Acceleration improvements and fixes

.github/actions/build-plugin/action.yaml

@@ -11,7 +11,7 @@ inputs:
   acceleration:
     description: 'Enable acceleration'
     required: false
-    default: 'cpu'
+    default: 'generic'
   codesign:
     description: 'Enable codesigning (macOS only)'
     required: false
@@ -59,6 +59,8 @@ runs:
       if: runner.os == 'Linux'
       shell: zsh --no-rcs --errexit --pipefail {0}
       working-directory: ${{ inputs.workingDirectory }}
+      env:
+        ACCELERATION: ${{ inputs.acceleration }}
       run: |
         : Run Ubuntu Build
 

.github/scripts/.Aptfile

@@ -7,6 +7,8 @@ package 'libicu-dev'
 package 'libopenblas-dev'
 package 'libopenblas-openmp-dev'
 package 'libsimde-dev'
-package 'libvulkan-dev'
 package 'ninja-build', bin: 'ninja'
+package 'nvidia-opencl-dev'
+package 'ocl-icd-opencl-dev'
+package 'opencl-headers'
 package 'pkg-config'

.github/scripts/.build.zsh

@@ -65,7 +65,12 @@ build() {
     linux-aarch64
   )
   local target
-  local config='RelWithDebInfo'
+  if [[ "${+ACCELERATION}" == "amd" ]]
+  then
+    local config='Release'
+  else
+    local config='RelWithDebInfo'
+  fi
   local -r -a _valid_configs=(Debug RelWithDebInfo Release MinSizeRel)
   local -i codesign=0
 
@@ -184,7 +189,7 @@ ${_usage_host:-}"
     setup_ccache
   }
 
-  if [[ ${host_os} == linux ]] {
+  if [[ ${host_os} == linux && ("$(lsb_release -i)" == "Distributor ID: Ubuntu") ]] {
     autoload -Uz setup_linux && setup_linux
   }
 
@@ -230,7 +235,7 @@ ${_usage_host:-}"
           -DCODESIGN_IDENTITY=${CODESIGN_IDENT:--}
         )
 
-        cmake_build_args+=(--preset ${_preset} --parallel --config ${config} -- ONLY_ACTIVE_ARCH=YES)
+        cmake_build_args+=(--verbose --preset ${_preset} --parallel --config ${config} -- ONLY_ACTIVE_ARCH=YES)
         # check the MACOS_ARCH env var to determine the build architecture
         if [[ -n ${MACOS_ARCH} ]] {
           cmake_build_args+=(-arch ${MACOS_ARCH})

.github/scripts/.package.zsh

@@ -256,7 +256,7 @@ ${_usage_host:-}"
       popd
     } else {
       log_group "Archiving ${product_name}..."
-      local output_name="${product_name}-${product_version}-${target##*-}-linux-gnu"
+      local output_name="${product_name}-${product_version}-${target##*-}-${+ACCELERATION}-linux-gnu"
       local _tarflags='cJf'
       if (( _loglevel > 1 || ${+CI} )) _tarflags="v${_tarflags}"
 

.github/scripts/Package-Windows.ps1

@@ -4,8 +4,8 @@ param(
     [string] $Target = 'x64',
     [ValidateSet('Debug', 'RelWithDebInfo', 'Release', 'MinSizeRel')]
     [string] $Configuration = 'RelWithDebInfo',
-    [ValidateSet('cpu', 'hipblas', 'vulkan', 'cuda')]
-    [string] $Acceleration = 'cpu',
+    [ValidateSet('generic', 'nvidia', 'amd')]
+    [string] $Acceleration = 'generic',
     [switch] $BuildInstaller,
     [switch] $SkipDeps
 )

.github/scripts/utils.zsh/check_linux

@@ -18,43 +18,42 @@ if [[ -f /etc/os_release ]] {
 
 log_debug 'Checking for apt-get...'
 if (( ! ${+commands[apt-get]} )) {
-  log_error 'No apt-get command found. Please install apt'
-  return 2
+  log_warning 'No apt-get command found. Please ensure all dependencies are correctly installed'
 } else {
   log_debug "Apt-get located at ${commands[apt-get]}"
-}
 
-local -a dependencies=("${(fA)$(<${SCRIPT_HOME}/.Aptfile)}")
-local -a install_list
-local binary
+  local -a dependencies=("${(fA)$(<${SCRIPT_HOME}/.Aptfile)}")
+  local -a install_list
+  local binary
 
-sudo apt-get update -qq
+  sudo apt-get update -qq
 
-for dependency (${dependencies}) {
-  local -a tokens=(${=dependency//(,|:|\')/})
+  for dependency (${dependencies}) {
+    local -a tokens=(${=dependency//(,|:|\')/})
 
-  if [[ ! ${tokens[1]} == 'package' ]] continue
+    if [[ ! ${tokens[1]} == 'package' ]] continue
 
-  if [[ ${#tokens} -gt 2 && ${tokens[3]} == 'bin' ]] {
-    binary=${tokens[4]}
-  } else {
-    binary=${tokens[2]}
-  }
+    if [[ ${#tokens} -gt 2 && ${tokens[3]} == 'bin' ]] {
+      binary=${tokens[4]}
+    } else {
+      binary=${tokens[2]}
+    }
 
-  if (( ! ${+commands[${binary}]} )) install_list+=(${tokens[2]})
-}
+    if (( ! ${+commands[${binary}]} )) install_list+=(${tokens[2]})
+  }
 
-log_debug "List of dependencies to install: ${install_list}"
-if (( ${#install_list} )) {
-  if (( ! ${+CI} )) log_warning 'Dependency installation via apt may require elevated privileges'
+  log_debug "List of dependencies to install: ${install_list}"
+  if (( ${#install_list} )) {
+    if (( ! ${+CI} )) log_warning 'Dependency installation via apt may require elevated privileges'
 
-  local -a apt_args=(
-    ${CI:+-y}
-    --no-install-recommends
-  )
-  if (( _loglevel == 0 )) apt_args+=(--quiet)
+    local -a apt_args=(
+      ${CI:+-y}
+      --no-install-recommends
+    )
+    if (( _loglevel == 0 )) apt_args+=(--quiet)
 
-  sudo apt-get ${apt_args} install ${install_list}
+    sudo apt-get ${apt_args} install ${install_list}
+  }
 }
 
 rehash

.github/workflows/build-project.yaml

@@ -202,14 +202,42 @@ jobs:
     name: Build for Ubuntu 🐧
     runs-on: ubuntu-22.04
     needs: check-event
+    strategy:
+      matrix:
+        acceleration: [generic, nvidia, amd]
     defaults:
       run:
         shell: bash
+    env:
+      CI: 1
     steps:
+      - name: Free Disk Space (Ubuntu)
+        if: ${{ matrix.acceleration == 'amd' || matrix.acceleration == 'nvidia' }}
+        uses: jlumbroso/free-disk-space@main
+        with:
+          # this might remove tools that are actually needed,
+          # if set to "true" but frees about 6 GB
+          tool-cache: false
+          # Needs to be false to prevent running out of memory
+          swap-storage: false
+          # large-packages takes a long time so don't remove them unless we need to
+          large-packages: false
+
+          android: true
+          dotnet: true
+          haskell: true
+          docker-images: true
+
       - uses: actions/checkout@v4
         with:
           submodules: recursive
           fetch-depth: 0
+
+      - name: "Setup additional apt repos, keys and pre-requisites"
+        run: |
+          wget -qO- https://packages.lunarg.com/lunarg-signing-key-pub.asc | sudo tee /etc/apt/trusted.gpg.d/lunarg.asc
+          sudo wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list http://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list
+          sudo apt update
 
       - uses: awalsh128/cache-apt-pkgs-action@v1.6.0
         with:
@@ -218,17 +246,25 @@ jobs:
             libcurl4-openssl-dev
             libicu-dev
             libgles2-mesa-dev
+            libopenblas0
+            libopenblas0-openmp
+            libopenblas0-pthread
             libopenblas-dev
             libopenblas-openmp-dev
             libqt6svg6-dev
             libsimde-dev
-            libvulkan-dev
+            nvidia-opencl-dev
             obs-studio
+            ocl-icd-opencl-dev
             pkg-config
+            python3-setuptools
+            python3-wheel
             qt6-base-dev
             qt6-base-private-dev
+            vulkan-sdk
           add-repository: ppa:obsproject/obs-studio
           version: 1.0
+          execute_install_scripts: true
 
       - name: Set Up Environment 🔧
         id: setup
@@ -244,6 +280,29 @@ jobs:
 
       - uses: actions-rust-lang/setup-rust-toolchain@v1
 
+      # CUDA toolkit needed for linking against so it can be found when looking up package dependencies
+      - name: "Install CUDA toolkit 12.8"
+        if: ${{ matrix.acceleration == 'nvidia' }}
+        run: |
+          wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
+          sudo dpkg -i cuda-keyring_1.1-1_all.deb
+          sudo apt update
+          sudo apt -y install cuda-12-8
+          export PATH=/usr/local/cuda-12.8/bin${PATH:+:${PATH}}
+          export LD_LIBRARY_PATH=/usr/local/cuda-12.8/lib ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
+
+      # hip SDK needed for linking against so it can be found when looking up package dependencies
+      - name: "Install AMD hip SDK v6.4.2"
+        if: ${{ matrix.acceleration == 'amd' }}
+        run: |
+          wget https://repo.radeon.com/amdgpu-install/6.4.2/ubuntu/jammy/amdgpu-install_6.4.60402-1_all.deb
+          sudo apt install ./amdgpu-install_6.4.60402-1_all.deb
+          sudo apt update
+          sudo apt install python3-setuptools python3-wheel
+          sudo usermod -a -G render,video $LOGNAME # Add the current user to the render and video groups
+          sudo apt install rocm
+          sudo apt update && sudo apt install hipblas
+
       - uses: actions/cache@v4
         id: ccache-cache
         with:
@@ -260,6 +319,7 @@ jobs:
         with:
           target: x86_64
           config: ${{ needs.check-event.outputs.config }}
+          acceleration: ${{ matrix.acceleration }}
 
       - name: Package Plugin 📀
         uses: ./.github/actions/package-plugin
@@ -271,29 +331,29 @@ jobs:
       - name: Upload Source Tarball 🗜️
         uses: actions/upload-artifact@v4
         with:
-          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-sources-${{ needs.check-event.outputs.commitHash }}
-          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-source.*
+          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-sources-${{ needs.check-event.outputs.commitHash }}
+          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-source.*
 
       - name: Upload Artifacts 📡
         uses: actions/upload-artifact@v4
         with:
-          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-ubuntu-22.04-x86_64-${{ needs.check-event.outputs.commitHash }}
-          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-x86_64*.*
+          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-ubuntu-22.04-x86_64-${{ needs.check-event.outputs.commitHash }}
+          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-x86_64*.*
 
       - name: Upload debug symbol artifacts 🪲
         uses: actions/upload-artifact@v4
-        if: ${{ fromJSON(needs.check-event.outputs.package) }}
+        if: ${{ fromJSON(needs.check-event.outputs.package) && (matrix.acceleration != 'amd') }}
         with:
-          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-ubuntu-22.04-x86_64-${{ needs.check-event.outputs.commitHash }}-dbgsym
-          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-x86_64*-dbgsym.ddeb
+          name: ${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-ubuntu-22.04-x86_64-${{ needs.check-event.outputs.commitHash }}-dbgsym
+          path: ${{ github.workspace }}/release/${{ steps.setup.outputs.pluginName }}-${{ steps.setup.outputs.pluginVersion }}-${{ matrix.acceleration }}-x86_64*-dbgsym.ddeb
 
   windows-build:
     name: Build for Windows 🪟
     runs-on: windows-2022
     needs: check-event
     strategy:
       matrix:
-        acceleration: [cpu, cuda, vulkan, hipblas]
+        acceleration: [generic, nvidia, amd]
     defaults:
       run:
         shell: pwsh