Merge branch 'release-2.8.0'

Author: alazzaro

.ci/daint.cscs.ch/ocl.build.sh

@@ -35,7 +35,7 @@ if [ ! -d "${HOME}/libxsmm" ]; then
 fi
 cd "${HOME}/libxsmm"
 git fetch
-git checkout d009b33e8742a93c9e1549323587fb6197451294
+git checkout 488aa88f2a9825e9f92a0cfc773c1aedf019f88a
 make -j
 cd ..
 

.github/workflows/testing-linux.yml

@@ -74,16 +74,16 @@ jobs:
         mv build/coverage.info build/coverage-Linux-${{ matrix.use_mpi }}-${{ matrix.use_openmp }}-${{ matrix.use_smm }}-cpu.info
 
     - name: Upload coverage data
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
-        name: coverage-data
+        name: coverage-data-${{ matrix.use_mpi }}-${{ matrix.use_openmp }}-${{ matrix.use_smm }}-${{ matrix.mpi_suffix }}
         path: build/coverage-*.info
 
     - name: Upload coverage data (generated files)
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       if: matrix.use_mpi == 'MPI=ON' && matrix.use_openmp == 'OPENMP=ON' && matrix.use_smm == 'SMM=blas' && matrix.mpi_suffix == 'openmpi'
       with:
-        name: coverage-data
+        name: coverage-data-${{ matrix.use_mpi }}-${{ matrix.use_openmp }}-${{ matrix.use_smm }}-${{ matrix.mpi_suffix }}-generated-files
         path: |
           build/src/dbcsr.h
           build/src/tensors/dbcsr_tensor.h
@@ -200,9 +200,10 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Download coverage data
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4.1.7
         with:
-          name: coverage-data
+          pattern: coverage-data-*
+          merge-multiple: true
 
       - name: Combine coverage
         run: |
@@ -213,7 +214,7 @@ jobs:
           lcov --summary merged.info
 
       - name: Upload merged HTML report
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: html-report
           path: htmlcov

.pre-commit-config.yaml

@@ -6,7 +6,7 @@ fail_fast: false
 minimum_pre_commit_version: 3.2.0
 repos:
 - repo: https://github.com/astral-sh/ruff-pre-commit
-  rev: 'v0.5.4'
+  rev: 'v0.8.2'
   hooks:
   - id: ruff
     args: [ --fix, --exit-non-zero-on-fix ]
@@ -15,19 +15,19 @@ repos:
         .cp2k/.*|
       )$
 - repo: https://github.com/psf/black
-  rev: 24.4.2
+  rev: 24.10.0
   hooks:
   - id: black
     name: Reformat Python files with the black code formatter
     files: '^.*(/PACKAGE)|(\.py)$'
 - repo: https://github.com/pre-commit/pre-commit-hooks
-  rev: v4.6.0
+  rev: v5.0.0
   hooks:
   - id: check-ast
   - id: check-yaml
   - id: check-symlinks
   - id: trailing-whitespace
-- repo: https://github.com/pseewald/fprettify
+- repo: https://github.com/fortran-lang/fprettify
   rev: v0.3.7
   hooks:
   - id: fprettify
@@ -65,4 +65,5 @@ repos:
     language: python
     files: \.(c|cc|cxx|cpp|cl|frag|glsl|h|hpp|hxx|ih|ispc|ipp|java|js|m|mm|proto|textproto|vert)$
     args: ['-i', '-fallback-style=none', '--style=file']
-    additional_dependencies: ['clang-format']
+    # specify version since clang-format is not stable version-to-version
+    additional_dependencies: ['clang-format~=19.1.0']

CMakeLists.txt

@@ -120,13 +120,7 @@ set_property(CACHE WITH_GPU PROPERTY STRINGS ${SUPPORTED_CUDA_ARCHITECTURES}
 
 option(WITH_CUDA_PROFILING "Enable profiling within CUDA" OFF)
 option(WITH_HIP_PROFILING "Enable profiling within HIP" OFF)
-option(WITH_G2G "Enable GPU aware MPI within CUDA/HIP backends" OFF)
 
-if (WITH_G2G AND ((NOT USE_ACCEL) OR ((NOT USE_ACCEL MATCHES "cuda")
-                                      AND (NOT USE_ACCEL MATCHES "hip"))))
-  message(
-    FATAL_ERROR "GPU aware MPI can only be enabled for HIP/CUDA GPU backends")
-endif ()
 # =================================================================================================
 # LANGUAGES AND TESTING
 enable_language(Fortran)
@@ -274,7 +268,6 @@ if (USE_ACCEL MATCHES "cuda")
   message(STATUS "Kernel parameters: " ${WITH_GPU_PARAMS})
   message(STATUS "GPU architecture number: " ${ACC_ARCH_NUMBER})
   message(STATUS "GPU profiling enabled: " ${WITH_CUDA_PROFILING})
-  message(STATUS "GPU aware MPI enabled: " ${WITH_G2G})
 endif ()
 
 if (USE_ACCEL MATCHES "hip")
@@ -319,7 +312,6 @@ if (USE_ACCEL MATCHES "hip")
   message(STATUS "Kernel parameters: " ${WITH_GPU_PARAMS})
   message(STATUS "GPU architecture number: " ${ACC_ARCH_NUMBER})
   message(STATUS "GPU profiling enabled: " ${WITH_HIP_PROFILING})
-  message(STATUS "GPU aware MPI enabled: " ${WITH_G2G})
 
   # =================================== BLAS on GPU backend
   find_package(hipblas CONFIG REQUIRED HINTS ${ROCM_PATH})

VERSION

@@ -1,8 +1,8 @@
 MAJOR = 2
-MINOR = 7
+MINOR = 8
 PATCH = 0
 # A specific DATE (YYYY-MM-DD) fixes an official release, otherwise
 # it is considered Development version.
-DATE  = 2024-07-29
+DATE  = 2024-12-11
 
 

cmake/CompilerConfiguration.cmake

@@ -51,9 +51,6 @@ if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
   if ((NOT (USE_MPI)) OR (NOT ("${MPI_Fortran_LIBRARY_VERSION_STRING}" MATCHES "Open MPI")))
     set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fsanitize=leak")
   endif ()
-  if (USE_ACCEL MATCHES "hip" AND  hip_VERSION GREATER_EQUAL 6.0.0) # Remove deprecated function error with ROCm v6+
-     set(CMAKE_CXX_FLAGS            "${CMAKE_CXX_FLAGS} -Wno-error=deprecated-declarations")
-  endif	()
 elseif (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
   set(CMAKE_CXX_FLAGS_RELEASE      "-O3 -funroll-loops")
   set(CMAKE_CXX_FLAGS_COVERAGE     "-O0 -g --coverage")

docs/guide/2-user-guide/1-installation/index.md

@@ -70,7 +70,6 @@ make
 -DUSE_ACCEL=<opencl|cuda|hip>
 -DWITH_CUDA_PROFILING=<OFF|ON>
 -DWITH_HIP_PROFILING=<OFF|ON>
--DWITH_G2G=<OFF|ON>
 -DWITH_C_API=<ON|OFF>
 -DWITH_EXAMPLES=<ON|OFF>
 -DWITH_GPU=<P100|K20X|K40|K80|V100|Mi50|Mi100|Mi250>

docs/guide/3-developer-guide/3-programming/1-overview/index.md

@@ -55,4 +55,3 @@ Assumed square matrix with 20x20 matrix with 5x5 blocks and a 2x2 processor grid
 | `__CUDA_PROFILING`  | To turn on Nvidia Tools Extensions. It requires to link `-lnvToolsExt` | Fortran, C, C++ |
 | `__CUDA` | Enable CUDA acceleration | C, C++ |
 | `__HIP`  | Enable HIP acceleration | C, C++ |
-| `__DBCSR_ACC_G2G` | Enable GPU Aware MPI in CUDA and HIP backends | Fortran, C, C++ |

docs/guide/3-developer-guide/3-programming/2-accelerator-backend/2-libsmm_acc/2-parameters.md

@@ -14,9 +14,3 @@ The batched matrix-matrix multiplication kernels are templated on:
 The batched transpose kernels are templated on:
 
 * the characteristic dimensions of the transpose: `m, n`
-
-## Predictive parameters
-
-The input features for the predictive models can be 'raw' parameters (left-most-column in the figure below), or hand-engineered features 'derived' from the raw features (matrix sizes, launch parameters and resource usage estimations).
-
-![libsmm_acc_predictive_modeling_features](../../../../../media/images/libsmm_acc_predictive_modeling_features.png)