Fixed windows rocm build and changed all builds to be closer to the original sd.cpp builds

.github/workflows/backends.yml

@@ -101,7 +101,7 @@ jobs:
           mkdir build
           cd build
           call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvars64.bat"
-          cmake .. -DSD_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES="61;75;86;89;100" -DSD_BUILD_SHARED_LIBS=ON
+          cmake .. -DSD_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES="61;70;75;80;86;89;90;100;120" -DCMAKE_CUDA_FLAGS='-Xcudafe \"--diag_suppress=177\" -Xcudafe \"--diag_suppress=550\"'" -DSD_BUILD_SHARED_LIBS=ON
           cmake --build . --config Release
       
       - name: Upload artifact
@@ -114,27 +114,11 @@ jobs:
   windows-hip:
     runs-on: windows-latest
 
-    steps:
+    env:
-      - name: Install
+      HIPSDK_INSTALLER_VERSION: "25.Q3"
-        id: depends
+      GPU_TARGETS: "gfx1151;gfx1200;gfx1201;gfx1100;gfx1101;gfx1102;gfx1030;gfx1031;gfx1032"
-        run: |
+      
-          $ErrorActionPreference = "Stop"
+    steps:          
-          write-host "Downloading AMD HIP SDK Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-25.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP SDK"
-          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
-          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
-          }
-          & $clangPath.FullName --version
-          
       - name: Checkout
         id: checkout
         uses: actions/checkout@v4.2.2
@@ -143,15 +127,66 @@ jobs:
           ref: '${{ github.event.inputs.commit }}'
           submodules: recursive  
 
-      - name: Build
+      - name: Cache ROCm Installation
-        id: cmake_build
+        id: cache-rocm
+        uses: actions/cache@v4
+        with:
+          path: C:\Program Files\AMD\ROCm
+          key: rocm-${{ env.HIPSDK_INSTALLER_VERSION }}-${{ runner.os }}
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.16
+        with:
+          key: windows-latest-cmake-hip-${{ env.HIPSDK_INSTALLER_VERSION }}-x64
+          evict-old-files: 1d
+
+      - name: Install ROCm
+        if: steps.cache-rocm.outputs.cache-hit != 'true'
+        run: |
+          $ErrorActionPreference = "Stop"
+          write-host "Downloading AMD HIP SDK Installer"
+          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-${{ env.HIPSDK_INSTALLER_VERSION }}-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
+        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
+          }
+          & $clangPath.FullName --version
+          # Set HIP_PATH environment variable for later steps
+          echo "HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)" >> $env:GITHUB_ENV
+
+      - name: Build
         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}"
           mkdir build
           cd build
-          cmake -G "Unix Makefiles" .. -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS="gfx1100;gfx1101;gfx1030" -DGGML_RPC=ON -DSD_BUILD_SHARED_LIBS=ON
+          $env:CMAKE_PREFIX_PATH="${env:HIP_PATH}"
-          cmake --build . --config Release
+          cmake .. `
+            -G "Unix Makefiles" `
+            -DSD_HIPBLAS=ON `
+            -DSD_BUILD_SHARED_LIBS=ON `
+            -DGGML_NATIVE=OFF `
+            -DCMAKE_C_COMPILER=clang `
+            -DCMAKE_CXX_COMPILER=clang++ `
+            -DCMAKE_BUILD_TYPE=Release `
+            -DGPU_TARGETS="${{ env.GPU_TARGETS }}"
+          cmake --build . --config Release --parallel ${env:NUMBER_OF_PROCESSORS}
               
       - name: Upload artifact
         id: upload_artifact
@@ -222,6 +257,12 @@ jobs:
           ref: '${{ github.event.inputs.commit }}'
           submodules: recursive
 
+      - name: Dependencies
+        id: depends
+        run: |
+          sudo apt-get update
+          sudo apt-get install build-essential
+
       - name: Build
         id: cmake_build
         run: |
@@ -237,6 +278,39 @@ jobs:
           name: linux-${{ matrix.build }}
           path: ./build/bin/libstable-diffusion.so
           
+  linux-vulkan:
+    runs-on: ubuntu-latest
+            
+    steps:
+      - name: Checkout
+        id: checkout
+        uses: actions/checkout@v4.2.2
+        with:
+          repository: 'leejet/stable-diffusion.cpp'
+          ref: '${{ github.event.inputs.commit }}'
+          submodules: recursive
+
+      - name: Dependencies
+        id: depends
+        run: |
+          sudo apt-get update
+          sudo apt-get install build-essential libvulkan-dev glslc
+          
+      - name: Build
+        id: cmake_build
+        run: |
+          mkdir build
+          cd build
+          cmake .. -DSD_BUILD_SHARED_LIBS=ON -DSD_VULKAN=ON
+          cmake --build . --config Release
+                  
+      - name: Upload artifact
+        id: upload_artifact
+        uses: actions/upload-artifact@v4.6.1
+        with:
+          name: linux-vulkan
+          path: ./build/bin/libstable-diffusion.so
+          
   linux-cuda12:
     runs-on: ubuntu-latest
     container: nvidia/cuda:12.8.0-devel-ubuntu24.04
@@ -313,31 +387,96 @@ jobs:
           path: ./build/bin/libstable-diffusion.so
           
   linux-hip:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
-    container: rocm/dev-ubuntu-22.04:6.4.4
+    container: rocm/dev-ubuntu-24.04:7.2
+
+    env:
+      ROCM_VERSION: "7.2"
+      UBUNTU_VERSION: "24.04"
+      GPU_TARGETS: "gfx1151;gfx1150;gfx1100;gfx1101;gfx1102;gfx1200;gfx1201"
    
     steps:
+      - run: apt-get update && apt-get install -y git
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+        with:
+          submodules: recursive
+
+      - name: Free disk space
+        run: |
+          # Remove preinstalled SDKs and caches not needed for this job
+          sudo rm -rf /usr/share/dotnet || true
+          sudo rm -rf /usr/local/lib/android || true
+          sudo rm -rf /opt/ghc || true
+          sudo rm -rf /usr/local/.ghcup || true
+          sudo rm -rf /opt/hostedtoolcache || true
+
+          # Remove old package lists and caches
+          sudo rm -rf /var/lib/apt/lists/* || true
+          sudo apt clean
+
       - name: Dependencies
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install -y build-essential git cmake rocblas-dev hipblas-dev
+          sudo apt install -y \
-          hipconfig
+            cmake \
-          
+            hip-dev \
-      - name: Clone
+            hipblas-dev \
-        id: checkout
+            ninja-build \
-        uses: actions/checkout@v4.2.2
+            rocm-dev \
-        with:
+            zip
-          repository: 'leejet/stable-diffusion.cpp'
+          # Clean apt caches to recover disk space
-          ref: '${{ github.event.inputs.commit }}'
+          sudo apt clean
-          submodules: recursive
+          sudo rm -rf /var/lib/apt/lists/* || true
-                    
+
+      - name: Setup ROCm Environment
+        run: |
+          # Add ROCm to PATH for current session
+          echo "/opt/rocm/bin" >> $GITHUB_PATH
+
+          # Build regex pattern from ${{ env.GPU_TARGETS }} (match target as substring)
+          TARGET_REGEX="($(printf '%s' "${{ env.GPU_TARGETS }}" | sed 's/;/|/g'))"
+
+          # Remove library files for architectures we're not building for to save disk space
+          echo "Cleaning up unneeded architecture files..."
+          cd /opt/rocm/lib/rocblas/library
+          # Keep only our target architectures
+          for file in *; do
+            if printf '%s' "$file" | grep -q 'gfx'; then
+              if ! printf '%s' "$file" | grep -Eq "$TARGET_REGEX"; then
+                echo "Removing $file" &&
+                sudo rm -f "$file";
+              fi
+            fi
+          done
+
+          cd /opt/rocm/lib/hipblaslt/library
+          for file in *; do
+            if printf '%s' "$file" | grep -q 'gfx'; then
+              if ! printf '%s' "$file" | grep -Eq "$TARGET_REGEX"; then
+                echo "Removing $file" &&
+                sudo rm -f "$file";
+              fi
+            fi
+          done
+
       - name: Build
         id: cmake_build
         run: |
           mkdir build
           cd build
-          cmake .. -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIP=ON -DSD_BUILD_SHARED_LIBS=ON
+          cmake .. -G Ninja \
+            -DCMAKE_CXX_COMPILER=amdclang++ \
+            -DCMAKE_C_COMPILER=amdclang \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DSD_HIPBLAS=ON \
+            -DGPU_TARGETS="${{ env.GPU_TARGETS }}" \
+            -DAMDGPU_TARGETS="${{ env.GPU_TARGETS }}" \
+            -DCMAKE_BUILD_WITH_INSTALL_RPATH=ON \
+            -DCMAKE_POSITION_INDEPENDENT_CODE=ON \
+            -DSD_BUILD_SHARED_LIBS=ON
           cmake --build . --config Release
                
       - name: Upload artifact Rocm