diff --git a/.github/workflows/intel.yml b/.github/workflows/intel.yml
index 7a1e97a80e8..814bb2e33ba 100644
--- a/.github/workflows/intel.yml
+++ b/.github/workflows/intel.yml
@@ -21,9 +21,9 @@ jobs:
       fail-fast: false
       matrix:
         config:
-        - {compiler: "dpcpp", build_type: "Release", name: "intel/dpcpp/release/shared", mixed: "ON"}
-        - {compiler: "icpx", build_type: "Release", name: "intel/icpx/release/shared", mixed: "OFF"}
-    name: ${{ matrix.config.name }}
+          - { preset: "ci/sycl/dpcpp/nompi/release/singlemode" }
+          - { preset: "ci/sycl/icpx/nompi/release/singlemode/nomixed" }
+    name: ${{ matrix.config.preset }}
     runs-on: [gpu_intel]
 
     steps:
@@ -36,8 +36,8 @@ jobs:
         module load intel-oneapi-compilers intel-oneapi-dpl intel-oneapi-mkl intel-oneapi-tbb cmake
         mkdir build
         cd build
-        cmake .. -DCMAKE_INSTALL_PREFIX=install_ginkgo -DCMAKE_CXX_FLAGS="-Wpedantic -ffp-model=precise" -DCMAKE_CXX_COMPILER=${{ matrix.config.compiler }} -DCMAKE_BUILD_TYPE=${{ matrix.config.build_type }} -DGINKGO_MIXED_PRECISION=${{ matrix.config.mixed }} -DGINKGO_BUILD_CUDA=OFF -DGINKGO_BUILD_HIP=OFF -DGINKGO_BUILD_MPI=OFF -DGINKGO_DPCPP_SINGLE_MODE=ON
-        make -j8
+        cmake -S .. -B . --preset ${{ matrix.config.preset }} -DCMAKE_INSTALL_PREFIX=install_ginkgo
+        cmake --build . -j8
         ONEAPI_DEVICE_SELECTOR=level_zero:gpu ctest -j10 --output-on-failure
 
     - name: install
diff --git a/.github/workflows/osx.yml b/.github/workflows/osx.yml
index 441be04d842..001297ea5f5 100644
--- a/.github/workflows/osx.yml
+++ b/.github/workflows/osx.yml
@@ -27,9 +27,9 @@ jobs:
       fail-fast: false
       matrix:
         config:
-        - {shared: "ON", build_type: "Debug", name: "omp/debug/shared", "mixed": "OFF"}
-        - {shared: "OFF", build_type: "Release", name: "omp/release/static", "mixed": "ON"}
-    name: ${{ matrix.config.name }}
+          - { preset: "ci/omp/clang/nompi/debug/nomixed" }
+          - { preset: "ci/omp/clang/nompi/release/static" }
+    name: ${{ matrix.config.preset }}
     runs-on: [macos-latest]
 
     steps:
@@ -42,7 +42,7 @@ jobs:
 
     - name: info
       run: |
-        g++ -v
+        clang++ -v
         cmake --version
 
     - name: Debug over SSH (tmate)
@@ -55,12 +55,12 @@ jobs:
         mkdir install
         export INSTALL_PREFIX=`pwd`/install
         cd build
-        cmake .. -DCMAKE_PREFIX_PATH=/opt/homebrew/opt/libomp/ -DCMAKE_CXX_FLAGS=-Wpedantic -DBUILD_SHARED_LIBS=${{ matrix.config.shared }} -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX} -DCMAKE_BUILD_TYPE=${{ matrix.config.build_type }} -DGINKGO_MIXED_PRECISION=${{ matrix.config.mixed }}
-        make -j8
+        cmake -S .. -B . --preset ${{ matrix.config.preset }} -DCMAKE_PREFIX_PATH=/opt/homebrew/opt/libomp/ -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX}
+        cmake --build . -j8
         ctest -j10 --output-on-failure
 
     - name: install
       run: |
         cd build
-        make install
-        make test_install
+        cmake --install .
+        cmake --build . --target test_install
diff --git a/.github/workflows/windows-mingw.yml b/.github/workflows/windows-mingw.yml
index 1c859661562..31adbba8a4d 100644
--- a/.github/workflows/windows-mingw.yml
+++ b/.github/workflows/windows-mingw.yml
@@ -28,8 +28,8 @@ jobs:
       fail-fast: false
       matrix:
         config:
-        - {shared: "OFF", build_type: "Release", name: "omp/release/static", cflags: ""}
-    name: mingw/${{ matrix.config.name }}
+          - { preset: "ci/omp/mingwgcc/nompi/release/static" }
+    name: ${{ matrix.config.preset }}
     runs-on: [windows-latest]
     steps:
     - name: Checkout the latest code (shallow clone)
@@ -50,7 +50,7 @@ jobs:
         set PATH=C:\ProgramData\chocolatey\lib\mingw\tools\install\mingw64\bin;%PATH%
         mkdir build
         cd build
-        cmake -G "MinGW Makefiles" -DBUILD_SHARED_LIBS=${{ matrix.config.shared }} -DCMAKE_BUILD_TYPE=${{ matrix.config.build_type }} -DCMAKE_CXX_FLAGS=${{ matrix.config.cflags }} ..
+        cmake -S .. -B . --preset ${{ matrix.config.preset }}
         cmake --build . -j4
       shell: cmd
 
diff --git a/.github/workflows/windows-msvc-cuda.yml b/.github/workflows/windows-msvc-cuda.yml
index efa637b2bf9..5f7f8db984b 100644
--- a/.github/workflows/windows-msvc-cuda.yml
+++ b/.github/workflows/windows-msvc-cuda.yml
@@ -28,10 +28,13 @@ jobs:
       fail-fast: false
       matrix:
         config:
-        - {version: "latest", name: "cuda-latest/release/shared", "mixed": "ON"}
-    name: msvc/${{ matrix.config.name }} (only compile)
+        - {version: "latest", preset: "ci/cuda/msvc/nompi/release"}
+    name: ${{ matrix.config.preset }} (only compile)
     runs-on: [windows-2019]
 
+    env:
+      CUDA_ARCH: "60"
+
     steps:
     - name: Checkout the latest code (shallow clone)
       uses: actions/checkout@v4
@@ -58,5 +61,5 @@ jobs:
         refreshenv
         mkdir build
         cd build
-        cmake -DGINKGO_BUILD_CUDA=ON -DGINKGO_BUILD_OMP=OFF -DGINKGO_MIXED_PRECISION=${{ matrix.config.mixed }} -DGINKGO_CUDA_ARCHITECTURES=60 ..
+        cmake -S .. -B . --preset ${{ matrix.config.preset }}
         cmake --build . -j4 --config Release
diff --git a/.github/workflows/windows-msvc-ref.yml b/.github/workflows/windows-msvc-ref.yml
index 60a811bb99b..308af1aa1b0 100644
--- a/.github/workflows/windows-msvc-ref.yml
+++ b/.github/workflows/windows-msvc-ref.yml
@@ -30,11 +30,11 @@ jobs:
         config:
         # Debug shared exceeds symbol limit
         # - {shared: "ON", build_type: "Debug", name: "reference/debug/shared"}
-        - {shared: "OFF", build_type: "Release", name: "reference/release/static"}
-        - {shared: "ON", build_type: "Release", name: "reference/release/shared"}
+          - { preset: "ci/ref/msvc/nompi/release/static" }
+          - { preset: "ci/ref/msvc/nompi/release" }
         # Debug static needs too much storage
         # - {shared: "OFF", build_type: "Debug", name: "reference/debug/static"}
-    name: msvc/${{ matrix.config.name }}
+    name: ${{ matrix.config.preset }}
     runs-on: [windows-latest]
     steps:
     - name: Checkout the latest code (shallow clone)
@@ -50,13 +50,13 @@ jobs:
       run: |
         mkdir build
         cd build
-        cmake  -DBUILD_SHARED_LIBS=${{ matrix.config.shared }} -DCMAKE_CXX_FLAGS_DEBUG='/MDd /Zi /Ob1 /O1 /Od /RTC1' -DGINKGO_BUILD_CUDA=OFF -DGINKGO_BUILD_OMP=OFF ..
-        cmake --build . -j4 --config ${{ matrix.config.build_type }}
-        ctest . -C ${{ matrix.config.build_type }} --output-on-failure
+        cmake -S .. -B . --preset ${{ matrix.config.preset }}
+        cmake --build . -j4
+        ctest . --output-on-failure
 
     - name: install
       run: |
         $env:PATH="$env:PATH;C:\Program Files (x86)\Ginkgo\bin"
         cd build
-        cmake --install . --config ${{ matrix.config.build_type }}
-        cmake --build . --target test_install --config ${{ matrix.config.build_type }}
+        cmake --install .
+        cmake --build . --target test_install
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 2f8e3a892a5..05861872d37 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -2,6 +2,7 @@ stages:
   - init-status
   - sync
   - trigger_pipeline
+  - configure
   - build
   - test
   - code_quality
@@ -90,6 +91,27 @@ trigger_pipeline:
         echo "Can not find the corresponding Pull Request"
       fi
 
+configure/cmake-3.16:
+  extends:
+    - .quick_test_condition
+    - .configure_template
+    - .use_gko-rocm45-nompi-gnu8-llvm8
+  variables:
+    BUILD_CUDA: "OFF"
+    BUILD_HIP: "ON"
+    BUILD_SYCL: "OFF"
+    BUILD_MPI: "ON"
+
+configure/cmake-3.18:
+  extends:
+    - .quick_test_condition
+    - .configure_template
+    - .use_gko-cuda102-nompi-gnu8-llvm8-intel2019
+  variables:
+    BUILD_CUDA: "ON"
+    BUILD_HIP: "OFF"
+    BUILD_SYCL: "OFF"
+    BUILD_MPI: "OFF"
 
 # Build jobs
 # Job with example runs.
@@ -101,12 +123,7 @@ build/cuda110/mvapich2/gcc/cuda/debug/shared:
     - .quick_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_MPI: "ON"
-    BUILD_TYPE: "Debug"
-    FAST_TESTS: "ON"
-    NONDEFAULT_STREAM: "ON"
+    CI_CMAKE_PRESET: "ci/cuda+omp/gcc/mpi/debug/fasttest/nondefault_stream"
     CUDA_ARCH: 80
     USE_NAME: "cuda110-mvapich2-gcc-${CI_PIPELINE_ID}"
     KEEP_CONTAINER: "ON"
@@ -134,11 +151,7 @@ build/cuda110/nompi/clang/cuda/release/static:
     - .full_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "OFF"
+    CI_CMAKE_PRESET: "ci/cuda+omp/clang/nompi/release/static"
     CUDA_ARCH: 80
     USE_NAME: "cuda110-nompi-clang-${CI_PIPELINE_ID}"
     KEEP_CONTAINER: "ON"
@@ -157,7 +170,7 @@ test/cuda110/nompi/clang/cuda/release/static:
     SLURM_TIME: "01:30:00"
   dependencies: null
   needs: [ "build/cuda110/nompi/clang/cuda/release/static" ]
-  
+
 
 build/cuda110/nompi/clang/cuda/release/shared:
   extends:
@@ -166,14 +179,8 @@ build/cuda110/nompi/clang/cuda/release/shared:
     - .quick_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    CXX_COMPILER: "clang++"
+    CI_CMAKE_PRESET: "ci/cuda+omp/clang/nompi/release/fasttest/shared"
     CUDA_ARCH: 52
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_TYPE: "Release"
-    FAST_TESTS: "ON"
-    # disable spurious unused argument warning
-    EXTRA_CMAKE_FLAGS: "-DCMAKE_CUDA_FLAGS=-diag-suppress=177"
 
 
 # cuda 11.4 and friends
@@ -184,14 +191,7 @@ build/cuda114/nompi/gcc/cuda/debug/shared:
     - .quick_test_condition
     - .use_gko_cuda114-openmpi-gnu10-llvm12
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_TYPE: "Debug"
-    FAST_TESTS: "ON"
-    # fix gtest issue https://github.com/google/googletest/issues/3514
-    CXX_FLAGS: "-Wno-error=maybe-uninitialized"
-    # disable spurious unused argument warning
-    EXTRA_CMAKE_FLAGS: "-DCMAKE_CUDA_FLAGS=-diag-suppress=177"
+    CI_CMAKE_PRESET: "ci/cuda+omp/gcc/nompi/debug/fasttest"
 
 
 # nvhpc and friends
@@ -202,17 +202,7 @@ build/nvhpc233/cuda120/nompi/nvcpp/release/static:
     - .quick_test_condition
     - .use_gko_nvhpc233-cuda120-openmpi-gnu12-llvm16
   variables:
-    CXX_COMPILER: "nvc++"
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_MPI: "OFF"
-    BUILD_TYPE: "Release"
-    FAST_TESTS: "ON"
-    BUILD_SHARED_LIBS: "OFF"
-    # disable common gflags warnings
-    CXX_FLAGS: "--diag_suppress=useless_using_declaration,declared_but_not_referenced"
-    # disable spurious unused argument warning
-    EXTRA_CMAKE_FLAGS: "-DCMAKE_CUDA_FLAGS=-diag-suppress=177"
+    CI_CMAKE_PRESET: "ci/cuda+omp/nvhpc/nompi/release/static/fasttest"
 
 build/nvhpc227/cuda117/nompi/nvcpp/debug/shared:
   extends:
@@ -221,16 +211,7 @@ build/nvhpc227/cuda117/nompi/nvcpp/debug/shared:
     - .quick_test_condition
     - .use_gko_nvhpc227-cuda117-openmpi-gnu11-llvm14
   variables:
-    CXX_COMPILER: "nvc++"
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_MPI: "OFF"
-    BUILD_TYPE: "Debug"
-    FAST_TESTS: "ON"
-    # disable common gflags warnings
-    CXX_FLAGS: "--diag_suppress=useless_using_declaration,declared_but_not_referenced"
-    # disable spurious unused argument warning
-    EXTRA_CMAKE_FLAGS: "-DCMAKE_CUDA_FLAGS=-diag-suppress=177"
+    CI_CMAKE_PRESET: "ci/cuda+omp/nvhpc/nompi/debug/fasttest"
 
 # ROCm 4.5 and friends
 build/amd/nompi/gcc/rocm45/release/shared:
@@ -240,11 +221,7 @@ build/amd/nompi/gcc/rocm45/release/shared:
     - .quick_test_condition
     - .use_gko-rocm45-nompi-gnu8-llvm8
   variables:
-    BUILD_OMP: "ON"
-    BUILD_HIP: "ON"
-    NONDEFAULT_STREAM: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_HWLOC: "OFF"
+    CI_CMAKE_PRESET: "ci/hip+omp/gcc/nompi/release/fasttest/nondefault_stream"
 
 build/amd/nompi/clang/rocm45/release/static:
   extends:
@@ -253,11 +230,7 @@ build/amd/nompi/clang/rocm45/release/static:
     - .full_test_condition
     - .use_gko-rocm45-nompi-gnu8-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_HIP: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "OFF"
+    CI_CMAKE_PRESET: "ci/hip+omp/clang/nompi/release/static"
 
 build/amd/nompi/clang/rocm45/debug/shared:
   extends:
@@ -266,10 +239,7 @@ build/amd/nompi/clang/rocm45/debug/shared:
     - .quick_test_condition
     - .use_gko-rocm45-nompi-gnu8-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_HIP: "ON"
-    BUILD_TYPE: "Debug"
+    CI_CMAKE_PRESET: "ci/hip+omp/clang/nompi/debug"
 
 # ROCm 5.1.4 and friends
 build/amd/nompi/gcc/rocm514/debug/static:
@@ -279,10 +249,7 @@ build/amd/nompi/gcc/rocm514/debug/static:
     - .full_test_condition
     - .use_gko-rocm514-nompi-gnu11-llvm11
   variables:
-    BUILD_OMP: "ON"
-    BUILD_HIP: "ON"
-    BUILD_TYPE: "Debug"
-    BUILD_SHARED_LIBS: "OFF"
+    CI_CMAKE_PRESET: "ci/hip+omp/gcc/nompi/debug/static"
 
 build/amd/nompi/clang/rocm514/release/shared:
   extends:
@@ -291,10 +258,7 @@ build/amd/nompi/clang/rocm514/release/shared:
     - .quick_test_condition
     - .use_gko-rocm514-nompi-gnu11-llvm11
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_HIP: "ON"
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/hip+omp/clang/nompi/release"
 
 # without omp
 build/amd/nompi/gcc/rocm514_wo_omp/release/shared:
@@ -304,10 +268,7 @@ build/amd/nompi/gcc/rocm514_wo_omp/release/shared:
     - .full_test_condition
     - .use_gko-rocm514-nompi-gnu11-llvm11
   variables:
-    BUILD_OMP: "OFF"
-    BUILD_MPI: "OFF"
-    BUILD_HIP: "ON"
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/hip/gcc/nompi/release"
 
 # no cuda but latest gcc and clang
 build/nocuda/nompi/gcc/core/debug/static:
@@ -317,10 +278,7 @@ build/nocuda/nompi/gcc/core/debug/static:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    BUILD_TYPE: "Debug"
-    FAST_TESTS: "ON"
-    BUILD_SHARED_LIBS: "OFF"
-    BUILD_HWLOC: "OFF"
+    CI_CMAKE_PRESET: "ci/ref/gcc/nompi/debug/static/fasttest"
 
 build/nocuda/nompi/clang/core/release/shared:
   extends:
@@ -329,8 +287,7 @@ build/nocuda/nompi/clang/core/release/shared:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/ref/clang/nompi/release"
 
 build/nocuda/nompi/gcc/omp/release/shared:
   extends:
@@ -339,8 +296,7 @@ build/nocuda/nompi/gcc/omp/release/shared:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    BUILD_OMP: "ON"
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/omp/gcc/nompi/release"
 
 build/nocuda/openmpi/clang/omp/debug/static:
   extends:
@@ -349,13 +305,7 @@ build/nocuda/openmpi/clang/omp/debug/static:
     - .full_test_condition
     - .use_gko-nocuda-openmpi-gnu9-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    MPI_AS_ROOT: "ON"
-    BUILD_MPI: "ON"
-    BUILD_TYPE: "Debug"
-    FAST_TESTS: "ON"
-    BUILD_SHARED_LIBS: "OFF"
+    CI_CMAKE_PRESET: "ci/omp/clang/mpi/debug/static/fasttest"
 
 build/nocuda/openmpi/clang/omp/glibcxx-debug-release/shared:
   extends:
@@ -364,13 +314,7 @@ build/nocuda/openmpi/clang/omp/glibcxx-debug-release/shared:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    MPI_AS_ROOT: "ON"
-    BUILD_MPI: "ON"
-    CXX_FLAGS: "-Wpedantic -D_GLIBCXX_DEBUG=1"
-    # The tests are prohibitively slow in Debug
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/omp/clang/mpi/glibcxx+debug/release"
 
 # nocuda with old compiler
 build/nocuda/nompi/gcc/omp/release/static:
@@ -380,9 +324,7 @@ build/nocuda/nompi/gcc/omp/release/static:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    BUILD_OMP: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "OFF"
+    CI_CMAKE_PRESET: "ci/omp/gcc/nompi/release/static"
 
 build/nocuda-nomixed/nompi/clang/omp/release/static:
   extends:
@@ -391,11 +333,7 @@ build/nocuda-nomixed/nompi/clang/omp/release/static:
     - .full_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "OFF"
-    MIXED_PRECISION: "OFF"
+    CI_CMAKE_PRESET: "ci/omp/clang/nompi/release/static/nomixed"
 
 build/nocuda-nomixed/openmpi/gcc/omp/release/shared:
   extends:
@@ -404,11 +342,7 @@ build/nocuda-nomixed/openmpi/gcc/omp/release/shared:
     - .quick_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    BUILD_MPI: "ON"
-    MPI_AS_ROOT: "ON"
-    BUILD_OMP: "ON"
-    BUILD_TYPE: "Release"
-    MIXED_PRECISION: "OFF"
+    CI_CMAKE_PRESET: "ci/omp/gcc/mpi/release/nomixed"
 
 build/nocuda-nomixed/nompi/clang/omp/debug/static:
   extends:
@@ -417,11 +351,7 @@ build/nocuda-nomixed/nompi/clang/omp/debug/static:
     - .full_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    CXX_COMPILER: "clang++"
-    BUILD_OMP: "ON"
-    BUILD_TYPE: "Debug"
-    BUILD_SHARED_LIBS: "OFF"
-    MIXED_PRECISION: "OFF"
+    CI_CMAKE_PRESET: "ci/omp/clang/nompi/debug/static/nomixed"
 
 # spack oneapi 2023.1
 build/icpx20231/igpu/release/shared:
@@ -431,14 +361,8 @@ build/icpx20231/igpu/release/shared:
     - .quick_test_condition
     - .use_gko-oneapi20231-igpu
   variables:
-    CXX_COMPILER: "icpx"
-    CXX_FLAGS: "-Wpedantic -ffp-model=precise"
-    BUILD_SYCL: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "ON"
-    DPCPP_SINGLE_MODE: "ON"
+    CI_CMAKE_PRESET: "ci/sycl/icpx/nompi/release/shared/singlemode"
     ONEAPI_DEVICE_SELECTOR: "opencl:gpu"
-    BUILD_HWLOC: "OFF"
 
 # TODO: Enable when debug shared library size issues are fixed
 # build/dpcpp/level_zero_igpu/debug/shared:
@@ -464,14 +388,8 @@ build/icpx/igpu/release/static:
     - .full_test_condition
     - .use_gko-oneapi-igpu
   variables:
-    CXX_COMPILER: "dpcpp"
-    CXX_FLAGS: "-Wpedantic -ffp-model=precise"
-    BUILD_SYCL: "ON"
-    BUILD_TYPE: "Release"
-    BUILD_SHARED_LIBS: "OF"
-    DPCPP_SINGLE_MODE: "ON"
+    CI_CMAKE_PRESET: "ci/sycl/dpcpp/nompi/release/static/singlemode"
     ONEAPI_DEVICE_SELECTOR: "*:gpu"
-    BUILD_HWLOC: "OFF"
 
 # windows jobs
 # Note that this is using Powershell, not bash
@@ -484,7 +402,7 @@ build/windows-cuda/release/shared:
     - if (Test-Path install) { rm -r -fo install }
     - mkdir build
     - mkdir install
-    - cmake -B build -DBUILD_SHARED_LIBS=ON -DGINKGO_BUILD_CUDA=ON "-DCMAKE_INSTALL_PREFIX=$pwd\install" .
+    - cmake -B build --preset "ci/cuda/msvc/nompi/release/shared" "-DCMAKE_INSTALL_PREFIX=$pwd\install" .
     - cmake --build build --config Release -j16
 # we disable these tests until the triangular solver issues are resolved
 #   - ctest --test-dir build -C Release --no-tests=error --output-on-failure
@@ -503,7 +421,7 @@ build/windows/release/shared:
     - if (Test-Path install) { rm -r -fo install }
     - mkdir build
     - mkdir install
-    - cmake -B build -DBUILD_SHARED_LIBS=ON -DGINKGO_BUILD_CUDA=OFF "-DCMAKE_INSTALL_PREFIX=$pwd\install" .
+    - cmake -B build --preset "ci/ref/msvc/nompi/release" "-DCMAKE_INSTALL_PREFIX=$pwd\install" .
     - cmake --build build --config Release -j16
     - ctest --test-dir build -C Release --no-tests=error --output-on-failure -j16
     - $env:PATH+=";$pwd/install/bin"
@@ -521,9 +439,7 @@ warnings:
     - .full_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    CXX_FLAGS: "-Werror=pedantic -pedantic-errors"
+    CI_CMAKE_PRESET: "ci/warnings"
   allow_failure: yes
 
 # Ensure kernel modules do not depend on core
@@ -535,9 +451,7 @@ no-circular-deps:
     - .quick_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    EXTRA_CMAKE_FLAGS: '-DGINKGO_CHECK_CIRCULAR_DEPS=on'
+    CI_CMAKE_PRESET: "ci/no_circular_deps"
   allow_failure: no
 
 # Ensure Ginkgo builds from a subdirectory
@@ -549,8 +463,7 @@ subdir-build:
     - .full_test_condition
     - .use_gko-nocuda-nompi-gnu9-llvm8
   variables:
-    BUILD_OMP: "ON"
-    CI_PROJECT_PATH_SUFFIX: "/test_subdir"
+    CI_CMAKE_PRESET: "ci/subdir_build"
   allow_failure: no
 
 # Run clang-tidy and iwyu
@@ -562,9 +475,7 @@ clang-tidy:
     - .full_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    EXTRA_CMAKE_FLAGS: '-DGINKGO_WITH_CLANG_TIDY=ON'
+    CI_CMAKE_PRESET: "ci/clang_tidy"
   allow_failure: yes
 
 iwyu:
@@ -575,9 +486,7 @@ iwyu:
     - .full_test_condition
     - .use_gko-cuda110-mvapich2-gnu9-llvm9-intel2020
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    EXTRA_CMAKE_FLAGS: '-DGINKGO_WITH_IWYU=ON'
+    CI_CMAKE_PRESET: "ci/iwyu"
   allow_failure: yes
 
 # Code analysis, coverage and reporting tool
@@ -769,12 +678,7 @@ benchmark-cuda-spmv-build:
     - .benchmark-spmv-cuda-rules
   stage: benchmark-build
   variables:
-    BUILD_OMP: "ON"
-    BUILD_CUDA: "ON"
-    BUILD_MPI: "ON"
-    MPI_AS_ROOT: "ON"
-    BUILD_HIP: "OFF"
-    BUILD_TYPE: "Release"
+    CI_CMAKE_PRESET: "ci/cuda+omp/gcc/mpi/release/shared"
     CUDA_ARCH: 80
     USE_NAME: "benchmark-cuda-spmv-${CI_PIPELINE_ID}"
     KEEP_CONTAINER: "ON"
diff --git a/.gitlab/scripts.yml b/.gitlab/scripts.yml
index dda127ff535..cd00478ee3e 100644
--- a/.gitlab/scripts.yml
+++ b/.gitlab/scripts.yml
@@ -15,6 +15,43 @@
     - git config --global user.name "${BOT_USER}"
     - git config --global user.email "${BOT_EMAIL}"
 
+.environment_script: &environment_script
+  - |
+    # extracts the value of the given CMake cache variable
+    function get_cache_val() {
+        cmake -S ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX} -B . -L -N | grep $1 | sed -e 's/.*=//'
+    }
+  - if [ -n "${CUDA_ARCH}" ]; then
+    export CUDA_ARCH_STR=-DGINKGO_CUDA_ARCHITECTURES=${CUDA_ARCH};
+    fi
+  - |
+    export OMPI_ALLOW_RUN_AS_ROOT=1;
+    export OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1;
+  - export CC=${C_COMPILER} CXX=${CXX_COMPILER} CUDAHOSTCXX=${CXX_COMPILER} CUDACXX=${CUDA_COMPILER}
+
+
+.configure_template:
+  stage: configure
+  extends:
+    - .before_script_template
+  script:
+    - mkdir -p ${CI_JOB_NAME} && cd ${CI_JOB_NAME}
+    - *environment_script
+    - |
+      cmake -S ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX} -B . \
+      -DGINKGO_BUILD_REFERENCE=ON -DGINKGO_BUILD_OMP=ON \
+      -DGINKGO_BUILD_MPI=${BUILD_MPI} -DGINKGO_BUILD_CUDA=${BUILD_CUDA} \
+      -DGINKGO_BUILD_HIP=${BUILD_HIP} -DGINKGO_BUILD_SYCL=${BUILD_SYCL} \
+      -DGINKGO_MIXED_PRECISION=ON -DGINKGO_FORCE_GPU_AWARE_MPI=ON \
+      -DGINKGO_BUILD_TESTS=ON -DGINKGO_FAST_TESTS=ON -DGINKGO_BUILD_EXTLIB_EXAMPLE=OFF \
+      -DGINKGO_BUILD_BENCHMARKS=ON -DGINKGO_BENCHMARK_ENABLE_TUNING=ON \
+      -DGINKGO_DEVEL_TOOLS=OFF -DGINKGO_WITH_CLANG_TIDY=ON -DGINKGO_WITH_IWYU=ON \
+      -DGINKGO_CHECK_CIRCULAR_DEPS=ON -DGINKGO_WITH_CCACHE=ON \
+      -DGINKGO_BUILD_PAPI_SDE=OFF -DGINKGO_BUILD_HWLOC=ON \
+      -DGINKGO_BUILD_TESTS=ON -DGINKGO_BUILD_EXAMPLES=ON \
+      -DGINKGO_BUILD_DOC=OFF
+  dependencies: []
+
 .build_template:
   stage: build
   extends:
@@ -23,35 +60,8 @@
     - mkdir -p ${CI_JOB_NAME} && cd ${CI_JOB_NAME}
     - mkdir install_prefix
     - export INSTALL_PREFIX=`pwd`/install_prefix
-    - if [ -n "${CUDA_ARCH}" ]; then
-      export CUDA_ARCH_STR=-DGINKGO_CUDA_ARCHITECTURES=${CUDA_ARCH};
-      fi
-    - if [[ "${MPI_AS_ROOT}" == "ON" ]];then
-      export OMPI_ALLOW_RUN_AS_ROOT=1;
-      export OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1;
-      fi
-    - if [[ "${BUILD_MPI}" == "ON" ]]; then
-      MPI_STR=-DGINKGO_MPI_EXEC_SUFFIX=${MPI_SUFFIX};
-      fi
-    - export CXX=${CXX_COMPILER} CUDAHOSTCXX=${CXX_COMPILER} CUDACXX=${CUDA_COMPILER}
-    - cmake ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX}
-        -GNinja
-        -DCMAKE_BUILD_TYPE=${BUILD_TYPE}
-        -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX}
-        -DCMAKE_CXX_FLAGS="${CXX_FLAGS}" -DBUILD_SHARED_LIBS=${BUILD_SHARED_LIBS}
-        ${EXTRA_CMAKE_FLAGS} ${CUDA_ARCH_STR}
-        -DGINKGO_DEVEL_TOOLS=OFF -DGINKGO_BUILD_REFERENCE=${BUILD_REFERENCE}
-        -DGINKGO_BUILD_OMP=${BUILD_OMP} -DGINKGO_BUILD_CUDA=${BUILD_CUDA}
-        -DGINKGO_BUILD_HIP=${BUILD_HIP} -DGINKGO_BUILD_SYCL=${BUILD_SYCL}
-        -DGINKGO_BUILD_MPI=${BUILD_MPI} ${MPI_STR}
-        -DGINKGO_BUILD_HWLOC=${BUILD_HWLOC}
-        -DGINKGO_BUILD_PAPI_SDE=${BUILD_PAPI_SDE}
-        -DGINKGO_BUILD_TESTS=ON -DGINKGO_BUILD_EXAMPLES=ON
-        -DGINKGO_FAST_TESTS=${FAST_TESTS}
-        -DGINKGO_TEST_NONDEFAULT_STREAM=${NONDEFAULT_STREAM}
-        -DGINKGO_MIXED_PRECISION=${MIXED_PRECISION}
-        -DGINKGO_CONFIG_LOG_DETAILED=${CONFIG_LOG}
-        -DGINKGO_DPCPP_SINGLE_MODE=${DPCPP_SINGLE_MODE}
+    - *environment_script
+    - cmake -S ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX} -B . --preset ${CI_CMAKE_PRESET} -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX}
     - ninja -j${NUM_CORES} -l${CI_LOAD_LIMIT} install
     - awk '!/^#/ { print ($2 - $1)/1000 " " $4 }' .ninja_log | sort -nr
   dependencies: []
@@ -64,36 +74,8 @@
     - mkdir -p ${CI_JOB_NAME} && cd ${CI_JOB_NAME}
     - mkdir install_prefix
     - export INSTALL_PREFIX=`pwd`/install_prefix
-    - if [ -n "${CUDA_ARCH}" ]; then
-      export CUDA_ARCH_STR=-DGINKGO_CUDA_ARCHITECTURES=${CUDA_ARCH};
-      fi
-    - if [ -n "${SYCL_DEVICE_TYPE}" ]; then export SYCL_DEVICE_TYPE; fi
-    - if [ -n "${SYCL_DEVICE_FILTER}" ]; then export SYCL_DEVICE_FILTER; fi
-    - if [ -n "${ONEAPI_DEVICE_SELECTOR}" ]; then export ONEAPI_DEVICE_SELECTOR; fi
-    - if [[ "${MPI_AS_ROOT}" == "ON" ]];then
-      export OMPI_ALLOW_RUN_AS_ROOT=1;
-      export OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1;
-      fi
-    - if [[ "${BUILD_MPI}" == "ON" ]]; then
-      MPI_STR=-DGINKGO_MPI_EXEC_SUFFIX=${MPI_SUFFIX};
-      fi
-    - export CXX=${CXX_COMPILER} CUDAHOSTCXX=${CXX_COMPILER} CUDACXX=${CUDA_COMPILER}
-    - cmake ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX}
-        -GNinja -DCMAKE_BUILD_TYPE=${BUILD_TYPE}
-        -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX}
-        -DCMAKE_CXX_FLAGS="${CXX_FLAGS}" -DBUILD_SHARED_LIBS=${BUILD_SHARED_LIBS}
-        ${EXTRA_CMAKE_FLAGS} ${CUDA_ARCH_STR}
-        -DGINKGO_DEVEL_TOOLS=OFF -DGINKGO_BUILD_REFERENCE=${BUILD_REFERENCE}
-        -DGINKGO_BUILD_OMP=${BUILD_OMP} -DGINKGO_BUILD_CUDA=${BUILD_CUDA}
-        -DGINKGO_BUILD_HIP=${BUILD_HIP} -DGINKGO_BUILD_SYCL=${BUILD_SYCL}
-        -DGINKGO_BUILD_MPI=${BUILD_MPI} ${MPI_STR}
-        -DGINKGO_BUILD_HWLOC=${BUILD_HWLOC}
-        -DGINKGO_BUILD_PAPI_SDE=${BUILD_PAPI_SDE}
-        -DGINKGO_BUILD_TESTS=ON -DGINKGO_BUILD_EXAMPLES=ON
-        -DGINKGO_FAST_TESTS=${FAST_TESTS}
-        -DGINKGO_MIXED_PRECISION=${MIXED_PRECISION}
-        -DGINKGO_CONFIG_LOG_DETAILED=${CONFIG_LOG}
-        -DGINKGO_DPCPP_SINGLE_MODE=${DPCPP_SINGLE_MODE}
+    - *environment_script
+    - cmake -S ${CI_PROJECT_DIR}${CI_PROJECT_DIR_SUFFIX} -B . --preset ${CI_CMAKE_PRESET} -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX}
     - ninja -j${NUM_CORES} -l${CI_LOAD_LIMIT} install
     - awk '!/^#/ { print ($2 - $1)/1000 " " $4 }' .ninja_log | sort -nr
     - |
@@ -106,7 +88,8 @@
     - popd
     - if [ -n "${SYCL_DEVICE_TYPE}" ]; then unset SYCL_DEVICE_TYPE; fi
     - if [ -n "${SYCL_DEVICE_FILTER}" ]; then unset SYCL_DEVICE_FILTER; fi
-    - if [[ "${BUILD_SHARED_LIBS}" == "ON" || "${BUILD_HIP}" != "ON" ]]; then
+    - if [ "$(get_cache_val GINKGO_EXPORT_BUILD_DIR)" == "ON" ]; then ninja test_exportbuild; fi
+    - if [[ "$(get_cache_val BUILD_SHARED_LIBS)" == "ON" || "$(get_cache_val GINKGO_BUILD_HIP)" != "ON" ]]; then
       PKG_CONFIG_PATH=${INSTALL_PREFIX}/lib/pkgconfig:$PKG_CONFIG_PATH LD_LIBRARY_PATH=${INSTALL_PREFIX}/lib:$LD_LIBRARY_PATH ninja test_pkgconfig;
       fi
   dependencies: []
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 10f76ac9a6c..27eee6d2abb 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -44,8 +44,9 @@ option(GINKGO_CHECK_CIRCULAR_DEPS
 option(GINKGO_CONFIG_LOG_DETAILED
     "Enable printing of detailed configuration log to screen in addition to the writing of files," OFF)
 option(GINKGO_BENCHMARK_ENABLE_TUNING
-    "Enable tuning variables in the benchmarks. For specific use cases, manual code changes could be required."
+    "Enable tuning variables in the benchmarks. For specific use cases, manual code changes could be required. WARNING: It is not possible to install Ginkgo with this option enabled."
     OFF)
+mark_as_advanced(GINKGO_BENCHMARK_ENABLE_TUNING)
 set(GINKGO_VERBOSE_LEVEL "1" CACHE STRING
     "Verbosity level. Put 0 to turn off. 1 activates a few important messages.")
 set(GINKGO_CUDA_ARCHITECTURES "Auto" CACHE STRING
@@ -70,6 +71,11 @@ option(GINKGO_EXTENSION_KOKKOS_CHECK_TYPE_ALIGNMENT "Enables mapping to Kokkos t
 gko_rename_cache(GINKGO_COMPILER_FLAGS CMAKE_CXX_FLAGS BOOL "Flags used by the CXX compiler during all build types.")
 gko_rename_cache(GINKGO_CUDA_COMPILER_FLAGS CMAKE_CUDA_FLAGS BOOL "Flags used by the CUDA compiler during all build types.")
 
+if(NOT DEFINED BUILD_SHARED_LIBS)
+    message(STATUS "BUILD_SHARED_LIBS is not set, so enabling it by default.")
+    set(BUILD_SHARED_LIBS ON)
+endif()
+
 # load executor-specific configuration
 if(GINKGO_BUILD_CUDA)
     include(cmake/cuda.cmake)
@@ -120,13 +126,6 @@ if(GINKGO_WITH_CCACHE)
     endif()
 endif()
 
-if(GINKGO_BENCHMARK_ENABLE_TUNING)
-    # In this state, the tests and examples cannot be compiled without extra
-    # complexity/intrusiveness, so we simply disable them.
-    set(GINKGO_BUILD_TESTS OFF)
-    set(GINKGO_BUILD_EXAMPLES OFF)
-endif()
-
 if(GINKGO_BUILD_TESTS)
     message(STATUS "GINKGO_BUILD_TESTS is ON, enabling GINKGO_BUILD_REFERENCE")
     set(GINKGO_BUILD_REFERENCE ON CACHE BOOL "Compile reference CPU kernels" FORCE)
diff --git a/CMakePresets.json b/CMakePresets.json
new file mode 100644
index 00000000000..e9a5aee3fc0
--- /dev/null
+++ b/CMakePresets.json
@@ -0,0 +1,76 @@
+{
+    "version": 4,
+    "include": ["cmake/presets/ci.json"],
+    "configurePresets": [
+        {
+            "name": "base",
+            "hidden": true,
+            "binaryDir": "${sourceDir}/build-${presetName}"
+        },
+        {
+            "name": "base-dev",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_BUILD_TYPE": "Debug",
+                "GINKGO_DEVEL_TOOLS": "ON",
+                "CMAKE_EXPORT_COMPILE_COMMANDS": "ON",
+                "CMAKE_COLOR_DIAGNOSTICS": "ON",
+                "CMAKE_LINK_DEPENDS_NO_SHARED": "ON"
+            },
+            "warnings": {
+                "dev": true,
+                "deprecated": true
+            }
+        },
+        {
+            "name": "base-dev-gcc-like",
+            "hidden": true,
+            "generator": "Ninja",
+            "cacheVariables": {
+                "CMAKE_CXX_FLAGS": "$env{CXXFLAGS} -g -fno-omit-frame-pointer -Wpedantic -Wall -Wextra -Wconversion -Wunused"
+            },
+            "condition": {
+                "type": "notEquals",
+                "lhs": "${hostSystemName}",
+                "rhs": "Windows"
+            }
+        },
+        {
+            "name": "release",
+            "inherits": "base",
+            "cacheVariables": {
+                "CMAKE_BUILD_TYPE": "Release"
+            },
+            "warnings": {
+                "dev": false
+            }
+        },
+        {
+            "name": "develop-gcc",
+            "inherits": ["base", "base-dev", "base-dev-gcc-like"],
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "g++"
+            }
+        },
+        {
+            "name": "develop-clang",
+            "inherits": ["base", "base-dev", "base-dev-gcc-like"],
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "clang++"
+            }
+        },
+        {
+            "name": "develop-msvc",
+            "inherits": ["base", "base-dev"],
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "msvc",
+                "CMAKE_CXX_FLAGS": "$env{CXXFLAGS} /permissive- /Oy-"
+            },
+            "condition": {
+                "type": "equals",
+                "lhs": "${hostSystemName}",
+                "rhs": "Windows"
+            }
+        }
+    ]
+}
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 66865585dca..4547986f3fe 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -568,9 +568,14 @@ files.
 
 ### Warnings
 
-By default, the `-DCMAKE_CXX_FLAGS` should be set to `-Wpedantic` to emit
-pedantic warnings by default. The CI system additionally also has a step
-where it compiles for pedantic warnings to be errors.
+By default, the `-Wpedantic` (or `/permissive-` on MSVC) is added to `CMAKE_CXX_FLAGS`
+and hence pedantic warnings are emitted by default. Some of these warnings 
+are false positives and a complete list of the resolved warnings and their solutions
+is listed in [Issue 174](https://github.com/ginkgo-project/ginkgo/issues/174).
+Specifically, when macros are being used, we have the issue of having `extra ;`
+warnings, which is resolved by adding a `static_assert()`. The CI system
+additionally also has a step where it compiles for pedantic warnings to be
+errors.
 
 ### Avoiding circular dependencies
 
diff --git a/INSTALL.md b/INSTALL.md
index 9719bdfb920..bab79203e93 100644
--- a/INSTALL.md
+++ b/INSTALL.md
@@ -89,7 +89,7 @@ Ginkgo adds the following additional switches to control what is being built:
     `COVERAGE`, `ASAN` (AddressSanitizer), `LSAN` (LeakSanitizer), `TSAN`
     (ThreadSanitizer) and `UBSAN` (undefined behavior sanitizer) types.
 *   `-DBUILD_SHARED_LIBS={ON, OFF}` builds ginkgo as shared libraries (`OFF`)
-    or as dynamic libraries (`ON`), default is `ON`.
+    or as dynamic libraries (`ON`), default is `ON`, see https://cmake.org/cmake/help/latest/variable/BUILD_SHARED_LIBS.html.
 *   `-DGINKGO_JACOBI_FULL_OPTIMIZATIONS={ON, OFF}` use all the optimizations
     for the CUDA Jacobi algorithm. `OFF` by default. Setting this option to `ON`
     may lead to very slow compile time (>20 minutes) for the
diff --git a/benchmark/CMakeLists.txt b/benchmark/CMakeLists.txt
index de6e74d464c..78fdfcaf219 100644
--- a/benchmark/CMakeLists.txt
+++ b/benchmark/CMakeLists.txt
@@ -6,7 +6,7 @@ endif()
 
 function(ginkgo_benchmark_add_tuning_maybe name)
     if(GINKGO_BENCHMARK_ENABLE_TUNING)
-        target_sources(${name} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/../utils/tuning_variables.cpp)
+        target_sources(${name} PRIVATE ${Ginkgo_SOURCE_DIR}/benchmark/utils/tuning_variables.cpp)
     endif()
 endfunction()
 
diff --git a/benchmark/conversion/conversion.cpp b/benchmark/conversion/conversion.cpp
index 59c052b11e7..c7de9a66506 100644
--- a/benchmark/conversion/conversion.cpp
+++ b/benchmark/conversion/conversion.cpp
@@ -23,11 +23,6 @@
 #include "benchmark/utils/types.hpp"
 
 
-#ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-#include "benchmark/utils/tuning_variables.hpp"
-#endif  // GINKGO_BENCHMARK_ENABLE_TUNING
-
-
 using Generator = DefaultSystemGenerator<>;
 
 
diff --git a/benchmark/matrix_generator/matrix_generator.cpp b/benchmark/matrix_generator/matrix_generator.cpp
index dc1cb3cc08a..895eec2325e 100644
--- a/benchmark/matrix_generator/matrix_generator.cpp
+++ b/benchmark/matrix_generator/matrix_generator.cpp
@@ -13,11 +13,6 @@
 #include "benchmark/utils/types.hpp"
 
 
-#ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-#include "benchmark/utils/tuning_variables.hpp"
-#endif  // GINKGO_BENCHMARK_ENABLE_TUNING
-
-
 namespace {
 std::string input_format =
     "  [\n"
diff --git a/benchmark/matrix_statistics/matrix_statistics.cpp b/benchmark/matrix_statistics/matrix_statistics.cpp
index 8eb847f42f2..7ddf408d565 100644
--- a/benchmark/matrix_statistics/matrix_statistics.cpp
+++ b/benchmark/matrix_statistics/matrix_statistics.cpp
@@ -16,11 +16,6 @@
 #include "benchmark/utils/types.hpp"
 
 
-#ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-#include "benchmark/utils/tuning_variables.hpp"
-#endif  // GINKGO_BENCHMARK_ENABLE_TUNING
-
-
 // See en.wikipedia.org/wiki/Five-number_summary
 // Quartile computation uses Method 3 from en.wikipedia.org/wiki/Quartile
 void compute_summary(const std::vector<gko::size_type>& dist, json& out)
diff --git a/benchmark/preconditioner/preconditioner.cpp b/benchmark/preconditioner/preconditioner.cpp
index b9dfe1d8369..a1283e50535 100644
--- a/benchmark/preconditioner/preconditioner.cpp
+++ b/benchmark/preconditioner/preconditioner.cpp
@@ -22,11 +22,6 @@
 #include "benchmark/utils/types.hpp"
 
 
-#ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-#include "benchmark/utils/tuning_variables.hpp"
-#endif  // GINKGO_BENCHMARK_ENABLE_TUNING
-
-
 // preconditioner generation and application
 std::string encode_parameters(const char* precond_name)
 {
diff --git a/benchmark/solver/solver_common.hpp b/benchmark/solver/solver_common.hpp
index 541a9f66287..1e77e13234b 100644
--- a/benchmark/solver/solver_common.hpp
+++ b/benchmark/solver/solver_common.hpp
@@ -16,11 +16,6 @@
 #include "benchmark/utils/runner.hpp"
 
 
-#ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-#include "benchmark/utils/tuning_variables.hpp"
-#endif  // GINKGO_BENCHMARK_ENABLE_TUNING
-
-
 // Command-line arguments
 DEFINE_uint32(max_iters, 1000,
               "Maximal number of iterations the solver will be run for");
diff --git a/benchmark/spmv/spmv_common.hpp b/benchmark/spmv/spmv_common.hpp
index 4ac777479b2..0c58cf30b7b 100644
--- a/benchmark/spmv/spmv_common.hpp
+++ b/benchmark/spmv/spmv_common.hpp
@@ -135,7 +135,6 @@ struct SpmvBenchmark : Benchmark<spmv_benchmark_state<Generator>> {
 
         // tuning run
 #ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-        auto& format_case = spmv_case[format_name];
         format_case["tuning"] = json::object();
         auto& tuning_case = format_case["tuning"];
         tuning_case["time"] = json::array();
diff --git a/cmake/build_helpers.cmake b/cmake/build_helpers.cmake
index 0985f089382..eb295719758 100644
--- a/cmake/build_helpers.cmake
+++ b/cmake/build_helpers.cmake
@@ -115,26 +115,6 @@ macro(ginkgo_modify_flags name)
     string(REPLACE "\"" "\\\"" ${name}_MODIFY "${${name}}")
 endmacro()
 
-# Extract the clang version from a clang executable path
-function(ginkgo_extract_clang_version CLANG_COMPILER GINKGO_CLANG_VERSION)
-    set(CLANG_VERSION_PROG "#include <cstdio>\n"
-        "int main() {printf(\"%d.%d.%d\", __clang_major__, __clang_minor__, __clang_patchlevel__)\;"
-        "return 0\;}")
-    file(WRITE "${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver.cpp" ${CLANG_VERSION_PROG})
-    execute_process(COMMAND ${CLANG_COMPILER} ${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver.cpp
-        -o ${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver
-        ERROR_VARIABLE CLANG_EXTRACT_VER_ERROR)
-    execute_process(COMMAND ${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver
-        OUTPUT_VARIABLE FOUND_CLANG_VERSION
-        OUTPUT_STRIP_TRAILING_WHITESPACE
-        ERROR_STRIP_TRAILING_WHITESPACE
-        )
-
-    set (${GINKGO_CLANG_VERSION} "${FOUND_CLANG_VERSION}" PARENT_SCOPE)
-    file(REMOVE ${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver.cpp)
-    file(REMOVE ${CMAKE_CURRENT_BINARY_DIR}/extract_clang_ver)
-endfunction()
-
 # Extract the DPC++ version
 function(ginkgo_extract_dpcpp_version DPCPP_COMPILER GINKGO_DPCPP_VERSION MACRO_VAR)
     set(DPCPP_VERSION_PROG "#include <CL/sycl.hpp>\n#include <iostream>\n"
diff --git a/cmake/get_info.cmake b/cmake/get_info.cmake
index 63f43c645f0..aee19cdd0c9 100644
--- a/cmake/get_info.cmake
+++ b/cmake/get_info.cmake
@@ -133,7 +133,7 @@ foreach(log_type ${log_types})
         "GINKGO_MIXED_PRECISION;GINKGO_HAVE_GPU_AWARE_MPI")
     ginkgo_print_module_footer(${${log_type}} "  Tests, benchmarks and examples:")
     ginkgo_print_foreach_variable(${${log_type}}
-        "GINKGO_BUILD_TESTS;GINKGO_FAST_TESTS;GINKGO_BUILD_EXAMPLES;GINKGO_EXTLIB_EXAMPLE;GINKGO_BUILD_BENCHMARKS;GINKGO_BENCHMARK_ENABLE_TUNING")
+        "GINKGO_BUILD_TESTS;GINKGO_FAST_TESTS;GINKGO_BUILD_EXAMPLES;GINKGO_BUILD_EXTLIB_EXAMPLE;GINKGO_BUILD_BENCHMARKS;GINKGO_BENCHMARK_ENABLE_TUNING")
     ginkgo_print_module_footer(${${log_type}} "  Documentation:")
     ginkgo_print_foreach_variable(${${log_type}} "GINKGO_BUILD_DOC;GINKGO_VERBOSE_LEVEL")
     ginkgo_print_module_footer(${${log_type}} "")
diff --git a/cmake/install_helpers.cmake b/cmake/install_helpers.cmake
index 7165cad0c2b..5e7ea305c55 100644
--- a/cmake/install_helpers.cmake
+++ b/cmake/install_helpers.cmake
@@ -1,6 +1,13 @@
 include(CMakePackageConfigHelpers)
 include(GNUInstallDirs)
 
+if(GINKGO_BENCHMARK_ENABLE_TUNING)
+    install(CODE [=[
+            message(FATAL_ERROR
+                    "Skipping install configuration due to enabled GINKGO_BENCHMARK_ENABLE_TUNING. "
+                    "Please deactivate that variable if you want to install Ginkgo")
+            ]=])
+endif()
 
 set(GINKGO_INSTALL_CONFIG_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/Ginkgo")
 set(GINKGO_INSTALL_MODULE_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/Ginkgo/Modules")
diff --git a/cmake/presets/ci.json b/cmake/presets/ci.json
new file mode 100644
index 00000000000..f0d2c5475c1
--- /dev/null
+++ b/cmake/presets/ci.json
@@ -0,0 +1,575 @@
+{
+    "version": 4,
+    "configurePresets": [
+        {
+            "name": "base-ci",
+            "hidden": true,
+            "binaryDir": "${sourceDir}/build-${presetName}",
+            "generator": "Ninja",
+            "environment": {
+                "MPI_AS_ROOT": "OFF"
+            },
+            "cacheVariables": {
+                "CMAKE_CXX_FLAGS": "",
+                "BUILD_SHARED_LIBS": "ON",
+                "GINKGO_BUILD_REFERENCE": "ON",
+                "GINKGO_BUILD_OMP": "OFF",
+                "GINKGO_BUILD_CUDA": "OFF",
+                "GINKGO_BUILD_HIP": "OFF",
+                "GINKGO_BUILD_SYCL": "OFF",
+                "GINKGO_BUILD_HWLOC": "ON",
+                "GINKGO_BUILD_PAPI_SDE": "OFF",
+                "GINKGO_BUILD_MPI": "OFF",
+                "GINKGO_BUILD_TESTS": "ON",
+                "GINKGO_CONFIG_LOG_DETAILED": "ON",
+                "GINKGO_DEVEL_TOOLS": "OFF",
+                "GINKGO_DPCPP_SINGLE_MODE": "OFF",
+                "GINKGO_FAST_TESTS": "OFF",
+                "GINKGO_MIXED_PRECISION": "ON",
+                "GINKGO_TEST_NONDEFAULT_STREAM": "OFF",
+                "GINKGO_RUN_EXAMPLES": "OFF"
+            }
+        },
+        {
+            "name": "enable-release",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_BUILD_TYPE": "Release",
+                "CMAKE_CONFIGURATION_TYPES": "Release"
+            }
+        },
+        {
+            "name": "enable-debug",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_BUILD_TYPE": "Debug",
+                "CMAKE_CONFIGURATION_TYPES": "Debug"
+            }
+        },
+        {
+            "name": "enable-cuda",
+            "hidden": true,
+            "cacheVariables": {
+                "GINKGO_BUILD_CUDA": "ON",
+                "CMAKE_CUDA_COMPILER": "nvcc",
+                "CMAKE_CUDA_ARCHITECTURES": "$env{CUDA_ARCH}"
+            }
+        },
+        {
+            "name": "enable-sycl",
+            "hidden": true,
+            "cacheVariables": {
+                "GINKGO_BUILD_SYCL": "ON",
+                "CMAKE_CXX_FLAGS": "-Wpedantic -ffp-model=precise",
+                "GINKGO_DPCPP_SINGLE_MODE": "ON"
+            }
+        },
+        {
+            "name": "enable-hip",
+            "hidden": true,
+            "cacheVariables": {
+                "GINKGO_BUILD_HIP": "ON"
+            }
+        },
+        {
+            "name": "enable-openmp",
+            "hidden": true,
+            "cacheVariables": {
+                "GINKGO_BUILD_OMP": "ON"
+            }
+        },
+        {
+            "name": "enable-mpi",
+            "hidden": true,
+            "environment": {
+                "MPI_AS_ROOT": "ON"
+            },
+            "cacheVariables": {
+                "GINKGO_BUILD_MPI": "ON"
+            }
+        },
+        {
+            "name": "enable-gcc",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "g++"
+            }
+        },
+        {
+            "name": "enable-clang",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "clang++"
+            }
+        },
+        {
+            "name": "enable-intel",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "icpx"
+            }
+        },
+        {
+            "name": "enable-dpcpp",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "dpcpp"
+            }
+        },
+        {
+            "name": "enable-nvhpc",
+            "hidden": true,
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "nvc++",
+                "CMAKE_CXX_FLAGS": "--diag_suppress=useless_using_declaration,declared_but_not_referenced",
+                "CMAKE_CUDA_FLAGS": "-diag-suppress=177"
+            }
+        },
+        {
+            "name": "enable-msvc",
+            "hidden": true,
+            "generator": "Visual Studio 17 2022",
+            "cacheVariables": {
+                "CMAKE_CXX_COMPILER": "msvc",
+                "CMAKE_CXX_FLAGS_DEBUG": "/MDd /Zi /Ob1 /O1 /Od /RTC1"
+            }
+        },
+        {
+            "name": "ci/cuda/msvc/nompi/release",
+            "inherits": [
+                "enable-cuda",
+                "enable-msvc",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/cuda+omp/gcc/mpi/release/shared",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-gcc",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/cuda+omp/gcc/mpi/debug/fasttest/nondefault_stream",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-gcc",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "GINKGO_TEST_NONDEFAULT_STREAM": "ON"
+            }
+        },
+        {
+            "name": "ci/cuda+omp/clang/nompi/release/static",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/cuda+omp/clang/nompi/release/fasttest/shared",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "CMAKE_CXX_FLAGS": "-Wno-error=maybe-uninitialized",
+                "CMAKE_CUDA_FLAGS": "-diag-suppress=177"
+            }
+        },
+        {
+            "name": "ci/cuda+omp/gcc/nompi/debug/fasttest",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-gcc",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "CMAKE_CXX_FLAGS": "-Wno-error=maybe-uninitialized",
+                "CMAKE_CUDA_FLAGS": "-diag-suppress=177"
+            }
+        },
+        {
+            "name": "ci/cuda+omp/nvhpc/nompi/release/static/fasttest",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-nvhpc",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/cuda+omp/nvhpc/nompi/debug/fasttest",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "enable-nvhpc",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON"
+            }
+        },
+        {
+            "name": "ci/cuda/msvc/nompi/release/shared",
+            "inherits": [
+                "enable-cuda",
+                "enable-msvc",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/hip+omp/gcc/nompi/debug/static",
+            "inherits": [
+                "enable-hip",
+                "enable-openmp",
+                "enable-gcc",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF",
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/hip+omp/gcc/nompi/release/fasttest/nondefault_stream",
+            "inherits": [
+                "enable-hip",
+                "enable-openmp",
+                "enable-gcc",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "GINKGO_TEST_NONDEFAULT_STREAM": "ON",
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/hip+omp/clang/nompi/debug",
+            "inherits": [
+                "enable-hip",
+                "enable-openmp",
+                "enable-clang",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/hip+omp/clang/nompi/release",
+            "inherits": [
+                "enable-hip",
+                "enable-openmp",
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/hip+omp/clang/nompi/release/static",
+            "inherits": [
+                "ci/hip+omp/clang/nompi/release"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF",
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/hip/gcc/nompi/release",
+            "inherits": [
+                "enable-hip",
+                "enable-gcc",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_RUN_EXAMPLES": "ON"
+            }
+        },
+        {
+            "name": "ci/ref/clang/nompi/release",
+            "inherits": [
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/ref/gcc/nompi/debug/static/fasttest",
+            "inherits": [
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_FAST_TESTS": "ON",
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/ref/msvc/nompi/release",
+            "generator": "Visual Studio 17 2022",
+            "inherits": [
+                "enable-msvc",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/ref/msvc/nompi/release/static",
+            "generator": "Visual Studio 17 2022",
+            "inherits": [
+                "ci/ref/msvc/nompi/release"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/clang/mpi/debug/static/fasttest",
+            "inherits": [
+                "enable-openmp",
+                "enable-clang",
+                "enable-mpi",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF",
+                "GINKGO_FAST_TESTS": "ON"
+            }
+        },
+        {
+            "name": "ci/omp/clang/mpi/glibcxx+debug/release",
+            "inherits": [
+                "enable-openmp",
+                "enable-clang",
+                "enable-mpi",
+                "enable-debug",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "CMAKE_CXX_FLAGS": "-Wpedantic -D_GLIBCXX_DEBUG=1"
+            }
+        },
+        {
+            "name": "ci/omp/clang/nompi/debug/nomixed",
+            "inherits": [
+                "enable-openmp",
+                "enable-clang",
+                "enable-debug",
+                "base-ci"
+            ],
+            "generator": "Unix Makefiles",
+            "cacheVariables": {
+                "GINKGO_MIXED_PRECISION": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/clang/nompi/debug/static/nomixed",
+            "inherits": [
+                "ci/omp/clang/nompi/debug/nomixed"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/clang/nompi/release/static",
+            "inherits": [
+                "enable-openmp",
+                "enable-clang",
+                "enable-release",
+                "base-ci"
+            ],
+            "generator": "Unix Makefiles",
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/clang/nompi/release/static/nomixed",
+            "inherits": [
+                "ci/omp/clang/nompi/release/static"
+            ],
+            "cacheVariables": {
+                "GINKGO_MIXED_PRECISION": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/gcc/nompi/release",
+            "inherits": [
+                "enable-openmp",
+                "enable-gcc",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/omp/gcc/nompi/release/static",
+            "inherits": [
+                "ci/omp/gcc/nompi/release"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/gcc/mpi/release/nomixed",
+            "inherits": [
+                "enable-openmp",
+                "enable-clang",
+                "enable-mpi",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_MIXED_PRECISION": "OFF"
+            }
+        },
+        {
+            "name": "ci/omp/mingwgcc/nompi/release/static",
+            "generator": "MinGW Makefiles",
+            "inherits": [
+                "enable-openmp",
+                "enable-gcc",
+                "enable-release",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/sycl/icpx/nompi/release/shared/singlemode",
+            "inherits": [
+                "enable-sycl",
+                "enable-intel",
+                "enable-mpi",
+                "enable-release",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/sycl/dpcpp/nompi/release/singlemode",
+            "inherits": [
+                "enable-sycl",
+                "enable-dpcpp",
+                "enable-mpi",
+                "enable-release",
+                "base-ci"
+            ],
+            "generator": "Unix Makefiles"
+        },
+        {
+            "name": "ci/sycl/dpcpp/nompi/release/static/singlemode",
+            "inherits": [
+                "ci/sycl/dpcpp/nompi/release/singlemode"
+            ],
+            "cacheVariables": {
+                "BUILD_SHARED_LIBS": "OFF"
+            }
+        },
+        {
+            "name": "ci/sycl/icpx/nompi/release/singlemode/nomixed",
+            "inherits": [
+                "ci/sycl/icpx/nompi/release/shared/singlemode"
+            ],
+            "generator": "Unix Makefiles",
+            "cacheVariables": {
+                "GINKGO_MIXED_PRECISION": "OFF"
+            }
+        },
+        {
+            "name": "ci/warnings",
+            "inherits": [
+                "enable-cuda",
+                "enable-hip",
+                "enable-openmp",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "CMAKE_CXX_FLAGS": "-Wpedantic -Werror=pedantic -pedantic-errors"
+            }
+        },
+        {
+            "name": "ci/no_circular_deps",
+            "inherits": [
+                "enable-cuda",
+                "enable-hip",
+                "enable-openmp",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_CHECK_CIRCULAR_DEPS": "ON"
+            }
+        },
+        {
+            "name": "ci/subdir_build",
+            "inherits": [
+                "enable-openmp",
+                "base-ci"
+            ]
+        },
+        {
+            "name": "ci/clang_tidy",
+            "inherits": [
+                "enable-cuda",
+                "enable-hip",
+                "enable-openmp",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_WITH_CLANG_TIDY": "ON"
+            }
+        },
+        {
+            "name": "ci/iwyu",
+            "inherits": [
+                "enable-cuda",
+                "enable-openmp",
+                "base-ci"
+            ],
+            "cacheVariables": {
+                "GINKGO_WITH_IWYU": "ON"
+            }
+        }
+    ]
+}
diff --git a/cmake/rename.cmake b/cmake/rename.cmake
index d2751bf6623..8964240cea9 100644
--- a/cmake/rename.cmake
+++ b/cmake/rename.cmake
@@ -2,7 +2,7 @@
 macro(gko_rename_cache deprecated actual type doc_string)
     if(DEFINED ${deprecated})
         if(DEFINED ${actual})
-            message("actual ${actual} and deprecated ${deprecated}")
+            message(TRACE "actual ${actual} and deprecated ${deprecated}")
             if("${${actual}}" STREQUAL "${${deprecated}}")
                 # They are the same, so only throw warning
                 message(WARNING "${deprecated} was deprecated, please only use ${actual} instead. We removed the old variable.")
diff --git a/cuda/CMakeLists.txt b/cuda/CMakeLists.txt
index d4a94eda802..a83cf7434c3 100644
--- a/cuda/CMakeLists.txt
+++ b/cuda/CMakeLists.txt
@@ -31,6 +31,13 @@ target_sources(ginkgo_cuda
     ${GKO_UNIFIED_COMMON_SOURCES}
     ${GKO_CUDA_HIP_COMMON_SOURCES}
     )
+if(GINKGO_BENCHMARK_ENABLE_TUNING)
+    target_sources(ginkgo_cuda PRIVATE ${Ginkgo_SOURCE_DIR}/benchmark/utils/tuning_variables.cpp)
+endif()
+# override the default language mapping for the common files, set them to CUDA
+foreach(source_file IN LISTS GKO_UNIFIED_COMMON_SOURCES)
+    set_source_files_properties(${source_file} PROPERTIES LANGUAGE CUDA)
+endforeach(source_file)
 if(GINKGO_JACOBI_FULL_OPTIMIZATIONS)
     set(GKO_CUDA_JACOBI_BLOCK_SIZES)
     foreach(blocksize RANGE 1 32)
diff --git a/cuda/components/format_conversion.cuh b/cuda/components/format_conversion.cuh
index 9ece2cdffe4..9bb57079edd 100644
--- a/cuda/components/format_conversion.cuh
+++ b/cuda/components/format_conversion.cuh
@@ -87,7 +87,7 @@ __host__ size_type calculate_nwarps(std::shared_ptr<const CudaExecutor> exec,
         multiple = 32;
     }
 #ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-    if (_tuning_flag) {
+    if (_tuning_flag && _tuned_value > 0) {
         multiple = _tuned_value;
     }
 #endif  // GINKGO_BENCHMARK_ENABLE_TUNING
diff --git a/dpcpp/CMakeLists.txt b/dpcpp/CMakeLists.txt
index 851ef9a3dc6..c73d5c1c8e4 100644
--- a/dpcpp/CMakeLists.txt
+++ b/dpcpp/CMakeLists.txt
@@ -1,6 +1,6 @@
 find_package(MKL CONFIG REQUIRED HINTS "$ENV{MKLROOT}" "$ENV{MKL_ROOT}")
 find_package(oneDPL REQUIRED HINTS "$ENV{DPL_ROOT}" "$ENV{DPLROOT}")
-# use the parameter from cmake 
+# use the parameter from cmake
 set(GINKGO_MKL_ROOT "${MKL_DIR}" PARENT_SCOPE)
 set(GINKGO_DPL_ROOT "${oneDPL_DIR}" PARENT_SCOPE)
 
@@ -69,6 +69,9 @@ target_sources(ginkgo_dpcpp
     ${GKO_UNIFIED_COMMON_SOURCES}
     ${DENSE_INSTANTIATE}
     )
+if(GINKGO_BENCHMARK_ENABLE_TUNING)
+    target_sources(ginkgo_dpcpp PRIVATE ${Ginkgo_SOURCE_DIR}/benchmark/utils/tuning_variables.cpp)
+endif()
 
 # TODO: adjust it when dpcpp jacobi supports more block size
 set(GKO_DPCPP_JACOBI_BLOCK_SIZES 32)
diff --git a/dpcpp/components/format_conversion.dp.hpp b/dpcpp/components/format_conversion.dp.hpp
index 17cf55389df..4adf706c2a5 100644
--- a/dpcpp/components/format_conversion.dp.hpp
+++ b/dpcpp/components/format_conversion.dp.hpp
@@ -48,7 +48,7 @@ size_type calculate_nwarps(std::shared_ptr<const DpcppExecutor> exec,
         multiple = 32;
     }
 #ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-    if (_tuning_flag) {
+    if (_tuning_flag && _tuned_value > 0) {
         multiple = _tuned_value;
     }
 #endif  // GINKGO_BENCHMARK_ENABLE_TUNING
diff --git a/hip/CMakeLists.txt b/hip/CMakeLists.txt
index 30e675509d5..319dee6b58c 100644
--- a/hip/CMakeLists.txt
+++ b/hip/CMakeLists.txt
@@ -28,6 +28,9 @@ set(GINKGO_HIP_SOURCES
     ${GKO_UNIFIED_COMMON_SOURCES}
     ${GKO_CUDA_HIP_COMMON_SOURCES}
     )
+if(GINKGO_BENCHMARK_ENABLE_TUNING)
+    list(APPEND GINKGO_HIP_SOURCES ${Ginkgo_SOURCE_DIR}/benchmark/utils/tuning_variables.cpp)
+endif()
 
 if(hipfft_FOUND)
     list(APPEND GINKGO_HIP_SOURCES matrix/fft_kernels.hip.cpp)
diff --git a/hip/components/format_conversion.hip.hpp b/hip/components/format_conversion.hip.hpp
index 2e6c4eb1236..c11a0364f3a 100644
--- a/hip/components/format_conversion.hip.hpp
+++ b/hip/components/format_conversion.hip.hpp
@@ -97,7 +97,7 @@ __host__ size_type calculate_nwarps(std::shared_ptr<const HipExecutor> exec,
     }
 #endif  // GINKGO_HIP_PLATFORM_NVCC
 #ifdef GINKGO_BENCHMARK_ENABLE_TUNING
-    if (_tuning_flag) {
+    if (_tuning_flag && _tuned_value > 0) {
         multiple = _tuned_value;
     }
 #endif  // GINKGO_BENCHMARK_ENABLE_TUNING