 
 # generate abacus_env.sh
 cat << EOF > "${TOOL}/abacus_env.sh"

diff --git a/toolchain/install_abacus_toolchain.sh b/toolchain/install_abacus_toolchain.sh
@@ -118,7 +118,7 @@ The --enable-FEATURE options follow the rules:
   --enable-FEATURE=no     Disable this particular feature
   --enable-FEATURE        The option keyword alone is equivalent to
                           --enable-FEATURE=yes
-
+  ===== NOTICE: THESE FEATURE AER NOT INCLUDED IN ABACUS =====
   --enable-cuda           Turn on GPU (CUDA) support (can be combined
                           with --enable-opencl).
                           Default = no
@@ -302,9 +302,9 @@ enable_tsan="__FALSE__"
 enable_opencl="__FALSE__"
 enable_cuda="__FALSE__"
 enable_hip="__FALSE__"
-export intel_classic="yes" 
+export intel_classic="no" 
 # no, then icc->icx, icpc->icpx, 
-# which cannot compile elpa-2021 and fftw.3.3.10 in some place
+# which cannot compile elpa in AMD server
 # due to some so-called cross-compile problem
 # and will lead to problem in force calculation
 # but icx is recommended by intel compiler

diff --git a/toolchain/scripts/VERSION b/toolchain/scripts/VERSION
@@ -1,2 +1,2 @@
 # version file to force a rebuild of the entire toolchain
-VERSION="2023.4"
+VERSION="2023.5"
diff --git a/toolchain/scripts/stage0/install_cmake.sh b/toolchain/scripts/stage0/install_cmake.sh
@@ -20,13 +20,13 @@ cd "${BUILDDIR}"
 case "${with_cmake}" in
   __INSTALL__)
     echo "==================== Installing CMake ===================="
-    cmake_ver="3.26.3"
+    cmake_ver="3.27.6"
     if [ "${OPENBLAS_ARCH}" = "arm64" ]; then
       cmake_arch="linux-aarch64"
-      cmake_sha256="b002c22b926aacd6fefe64bcf08620216088eb72f55ac532b7bcfd4d93443d50"
+      cmake_sha256="a83e01ed1cdf44c2e33e0726513b9a35a8c09e3b5a126fd720b3c8a9d5552368"
     elif [ "${OPENBLAS_ARCH}" = "x86_64" ]; then
       cmake_arch="linux-x86_64"
-      cmake_sha256="8ec0ef24375a1d0e78de2f790b4545d0718acc55fd7e2322ecb8e135696c77fe"
+      cmake_sha256="8c449dabb2b2563ec4e6d5e0fb0ae09e729680efab71527b59015131cea4a042"
     else
       report_error ${LINENO} \
         "cmake installation for ARCH=${ARCH} is not supported. You can try to use the system installation using the flag --with-cmake=system instead."

diff --git a/toolchain/scripts/stage0/install_gcc.sh b/toolchain/scripts/stage0/install_gcc.sh
@@ -6,8 +6,8 @@
 [ "${BASH_SOURCE[0]}" ] && SCRIPT_NAME="${BASH_SOURCE[0]}" || SCRIPT_NAME=$0
 SCRIPT_DIR="$(cd "$(dirname "$SCRIPT_NAME")/.." && pwd -P)"
 
-gcc_ver="13.1.0"
-gcc_sha256="bacd4c614d8bd5983404585e53478d467a254249e0f1bb747c8bc6d787bd4fa2"
+gcc_ver="13.2.0"
+gcc_sha256="8cb4be3796651976f94b9356fa08d833524f62420d6292c5033a9a26af315078"
 
 source "${SCRIPT_DIR}"/common_vars.sh
 source "${SCRIPT_DIR}"/tool_kit.sh

diff --git a/toolchain/scripts/stage2/install_openblas.sh b/toolchain/scripts/stage2/install_openblas.sh
@@ -6,8 +6,8 @@
 [ "${BASH_SOURCE[0]}" ] && SCRIPT_NAME="${BASH_SOURCE[0]}" || SCRIPT_NAME=$0
 SCRIPT_DIR="$(cd "$(dirname "$SCRIPT_NAME")/.." && pwd -P)"
 
-openblas_ver="0.3.23" # Keep in sync with get_openblas_arch.sh
-openblas_sha256="5d9491d07168a5d00116cdc068a40022c3455bf9293c7cb86a65b1054d7e5114"
+openblas_ver="0.3.24" # Keep in sync with get_openblas_arch.sh
+openblas_sha256="ceadc5065da97bd92404cac7254da66cc6eb192679cf1002098688978d4d5132"
 openblas_pkg="OpenBLAS-${openblas_ver}.tar.gz"
 
 source "${SCRIPT_DIR}"/common_vars.sh
@@ -76,7 +76,7 @@ case "${with_openblas}" in
         make -j $(get_nprocs) \
           MAKE_NB_JOBS=0 \
           TARGET=${TARGET} \
-          NUM_THREADS=64 \
+          NUM_THREADS=128 \
           USE_THREAD=1 \
           USE_OPENMP=1 \
           NO_AFFINITY=1 \
@@ -88,7 +88,7 @@ case "${with_openblas}" in
         make -j $(get_nprocs) \
           MAKE_NB_JOBS=0 \
           TARGET=NEHALEM \
-          NUM_THREADS=64 \
+          NUM_THREADS=128 \
           USE_THREAD=1 \
           USE_OPENMP=1 \
           NO_AFFINITY=1 \
@@ -100,7 +100,7 @@ case "${with_openblas}" in
       make -j $(get_nprocs) \
         MAKE_NB_JOBS=0 \
         TARGET=${TARGET} \
-        NUM_THREADS=64 \
+        NUM_THREADS=128 \
         USE_THREAD=1 \
         USE_OPENMP=1 \
         NO_AFFINITY=1 \