diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 71f7e5186d9..db5d66be0ff 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -8,7 +8,9 @@ on:
     paths: [
       '.github/workflows/build.yml',
       '.github/workflows/build-linux-cross.yml',
-      '.github/workflows/build-cmake-pkg.yml',
+      # Disable. There are some modifications in the fork
+      # so that ggml dynamic builds work with vcpkg.
+      # '.github/workflows/build-cmake-pkg.yml',
       '**/CMakeLists.txt',
       '**/.cmake',
       '**/*.h',
@@ -28,7 +30,9 @@ on:
     paths: [
       '.github/workflows/build.yml',
       '.github/workflows/build-linux-cross.yml',
-      '.github/workflows/build-cmake-pkg.yml',
+      # Disable. There are some modifications in the fork
+      # so that ggml dynamic builds work with vcpkg.
+      # '.github/workflows/build-cmake-pkg.yml',
       '**/CMakeLists.txt',
       '**/.cmake',
       '**/*.h',
@@ -291,946 +295,3 @@ jobs:
           cd build
           ctest -L main --verbose --timeout 900
 
-  ubuntu-latest-llguidance:
-    runs-on: ubuntu-latest
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        id: depends
-        run: |
-          sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
-
-      - name: Build
-        id: cmake_build
-        run: |
-          mkdir build
-          cd build
-          cmake .. \
-            -DLLAMA_FATAL_WARNINGS=ON \
-            -DLLAMA_LLGUIDANCE=ON
-          cmake --build . --config Release -j $(nproc)
-
-      - name: Test
-        id: cmake_test
-        run: |
-          cd build
-          ctest -L main --verbose --timeout 900
-
-  ubuntu-latest-cmake-rpc:
-    runs-on: ubuntu-latest
-
-    continue-on-error: true
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-latest-cmake-rpc
-          evict-old-files: 1d
-
-      - name: Dependencies
-        id: depends
-        run: |
-          sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
-
-      - name: Build
-        id: cmake_build
-        run: |
-          cmake -B build \
-            -DGGML_RPC=ON
-          cmake --build build --config Release -j $(nproc)
-
-      - name: Test
-        id: cmake_test
-        run: |
-          cd build
-          ctest -L main --verbose
-
-  ubuntu-22-cmake-vulkan:
-    runs-on: ubuntu-22.04
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-vulkan
-          evict-old-files: 1d
-
-      - name: Dependencies
-        id: depends
-        run: |
-          wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | sudo apt-key add -
-          sudo wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list
-          sudo apt-get update -y
-          sudo apt-get install -y build-essential mesa-vulkan-drivers vulkan-sdk libcurl4-openssl-dev
-
-      - name: Build
-        id: cmake_build
-        run: |
-          cmake -B build \
-            -DGGML_VULKAN=ON
-          cmake --build build --config Release -j $(nproc)
-
-      - name: Test
-        id: cmake_test
-        run: |
-          cd build
-          export GGML_VK_VISIBLE_DEVICES=0
-          # This is using llvmpipe and runs slower than other backends
-          ctest -L main --verbose --timeout 4200
-
-  ubuntu-22-cmake-webgpu:
-    runs-on: ubuntu-22.04
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-webgpu
-          evict-old-files: 1d
-
-      - name: Vulkan SDK Dependencies
-        id: vulkan-depends
-        run: |
-          wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | sudo apt-key add -
-          sudo wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list
-          sudo apt-get update -y
-          sudo apt-get install -y build-essential mesa-vulkan-drivers vulkan-sdk libcurl4-openssl-dev
-
-      - name: Dawn Dependency
-        id: dawn-depends
-        run: |
-          sudo apt-get install -y libxrandr-dev libxinerama-dev libxcursor-dev mesa-common-dev libx11-xcb-dev libxi-dev
-          DAWN_VERSION="v1.0.0"
-          DAWN_OWNER="reeselevine"
-          DAWN_REPO="dawn"
-          DAWN_ASSET_NAME="Dawn-a1a6b45cced25a3b7f4fb491e0ae70796cc7f22b-ubuntu-latest-Release.tar.gz"
-          echo "Fetching release asset from https://github.com/${DAWN_OWNER}/${DAWN_REPO}/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}"
-          curl -L -o artifact.tar.gz \
-            "https://github.com/${DAWN_OWNER}/${DAWN_REPO}/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}"
-          mkdir dawn
-          tar -xvf artifact.tar.gz -C dawn --strip-components=1
-
-      - name: Build
-        id: cmake_build
-        run: |
-          export Dawn_DIR=dawn/lib64/cmake/Dawn
-          cmake -B build -DGGML_WEBGPU=ON
-          cmake --build build --config Release -j $(nproc)
-
-      - name: Test
-        id: cmake_test
-        run: |
-          cd build
-          # This is using llvmpipe and runs slower than other backends
-          ctest -L main --verbose --timeout 3600
-
-  ubuntu-22-cmake-hip:
-    runs-on: ubuntu-22.04
-    container: rocm/dev-ubuntu-22.04:6.1.2
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        id: depends
-        run: |
-          sudo apt-get update
-          sudo apt-get install -y build-essential git cmake rocblas-dev hipblas-dev libcurl4-openssl-dev
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-hip
-          evict-old-files: 1d
-
-      - name: Build with native CMake HIP support
-        id: cmake_build
-        run: |
-          cmake -B build -S . \
-            -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" \
-            -DGGML_HIP_ROCWMMA_FATTN=ON \
-            -DGGML_HIP=ON
-          cmake --build build --config Release -j $(nproc)
-
-  ubuntu-22-cmake-musa:
-    runs-on: ubuntu-22.04
-    container: mthreads/musa:rc4.2.0-devel-ubuntu22.04-amd64
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        id: depends
-        run: |
-          apt-get update
-          apt-get install -y build-essential git cmake libcurl4-openssl-dev
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-musa
-          evict-old-files: 1d
-
-      - name: Build with native CMake MUSA support
-        id: cmake_build
-        run: |
-          cmake -B build -S . \
-            -DGGML_MUSA=ON
-          cmake --build build --config Release -j $(nproc)
-
-  ubuntu-22-cmake-sycl:
-    runs-on: ubuntu-22.04
-
-    continue-on-error: true
-
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: add oneAPI to apt
-        shell: bash
-        run: |
-          cd /tmp
-          wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          rm GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo add-apt-repository "deb https://apt.repos.intel.com/oneapi all main"
-
-      - name: install oneAPI dpcpp compiler
-        shell: bash
-        run: |
-          sudo apt update
-          sudo apt install intel-oneapi-compiler-dpcpp-cpp libcurl4-openssl-dev
-
-      - name: install oneAPI MKL library
-        shell: bash
-        run: |
-          sudo apt install intel-oneapi-mkl-devel
-
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-sycl
-          evict-old-files: 1d
-
-      - name: Build
-        id: cmake_build
-        run: |
-          source /opt/intel/oneapi/setvars.sh
-          cmake -B build \
-            -DGGML_SYCL=ON \
-            -DCMAKE_C_COMPILER=icx \
-            -DCMAKE_CXX_COMPILER=icpx
-          cmake --build build --config Release -j $(nproc)
-
-  ubuntu-22-cmake-sycl-fp16:
-    runs-on: ubuntu-22.04
-
-    continue-on-error: true
-
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: add oneAPI to apt
-        shell: bash
-        run: |
-          cd /tmp
-          wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          rm GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo add-apt-repository "deb https://apt.repos.intel.com/oneapi all main"
-
-      - name: install oneAPI dpcpp compiler
-        shell: bash
-        run: |
-          sudo apt update
-          sudo apt install intel-oneapi-compiler-dpcpp-cpp libcurl4-openssl-dev
-
-      - name: install oneAPI MKL library
-        shell: bash
-        run: |
-          sudo apt install intel-oneapi-mkl-devel
-
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ubuntu-22-cmake-sycl-fp16
-          evict-old-files: 1d
-
-      - name: Build
-        id: cmake_build
-        run: |
-          source /opt/intel/oneapi/setvars.sh
-          cmake -B build \
-            -DGGML_SYCL=ON \
-            -DCMAKE_C_COMPILER=icx \
-            -DCMAKE_CXX_COMPILER=icpx \
-            -DGGML_SYCL_F16=ON
-          cmake --build build --config Release -j $(nproc)
-
-  build-linux-cross:
-    uses: ./.github/workflows/build-linux-cross.yml
-
-  build-cmake-pkg:
-    uses: ./.github/workflows/build-cmake-pkg.yml
-
-  macOS-latest-cmake-ios:
-    runs-on: macos-latest
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: macOS-latest-cmake-ios
-          evict-old-files: 1d
-
-      - name: Dependencies
-        id: depends
-        continue-on-error: true
-        run: |
-          brew update
-
-      - name: Build
-        id: cmake_build
-        run: |
-          sysctl -a
-          cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
-            -DGGML_METAL_EMBED_LIBRARY=ON \
-            -DLLAMA_BUILD_COMMON=OFF \
-            -DLLAMA_BUILD_EXAMPLES=OFF \
-            -DLLAMA_BUILD_TOOLS=OFF \
-            -DLLAMA_BUILD_TESTS=OFF \
-            -DLLAMA_BUILD_SERVER=OFF \
-            -DCMAKE_SYSTEM_NAME=iOS \
-            -DCMAKE_OSX_DEPLOYMENT_TARGET=14.0 \
-            -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
-          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO
-
-  macOS-latest-cmake-tvos:
-    runs-on: macos-latest
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: macOS-latest-cmake-tvos
-          evict-old-files: 1d
-
-      - name: Dependencies
-        id: depends
-        continue-on-error: true
-        run: |
-          brew update
-
-      - name: Build
-        id: cmake_build
-        run: |
-          sysctl -a
-          cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
-            -DGGML_METAL_EMBED_LIBRARY=ON \
-            -DLLAMA_BUILD_COMMON=OFF \
-            -DLLAMA_BUILD_EXAMPLES=OFF \
-            -DLLAMA_BUILD_TOOLS=OFF \
-            -DLLAMA_BUILD_TESTS=OFF \
-            -DLLAMA_BUILD_SERVER=OFF \
-            -DCMAKE_SYSTEM_NAME=tvOS \
-            -DCMAKE_OSX_DEPLOYMENT_TARGET=14.0 \
-            -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
-          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO
-
-  macOS-latest-cmake-visionos:
-    runs-on: macos-latest
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        id: depends
-        continue-on-error: true
-        run: |
-          brew update
-
-      - name: Build
-        id: cmake_build
-        run: |
-          sysctl -a
-          cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
-            -DGGML_METAL_EMBED_LIBRARY=ON \
-            -DLLAMA_BUILD_COMMON=OFF \
-            -DLLAMA_BUILD_EXAMPLES=OFF \
-            -DLLAMA_BUILD_TOOLS=OFF \
-            -DLLAMA_BUILD_TESTS=OFF \
-            -DLLAMA_BUILD_SERVER=OFF \
-            -DCMAKE_SYSTEM_NAME=visionOS \
-            -DCMAKE_OSX_DEPLOYMENT_TARGET=1.0 \
-            -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
-          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO
-
-  macOS-latest-swift:
-    runs-on: macos-latest
-
-    strategy:
-      matrix:
-        destination: ['generic/platform=macOS', 'generic/platform=iOS', 'generic/platform=tvOS']
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: macOS-latest-swift
-          evict-old-files: 1d
-
-      - name: Dependencies
-        id: depends
-        continue-on-error: true
-        run: |
-          brew update
-
-      - name: Build llama.cpp with CMake
-        id: cmake_build
-        run: |
-          sysctl -a
-          cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
-            -DGGML_METAL_EMBED_LIBRARY=ON \
-            -DLLAMA_CURL=OFF \
-            -DLLAMA_BUILD_EXAMPLES=OFF \
-            -DLLAMA_BUILD_TOOLS=OFF \
-            -DLLAMA_BUILD_TESTS=OFF \
-            -DLLAMA_BUILD_SERVER=OFF \
-            -DCMAKE_OSX_ARCHITECTURES="arm64;x86_64"
-          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
-
-      - name: xcodebuild for swift package
-        id: xcodebuild
-        run: |
-          ./build-xcframework.sh
-
-  windows-msys2:
-    runs-on: windows-2025
-
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - { sys: UCRT64,  env: ucrt-x86_64,  build: Release }
-          - { sys: CLANG64, env: clang-x86_64, build: Release }
-
-    steps:
-      - name: Clone
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: windows-msys2
-          variant: ccache
-          evict-old-files: 1d
-
-      - name: Setup ${{ matrix.sys }}
-        uses: msys2/setup-msys2@v2
-        with:
-          update: true
-          msystem: ${{matrix.sys}}
-          install: >-
-            base-devel
-            git
-            mingw-w64-${{matrix.env}}-toolchain
-            mingw-w64-${{matrix.env}}-cmake
-            mingw-w64-${{matrix.env}}-openblas
-
-      - name: Build using CMake
-        shell: msys2 {0}
-        run: |
-            cmake -B build
-            cmake --build build --config ${{ matrix.build }} -j $(nproc)
-
-      - name: Clean after building using CMake
-        shell: msys2 {0}
-        run: |
-            rm -rf build
-
-      - name: Build using CMake w/ OpenBLAS
-        shell: msys2 {0}
-        run: |
-            cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS
-            cmake --build build --config ${{ matrix.build }} -j $(nproc)
-
-  windows-latest-cmake:
-    runs-on: windows-2025
-
-    env:
-      OPENBLAS_VERSION: 0.3.23
-      SDE_VERSION: 9.33.0-2024-01-07
-      VULKAN_VERSION: 1.4.313.2
-
-    strategy:
-      matrix:
-        include:
-          - build: 'cpu-x64 (static)'
-            arch: 'x64'
-            defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/x64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF'
-          - build: 'openblas-x64'
-            arch: 'x64'
-            defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/x64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_BACKEND_DL=ON -DGGML_CPU_ALL_VARIANTS=ON -DGGML_OPENMP=OFF -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS -DBLAS_INCLUDE_DIRS="$env:RUNNER_TEMP/openblas/include" -DBLAS_LIBRARIES="$env:RUNNER_TEMP/openblas/lib/openblas.lib"'
-          - build: 'vulkan-x64'
-            arch: 'x64'
-            defines: '-DCMAKE_BUILD_TYPE=Release -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_BACKEND_DL=ON -DGGML_CPU_ALL_VARIANTS=ON -DGGML_VULKAN=ON'
-          - build: 'llvm-arm64'
-            arch: 'arm64'
-            defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON'
-          - build: 'llvm-arm64-opencl-adreno'
-            arch: 'arm64'
-            defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/opencl-arm64-release" -DGGML_OPENCL=ON -DGGML_OPENCL_USE_ADRENO_KERNELS=ON'
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: windows-latest-cmake-${{ matrix.build }}
-          variant: ccache
-          evict-old-files: 1d
-
-      - name: Download OpenBLAS
-        id: get_openblas
-        if: ${{ matrix.build == 'openblas-x64' }}
-        run: |
-          curl.exe -o $env:RUNNER_TEMP/openblas.zip -L "https://github.com/xianyi/OpenBLAS/releases/download/v${env:OPENBLAS_VERSION}/OpenBLAS-${env:OPENBLAS_VERSION}-x64.zip"
-          curl.exe -o $env:RUNNER_TEMP/OpenBLAS.LICENSE.txt -L "https://github.com/xianyi/OpenBLAS/raw/v${env:OPENBLAS_VERSION}/LICENSE"
-          mkdir $env:RUNNER_TEMP/openblas
-          tar.exe -xvf $env:RUNNER_TEMP/openblas.zip -C $env:RUNNER_TEMP/openblas
-          $vcdir = $(vswhere -latest -products * -requires Microsoft.VisualStudio.Component.VC.Tools.x86.x64 -property installationPath)
-          $msvc = $(join-path $vcdir $('VC\Tools\MSVC\'+$(gc -raw $(join-path $vcdir 'VC\Auxiliary\Build\Microsoft.VCToolsVersion.default.txt')).Trim()))
-          $lib =  $(join-path $msvc 'bin\Hostx64\x64\lib.exe')
-          & $lib /machine:x64 "/def:${env:RUNNER_TEMP}/openblas/lib/libopenblas.def" "/out:${env:RUNNER_TEMP}/openblas/lib/openblas.lib" /name:openblas.dll
-
-      - name: Install Vulkan SDK
-        id: get_vulkan
-        if: ${{ matrix.build == 'vulkan-x64' }}
-        run: |
-          curl.exe -o $env:RUNNER_TEMP/VulkanSDK-Installer.exe -L "https://sdk.lunarg.com/sdk/download/${env:VULKAN_VERSION}/windows/vulkansdk-windows-X64-${env:VULKAN_VERSION}.exe"
-          & "$env:RUNNER_TEMP\VulkanSDK-Installer.exe" --accept-licenses --default-answer --confirm-command install
-          Add-Content $env:GITHUB_ENV "VULKAN_SDK=C:\VulkanSDK\${env:VULKAN_VERSION}"
-          Add-Content $env:GITHUB_PATH "C:\VulkanSDK\${env:VULKAN_VERSION}\bin"
-
-      - name: Install Ninja
-        id: install_ninja
-        run: |
-          choco install ninja
-
-      - name: Install OpenCL Headers and Libs
-        id: install_opencl
-        if: ${{ matrix.build == 'llvm-arm64-opencl-adreno' }}
-        run: |
-          git clone https://github.com/KhronosGroup/OpenCL-Headers
-          cd OpenCL-Headers
-          cmake -B build `
-            -DBUILD_TESTING=OFF `
-            -DOPENCL_HEADERS_BUILD_TESTING=OFF `
-            -DOPENCL_HEADERS_BUILD_CXX_TESTS=OFF `
-            -DCMAKE_INSTALL_PREFIX="$env:RUNNER_TEMP/opencl-arm64-release"
-          cmake --build build --target install
-          git clone https://github.com/KhronosGroup/OpenCL-ICD-Loader
-          cd OpenCL-ICD-Loader
-          cmake -B build-arm64-release `
-            -A arm64 `
-            -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/opencl-arm64-release" `
-            -DCMAKE_INSTALL_PREFIX="$env:RUNNER_TEMP/opencl-arm64-release"
-          cmake --build build-arm64-release --target install --config release
-
-      - name: libCURL
-        id: get_libcurl
-        uses: ./.github/actions/windows-setup-curl
-        with:
-          architecture: ${{ matrix.arch == 'x64' && 'win64' || 'win64a' }}
-
-      - name: Build
-        id: cmake_build
-        env:
-          CURL_PATH: ${{ steps.get_libcurl.outputs.curl_path }}
-        run: |
-          cmake -S . -B build ${{ matrix.defines }} `
-            -DCURL_LIBRARY="$env:CURL_PATH/lib/libcurl.dll.a" -DCURL_INCLUDE_DIR="$env:CURL_PATH/include"
-          cmake --build build --config Release -j ${env:NUMBER_OF_PROCESSORS}
-          cp $env:CURL_PATH/bin/libcurl-*.dll build/bin/Release
-
-      - name: Add libopenblas.dll
-        id: add_libopenblas_dll
-        if: ${{ matrix.build == 'openblas-x64' }}
-        run: |
-          cp $env:RUNNER_TEMP/openblas/bin/libopenblas.dll ./build/bin/Release/openblas.dll
-          cp $env:RUNNER_TEMP/OpenBLAS.LICENSE.txt ./build/bin/Release/OpenBLAS-${env:OPENBLAS_VERSION}.txt
-
-      - name: Test
-        id: cmake_test
-        if: ${{ matrix.arch == 'x64' }}
-        run: |
-          cd build
-          ctest -L main -C Release --verbose --timeout 900
-
-      # TODO: disabled for now, consider adding tests for all CPU variants instead
-      # - name: Test (Intel SDE)
-      #   id: cmake_test_sde
-      #   if: ${{ matrix.build == 'avx512-x64' && env.HAS_AVX512F == '0' }} # use Intel SDE for AVX-512 emulation
-      #   run: |
-      #     curl.exe -o $env:RUNNER_TEMP/sde.tar.xz -L "https://downloadmirror.intel.com/813591/sde-external-${env:SDE_VERSION}-win.tar.xz"
-      #     # for some weird reason windows tar doesn't like sde tar.xz
-      #     7z x "-o${env:RUNNER_TEMP}" $env:RUNNER_TEMP/sde.tar.xz
-      #     7z x "-o${env:RUNNER_TEMP}" $env:RUNNER_TEMP/sde.tar
-      #     $sde = $(join-path $env:RUNNER_TEMP sde-external-${env:SDE_VERSION}-win/sde.exe)
-      #     cd build
-      #     $env:LLAMA_SKIP_TESTS_SLOW_ON_EMULATOR = 1
-      #     & $sde -future -- ctest -L main -C Release --verbose --timeout 900
-
-  ubuntu-latest-cmake-cuda:
-    runs-on: ubuntu-latest
-    container: nvidia/cuda:12.6.2-devel-ubuntu24.04
-
-    steps:
-        - name: Clone
-          id: checkout
-          uses: actions/checkout@v4
-
-        - name: Install dependencies
-          env:
-            DEBIAN_FRONTEND: noninteractive
-          run: |
-              apt update
-              apt install -y cmake build-essential ninja-build libgomp1 git libcurl4-openssl-dev
-
-        - name: ccache
-          uses: ggml-org/ccache-action@v1.2.16
-          with:
-            key: ubuntu-latest-cmake-cuda
-            evict-old-files: 1d
-
-        - name: Build with CMake
-          run: |
-            cmake -S . -B build -G Ninja \
-              -DCMAKE_BUILD_TYPE=Release \
-              -DCMAKE_CUDA_ARCHITECTURES=89-real \
-              -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined \
-              -DLLAMA_FATAL_WARNINGS=ON \
-              -DGGML_NATIVE=OFF \
-              -DGGML_CUDA=ON
-            cmake --build build
-
-  windows-2022-cmake-cuda:
-    runs-on: windows-2022
-
-    strategy:
-      matrix:
-        cuda: ['12.4']
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Install ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: windows-cuda-${{ matrix.cuda }}
-          variant: ccache
-          evict-old-files: 1d
-
-      - name: Install Cuda Toolkit
-        uses: ./.github/actions/windows-setup-cuda
-        with:
-          cuda_version: ${{ matrix.cuda }}
-
-      - name: Install Ninja
-        id: install_ninja
-        run: |
-          choco install ninja
-
-      - name: libCURL
-        id: get_libcurl
-        uses: ./.github/actions/windows-setup-curl
-
-      - name: Build
-        id: cmake_build
-        shell: cmd
-        env:
-          CURL_PATH: ${{ steps.get_libcurl.outputs.curl_path }}
-        run: |
-          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64
-          cmake -S . -B build -G "Ninja Multi-Config" ^
-            -DLLAMA_BUILD_SERVER=ON ^
-            -DGGML_NATIVE=OFF ^
-            -DGGML_BACKEND_DL=ON ^
-            -DGGML_CPU_ALL_VARIANTS=ON ^
-            -DGGML_CUDA=ON ^
-            -DGGML_RPC=ON ^
-            -DCURL_LIBRARY="%CURL_PATH%/lib/libcurl.dll.a" -DCURL_INCLUDE_DIR="%CURL_PATH%/include"
-          set /A NINJA_JOBS=%NUMBER_OF_PROCESSORS%-1
-          cmake --build build --config Release -j %NINJA_JOBS% -t ggml
-          cmake --build build --config Release
-
-  windows-latest-cmake-sycl:
-    runs-on: windows-2022
-
-    defaults:
-      run:
-        shell: bash
-
-    env:
-      WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/7cd9bba0-7aab-4e30-b3ae-2221006a4a05/intel-oneapi-base-toolkit-2025.1.1.34_offline.exe
-      WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel:intel.oneapi.win.dnnl:intel.oneapi.win.tbb.devel
-      ONEAPI_ROOT: "C:/Program Files (x86)/Intel/oneAPI"
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: windows-latest-cmake-sycl
-          variant: ccache
-          evict-old-files: 1d
-
-      - name: Install
-        run:  |
-          scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL
-
-      # TODO: add libcurl support ; we will also need to modify win-build-sycl.bat to accept user-specified args
-
-      - name: Build
-        id: cmake_build
-        run:  examples/sycl/win-build-sycl.bat
-
-  windows-latest-cmake-hip:
-    if: ${{ github.event.inputs.create_release != 'true' }}
-    runs-on: windows-2022
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-
-      - name: Clone rocWMMA repository
-        id: clone_rocwmma
-        run: |
-          git clone https://github.com/rocm/rocwmma --branch rocm-6.2.4 --depth 1
-
-      - name: Cache ROCm Installation
-        id: cache-rocm
-        uses: actions/cache@v4
-        with:
-          path: C:\Program Files\AMD\ROCm
-          key: rocm-6.1-${{ runner.os }}-v1
-          restore-keys: |
-            rocm-6.1-${{ runner.os }}-
-
-      - name: Install ROCm
-        if: steps.cache-rocm.outputs.cache-hit != 'true'
-        id: depends
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "Downloading AMD HIP SDK Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP SDK"
-          $proc = Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -PassThru
-          $completed = $proc.WaitForExit(600000)
-          if (-not $completed) {
-              Write-Error "ROCm installation timed out after 10 minutes. Killing the process"
-              $proc.Kill()
-              exit 1
-          }
-          if ($proc.ExitCode -ne 0) {
-              Write-Error "ROCm installation failed with exit code $($proc.ExitCode)"
-              exit 1
-          }
-          write-host "Completed AMD HIP SDK installation"
-
-      - name: Verify ROCm
-        id: verify
-        run: |
-          # Find and test ROCm installation
-          $clangPath = Get-ChildItem 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | Select-Object -First 1
-          if (-not $clangPath) {
-            Write-Error "ROCm installation not found"
-            exit 1
-          }
-          & $clangPath.FullName --version
-
-      - name: Install ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: ${{ github.job }}
-          evict-old-files: 1d
-
-      - name: libCURL
-        id: get_libcurl
-        uses: ./.github/actions/windows-setup-curl
-
-      - name: Build
-        id: cmake_build
-        env:
-          CURL_PATH: ${{ steps.get_libcurl.outputs.curl_path }}
-        run: |
-          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake -G "Unix Makefiles" -B build -S . `
-            -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" `
-            -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" `
-            -DCMAKE_CXX_FLAGS="-I$($PWD.Path.Replace('\', '/'))/rocwmma/library/include/" `
-            -DCMAKE_BUILD_TYPE=Release `
-            -DGGML_HIP=ON `
-            -DGGML_HIP_ROCWMMA_FATTN=ON `
-            -DGGML_RPC=ON `
-            -DCURL_LIBRARY="$env:CURL_PATH/lib/libcurl.dll.a" -DCURL_INCLUDE_DIR="$env:CURL_PATH/include"
-          cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
-
-  ios-xcode-build:
-    runs-on: macos-latest
-
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Xcode
-        uses: maxim-lobanov/setup-xcode@v1
-        with:
-          xcode-version: latest-stable
-
-      - name: Build
-        id: cmake_build
-        run: |
-          sysctl -a
-          cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
-            -DGGML_METAL_EMBED_LIBRARY=ON \
-            -DLLAMA_CURL=OFF \
-            -DLLAMA_BUILD_EXAMPLES=OFF \
-            -DLLAMA_BUILD_TOOLS=OFF \
-            -DLLAMA_BUILD_TESTS=OFF \
-            -DLLAMA_BUILD_SERVER=OFF \
-            -DCMAKE_SYSTEM_NAME=iOS \
-            -DCMAKE_OSX_DEPLOYMENT_TARGET=14.0 \
-            -DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
-          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO
-
-      - name: xcodebuild for swift package
-        id: xcodebuild
-        run: |
-          ./build-xcframework.sh
-
-      - name: Build Xcode project
-        run: xcodebuild -project examples/llama.swiftui/llama.swiftui.xcodeproj -scheme llama.swiftui -sdk iphoneos CODE_SIGNING_REQUIRED=NO CODE_SIGN_IDENTITY= -destination 'generic/platform=iOS' FRAMEWORK_FOLDER_PATH=./build-ios build
-
-  android-build:
-    runs-on: ubuntu-latest
-
-    steps:
-      - name: Clone
-        uses: actions/checkout@v4
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.16
-        with:
-          key: android-build
-          evict-old-files: 1d
-
-      - name: Set up JDK
-        uses: actions/setup-java@v3
-        with:
-          java-version: 17
-          distribution: zulu
-
-      - name: Setup Android SDK
-        uses: android-actions/setup-android@v3
-        with:
-          log-accepted-android-sdk-licenses: false
-
-      - name: Build
-        run: |
-          cd examples/llama.android
-          ./gradlew build --no-daemon
-
-  openEuler-latest-cmake-cann:
-    if: ${{ github.event_name != 'pull_request' || contains(github.event.pull_request.labels.*.name, 'Ascend NPU') }}
-    defaults:
-      run:
-        shell: bash -el {0}
-    strategy:
-      matrix:
-        arch: [x86, aarch64]
-        cann:
-          - '8.1.RC1.alpha001-910b-openeuler22.03-py3.10'
-        device:
-          - 'ascend910b3'
-        build:
-          - 'Release'
-    runs-on: ${{ matrix.arch == 'aarch64' && 'ubuntu-24.04-arm' || 'ubuntu-24.04' }}
-    container: ascendai/cann:${{ matrix.cann }}
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Dependencies
-        run: |
-          yum update -y
-          yum install -y git gcc gcc-c++ make cmake libcurl-devel
-
-      - name: Build
-        run: |
-          export LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:${ASCEND_TOOLKIT_HOME}/$(uname -m)-linux/devlib/:${LD_LIBRARY_PATH}
-
-          cmake -S . -B build \
-              -DCMAKE_BUILD_TYPE=${{ matrix.build }} \
-              -DGGML_CANN=on \
-              -DSOC_TYPE=${{ matrix.device }}
-          cmake --build build -j $(nproc)
diff --git a/ggml/cmake/ggml-config.cmake.in b/ggml/cmake/ggml-config.cmake.in
index ef8fa816d2a..97e44a000fc 100644
--- a/ggml/cmake/ggml-config.cmake.in
+++ b/ggml/cmake/ggml-config.cmake.in
@@ -108,6 +108,9 @@ set_and_check(GGML_INCLUDE_DIR "@PACKAGE_GGML_INCLUDE_INSTALL_DIR@")
 set_and_check(GGML_LIB_DIR "@PACKAGE_GGML_LIB_INSTALL_DIR@")
 #set_and_check(GGML_BIN_DIR "@PACKAGE_GGML_BIN_INSTALL_DIR@")
 
+# Include the exported targets file
+include("${CMAKE_CURRENT_LIST_DIR}/ggml-targets.cmake")
+
 if(NOT TARGET ggml::ggml)
     find_package(Threads REQUIRED)
 
diff --git a/ggml/src/CMakeLists.txt b/ggml/src/CMakeLists.txt
index 420831493a7..2fff20752e1 100644
--- a/ggml/src/CMakeLists.txt
+++ b/ggml/src/CMakeLists.txt
@@ -183,8 +183,8 @@ endif()
 
 # ggml
 
-if (GGML_BACKEND_DL AND NOT BUILD_SHARED_LIBS)
-    message(FATAL_ERROR "GGML_BACKEND_DL requires BUILD_SHARED_LIBS")
+if (GGML_BACKEND_DL AND NOT BUILD_SHARED_LIBS AND NOT ${FORCE_BUILD_SHARED_LIBS})
+    message(FATAL_ERROR "GGML_BACKEND_DL requires BUILD_SHARED_LIBS or FORCE_BUILD_SHARED_LIBS")
 endif()
 
 add_library(ggml-base
@@ -229,17 +229,35 @@ if (CMAKE_SYSTEM_NAME MATCHES "Linux")
     target_link_libraries(ggml PRIVATE dl)
 endif()
 
+if (GGML_BACKEND_DL)
+    # Force shared libs
+    set(FORCE_BUILD_SHARED_LIBS ON CACHE BOOL "" FORCE)
+    message(STATUS "GGML_BACKEND_DL enabled: Forcing BUILD_SHARED_LIBS=ON for backends")
+else()
+    set(FORCE_BUILD_SHARED_LIBS OFF CACHE BOOL "" FORCE)
+endif()
+
 function(ggml_add_backend_library backend)
     if (GGML_BACKEND_DL)
         add_library(${backend} MODULE ${ARGN})
         # write the shared library to the output directory
         set_target_properties(${backend} PROPERTIES LIBRARY_OUTPUT_DIRECTORY ${CMAKE_RUNTIME_OUTPUT_DIRECTORY})
         target_compile_definitions(${backend} PRIVATE GGML_BACKEND_DL)
-        add_dependencies(ggml ${backend})
+        # Do not add dependency, the User will have to explicitely build and install
+        # the available `ggml::ggml-*` backend targets. This is for better integration
+        # with cmake-bare
+        # add_dependencies(ggml ${backend})
+
         if (GGML_BACKEND_DIR)
-            install(TARGETS ${backend} LIBRARY DESTINATION ${GGML_BACKEND_DIR})
+            install(TARGETS ${backend}
+                    EXPORT ggml-targets
+                    LIBRARY DESTINATION ${GGML_BACKEND_DIR}
+                    RUNTIME DESTINATION ${GGML_BACKEND_DIR})
         else()
-            install(TARGETS ${backend} LIBRARY DESTINATION ${CMAKE_INSTALL_BINDIR})
+            install(TARGETS ${backend}
+                    EXPORT ggml-targets
+                    LIBRARY DESTINATION ${CMAKE_BINARY_DIR}
+                    RUNTIME DESTINATION ${CMAKE_BINARY_DIR})
         endif()
     else()
         add_library(${backend} ${ARGN})
@@ -250,7 +268,7 @@ function(ggml_add_backend_library backend)
     target_link_libraries(${backend} PRIVATE ggml-base)
     target_include_directories(${backend} PRIVATE ..)
 
-    if (${BUILD_SHARED_LIBS})
+    if (${BUILD_SHARED_LIBS} OR ${FORCE_BUILD_SHARED_LIBS})
         target_compile_definitions(${backend} PRIVATE GGML_BACKEND_BUILD)
         target_compile_definitions(${backend} PUBLIC  GGML_BACKEND_SHARED)
     endif()
@@ -409,7 +427,7 @@ if(CMAKE_SYSTEM_NAME MATCHES "visionOS")
     target_compile_definitions(ggml-base PUBLIC _DARWIN_C_SOURCE)
 endif()
 
-if (BUILD_SHARED_LIBS)
+if (BUILD_SHARED_LIBS OR ${FORCE_BUILD_SHARED_LIBS})
     foreach (target ggml-base ggml)
         set_target_properties(${target} PROPERTIES POSITION_INDEPENDENT_CODE ON)
         target_compile_definitions(${target} PRIVATE GGML_BUILD)
diff --git a/ggml/src/ggml-backend-reg.cpp b/ggml/src/ggml-backend-reg.cpp
index 7002cb07e00..ccae66903cf 100644
--- a/ggml/src/ggml-backend-reg.cpp
+++ b/ggml/src/ggml-backend-reg.cpp
@@ -4,11 +4,13 @@
 #include <algorithm>
 #include <cstring>
 #include <filesystem>
+#include <limits>
 #include <memory>
 #include <string>
 #include <type_traits>
 #include <vector>
 #include <cctype>
+#include <regex>
 
 #ifdef _WIN32
 #    define WIN32_LEAN_AND_MEAN
@@ -499,7 +501,7 @@ static fs::path backend_filename_extension() {
 static ggml_backend_reg_t ggml_backend_load_best(const char * name, bool silent, const char * user_search_path) {
     // enumerate all the files that match [lib]ggml-name-*.[so|dll] in the search paths
     const fs::path name_path = fs::u8path(name);
-    const fs::path file_prefix = backend_filename_prefix().native() + name_path.native() + fs::u8path("-").native();
+    const fs::path file_prefix = backend_filename_prefix().native() + name_path.native();
     const fs::path file_extension = backend_filename_extension();
 
     std::vector<fs::path> search_paths;
@@ -510,6 +512,9 @@ static ggml_backend_reg_t ggml_backend_load_best(const char * name, bool silent,
         // default search paths: executable directory, current directory
         search_paths.push_back(get_executable_path());
         search_paths.push_back(fs::current_path());
+
+        // Android does not require prepending path, the .apk will have embedded the dynamic .so, only the name is needed for dlopen
+        // TODO add here prebuild/ search patch for Desktop platforms where we want to support dynamic loading
     } else {
         search_paths.push_back(fs::u8path(user_search_path));
     }
@@ -517,38 +522,41 @@ static ggml_backend_reg_t ggml_backend_load_best(const char * name, bool silent,
     int best_score = 0;
     fs::path best_path;
 
+    auto tryEntryWithScore = [&best_score, &best_path, silent, _func = __func__](const fs::path & entryPath,
+                                                                                 int              scoreOffset = 1) {
+        dl_handle_ptr handle{ dl_load_library(entryPath) };
+        if (!handle && !silent) {
+            GGML_LOG_ERROR("%s: failed to load %s\n", _func, path_str(entryPath).c_str());
+        }
+        if (handle) {
+            auto score_fn = (ggml_backend_score_t) dl_get_sym(handle.get(), "ggml_backend_score");
+            int  s        = 1;
+            if (score_fn) {
+                s = score_fn() + scoreOffset;
+            }
+#ifdef NDEBUG
+            GGML_LOG_DEBUG("%s: %s score: %d\n", _func, path_str(entryPath).c_str(), s);
+#endif
+            if (s > best_score) {
+                best_score = s;
+                best_path  = entryPath;
+            }
+        }
+    };
+
     for (const auto & search_path : search_paths) {
         if (!fs::exists(search_path)) {
             GGML_LOG_DEBUG("%s: search path %s does not exist\n", __func__, path_str(search_path).c_str());
             continue;
         }
+        GGML_LOG_INFO("%s: searching for %s in %s\n", __func__, path_str(name_path).c_str(), path_str(search_path).c_str());
         fs::directory_iterator dir_it(search_path, fs::directory_options::skip_permission_denied);
         for (const auto & entry : dir_it) {
             if (entry.is_regular_file()) {
                 auto filename = entry.path().filename();
                 auto ext = entry.path().extension();
                 if (filename.native().find(file_prefix) == 0 && ext == file_extension) {
-                    dl_handle_ptr handle { dl_load_library(entry) };
-                    if (!handle && !silent) {
-                        GGML_LOG_ERROR("%s: failed to load %s\n", __func__, path_str(entry.path()).c_str());
-                    }
-                    if (handle) {
-                        auto score_fn = (ggml_backend_score_t) dl_get_sym(handle.get(), "ggml_backend_score");
-                        if (score_fn) {
-                            int s = score_fn();
-#ifndef NDEBUG
-                            GGML_LOG_DEBUG("%s: %s score: %d\n", __func__, path_str(entry.path()).c_str(), s);
-#endif
-                            if (s > best_score) {
-                                best_score = s;
-                                best_path = entry.path();
-                            }
-                        } else {
-                            if (!silent) {
-                                GGML_LOG_INFO("%s: failed to find ggml_backend_score in %s\n", __func__, path_str(entry.path()).c_str());
-                            }
-                        }
-                    }
+                    tryEntryWithScore(entry.path());
                 }
             }
         }
@@ -563,7 +571,27 @@ static ggml_backend_reg_t ggml_backend_load_best(const char * name, bool silent,
                 return get_reg().load_backend(path, silent);
             }
         }
-        return nullptr;
+    }
+
+    // In the case of Android, we can load with just the library filename, without pre-pending any path
+    if (best_path.empty()) {
+        // From worst to best
+        std::vector<fs::path> names = { name_path };
+#ifdef __ANDROID__
+        if (strcmp(name, "cpu") == 0) {
+            names.emplace_back("cpu-android_armv8.0_1");
+            names.emplace_back("cpu-android_armv8.2_1");
+            names.emplace_back("cpu-android_armv8.2_2");
+            names.emplace_back("cpu-android_armv8.6_1");
+        }
+#endif
+        for (size_t scoreOffset = 0; scoreOffset < names.size(); ++scoreOffset) {
+            const auto & loopNamePath = names[scoreOffset];
+            // Try loading backend with just the library name, leave to dlopen path resolution.
+            fs::path     filename     = backend_filename_prefix().native() + loopNamePath.native() +
+                                backend_filename_extension().native();
+            tryEntryWithScore(filename, 1+scoreOffset);
+        }
     }
 
     return get_reg().load_backend(best_path, silent);
@@ -573,6 +601,51 @@ void ggml_backend_load_all() {
     ggml_backend_load_all_from_path(nullptr);
 }
 
+#ifdef __ANDROID__
+namespace {
+// Parses adreno version from gpu description or returns -1 if its not Adreno GPU or -3 if failed to parse the version
+int adrenoVersion(const std::string & gpuDescription) {
+    std::regex  adrenoRegex(R"((\d+))");
+    std::smatch matches;
+    if (gpuDescription.find("dreno") != std::string::npos && std::regex_search(gpuDescription, matches, adrenoRegex) && matches.size() > 1) {
+        try {
+            int adrenoVersion = std::stoi(matches[1].str());
+            return adrenoVersion;
+        } catch (std::invalid_argument & e) {
+            GGML_LOG_ERROR("%s: failed to parse adreno version from %s: %s\n", __func__, gpuDescription.c_str(),
+                           e.what());
+            return -3;
+        }
+    }
+    return -1;
+}
+
+// Returns smallest Adreno version among GPU devices or -1 if there is no adreno GPU
+int minAdrenoVersion(ggml_backend_reg_t vulkanBackend) {
+    if (!vulkanBackend) {
+        return -2;
+    }
+    int minFoundVersion = std::numeric_limits<int>::max();
+    for (size_t i = 0; i < vulkanBackend->iface.get_device_count(vulkanBackend); i++) {
+        ggml_backend_dev_t dev = vulkanBackend->iface.get_device(vulkanBackend, i);
+        if (!dev) {
+            continue;
+        }
+        auto description = std::string(dev->iface.get_description(dev));
+        GGML_LOG_INFO("%s: found device description: %s\n", __func__, description.c_str());
+        int devAdrenoVersion = adrenoVersion(description);
+        if (devAdrenoVersion > 0) {
+            minFoundVersion = std::min(minFoundVersion, devAdrenoVersion);
+        }
+    }
+    if (minFoundVersion < std::numeric_limits<int>::max()) {
+        return minFoundVersion;
+    }
+    return -1;
+}
+}  // namespace
+#endif
+
 void ggml_backend_load_all_from_path(const char * dir_path) {
 #ifdef NDEBUG
     bool silent = true;
@@ -588,7 +661,34 @@ void ggml_backend_load_all_from_path(const char * dir_path) {
     ggml_backend_load_best("rpc", silent, dir_path);
     ggml_backend_load_best("sycl", silent, dir_path);
     ggml_backend_load_best("vulkan", silent, dir_path);
-    ggml_backend_load_best("opencl", silent, dir_path);
+
+    bool useOpencl = true;
+
+#ifdef __ANDROID__
+    // Logic for buggy backends on Adreno GPUs
+    // Use Vulkan backend to obtain GPU information
+    ggml_backend_reg_t vulkanBackend           = ggml_backend_reg_by_name("vulkan");
+    int                devicesMinAdrenoVersion = minAdrenoVersion(vulkanBackend);
+    if (devicesMinAdrenoVersion <= 0) {
+        GGML_LOG_INFO(
+            "%s: no adreno GPU version found (%d) removing OpenCL backend (if any) to rely on Vulkan/cpu only\n",
+            __func__, devicesMinAdrenoVersion);
+        useOpencl = false;
+    } else if (devicesMinAdrenoVersion > 700) {
+        GGML_LOG_INFO("%s: Adreno GPU version %d found keeping OpenCL backend\n", __func__, devicesMinAdrenoVersion);
+    } else if (devicesMinAdrenoVersion > 600) {
+        GGML_LOG_INFO("%s: Adreno GPU version %d should rely on cpu only\n", __func__, devicesMinAdrenoVersion);
+        if (vulkanBackend) {
+            ggml_backend_unload(vulkanBackend);
+            GGML_LOG_INFO("%s: Vulkan backend removed\n", __func__);
+        }
+        useOpencl = false;
+    }
+#endif
+
+    if(useOpencl) {
+        ggml_backend_load_best("opencl", silent, dir_path);
+    }
     ggml_backend_load_best("musa", silent, dir_path);
     ggml_backend_load_best("cpu", silent, dir_path);
     // check the environment variable GGML_BACKEND_PATH to load an out-of-tree backend
@@ -596,4 +696,5 @@ void ggml_backend_load_all_from_path(const char * dir_path) {
     if (backend_path) {
         ggml_backend_load(backend_path);
     }
+
 }
diff --git a/tests/test-gguf.cpp b/tests/test-gguf.cpp
index 3f0c312e2f0..4d3e69fac1d 100644
--- a/tests/test-gguf.cpp
+++ b/tests/test-gguf.cpp
@@ -101,6 +101,12 @@ static bool expect_context_not_null(const enum handcrafted_file_type hft) {
 
 typedef std::pair<enum ggml_type, std::array<int64_t, GGML_MAX_DIMS>> tensor_config_t;
 
+// Helper function to safely cast to gguf_type, suppressing sanitizer warnings for intentional invalid values
+static inline enum gguf_type __attribute__((no_sanitize("undefined"))) 
+safe_cast_to_gguf_type(int value) {
+    return static_cast<enum gguf_type>(value);
+}
+
 static std::vector<tensor_config_t> get_tensor_configs(std::mt19937 & rng) {
     std::vector<tensor_config_t> tensor_configs;
     tensor_configs.reserve(100);
@@ -140,7 +146,9 @@ static std::vector<std::pair<enum gguf_type, enum gguf_type>> get_kv_types(std::
             continue;
         }
 
-        kv_types.push_back(std::make_pair(type, gguf_type(-1)));
+        // Intentionally create invalid enum value for testing error handling
+        // Suppress sanitizer warning as this is intentional undefined behavior for testing
+        kv_types.push_back(std::make_pair(type, safe_cast_to_gguf_type(-1)));
     }
     std::shuffle(kv_types.begin(), kv_types.end(), rng);
 
@@ -232,8 +240,10 @@ static FILE * get_handcrafted_file(const unsigned int seed, const enum handcraft
     }
 
     for (int i = 0; i < int(kv_types.size()); ++i) {
-        const enum gguf_type type     = gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].first);
-        const enum gguf_type type_arr = gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].second);
+        // Intentionally create invalid enum values for testing error handling
+        // Suppress sanitizer warning as this is intentional undefined behavior for testing
+        const enum gguf_type type     = safe_cast_to_gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].first);
+        const enum gguf_type type_arr = safe_cast_to_gguf_type(hft == HANDCRAFTED_KV_BAD_TYPE ? GGUF_TYPE_COUNT : kv_types[i].second);
 
         const std::string key = "my_key_" + std::to_string((hft == HANDCRAFTED_KV_DUPLICATE_KEY ? i/2 : i));
 
@@ -463,8 +473,9 @@ static bool handcrafted_check_kv(const gguf_context * gguf_ctx, const unsigned i
     bool ok = true;
 
     for (int i = 0; i < int(kv_types.size()); ++i) {
-        const enum gguf_type type     = gguf_type(kv_types[i].first);
-        const enum gguf_type type_arr = gguf_type(kv_types[i].second);
+        // Suppress sanitizer warning for intentional invalid enum values in test data
+        const enum gguf_type type     = safe_cast_to_gguf_type(kv_types[i].first);
+        const enum gguf_type type_arr = safe_cast_to_gguf_type(kv_types[i].second);
 
         const std::string key = "my_key_" + std::to_string(i);