CI: omit unused tools for faster release builds (#7432)

This leverages caching, and some reduced installer scope to try
to speed up builds. It also tidies up some windows build logic
that was only relevant for the older generate/cmake builds.

.github/workflows/release.yaml

@@ -1,5 +1,13 @@
 name: release
 
+env:
+  ROCM_WINDOWS_URL: https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe
+  MSYS2_URL: https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe
+  CUDA_12_WINDOWS_URL: https://developer.download.nvidia.com/compute/cuda/12.4.0/local_installers/cuda_12.4.0_551.61_windows.exe
+  CUDA_12_WINDOWS_VER: 12.4
+  CUDA_11_WINDOWS_URL: https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe
+  CUDA_11_WINDOWS_VER: 11.3
+
 on:
   push:
     tags:
@@ -60,70 +68,34 @@ jobs:
       KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
     steps:
       - uses: actions/checkout@v4
+      - name: Set make jobs default
+        run: |
+          echo "MAKEFLAGS=--jobs=$((Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
       - name: Set Version
         shell: bash
         run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
+      - name: Add msys paths
         run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - name: Install msys2
-        run: |
-          $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
-          write-host "Downloading msys2"
-          Invoke-WebRequest -Uri "${msys2_url}" -OutFile "${env:RUNNER_TEMP}\msys2.exe"
-          write-host "Installing msys2"
-          Start-Process "${env:RUNNER_TEMP}\msys2.exe" -ArgumentList @("in", "--confirm-command", "--accept-messages", "--root", "C:/msys64") -NoNewWindow -Wait
           echo "c:\msys64\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: Install msys2 tools
-        run: |
-          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang", "make") -NoNewWindow -Wait
           echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: verify tools
+      - name: Install msys2 tools
         run: |
-          get-command gcc
-          gcc --version
-          get-command make
-          make --version
+          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang") -NoNewWindow -Wait
       - uses: actions/setup-go@v5
         with:
           go-version-file: go.mod
           cache: true
-      - run: go get ./...
       - run: |
-          $gopath=(get-command go).source | split-path -parent
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          $cores = (Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }
-          make -j $cores
+          make
         name: make
       - uses: actions/upload-artifact@v4
         with:
           name: generate-windows-cpu
           path: |
             build/**/*
-            build/**/*.a
             dist/windows-amd64/**
 
   # ROCm generation step
@@ -134,77 +106,49 @@ jobs:
       KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
     steps:
       - uses: actions/checkout@v4
+      - name: Set make jobs default
+        run: |
+          echo "MAKEFLAGS=--jobs=$((Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
       - name: Set Version
         shell: bash
         run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - name: Install msys2
+      - name: Add msys paths
         run: |
-          $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
-          write-host "Downloading msys2"
-          Invoke-WebRequest -Uri "${msys2_url}" -OutFile "${env:RUNNER_TEMP}\msys2.exe"
-          write-host "Installing msys2"
-          Start-Process "${env:RUNNER_TEMP}\msys2.exe" -ArgumentList @("in", "--confirm-command", "--accept-messages", "--root", "C:/msys64") -NoNewWindow -Wait
           echo "c:\msys64\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: Install msys2 tools
-        run: |
-          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang", "make") -NoNewWindow -Wait
           echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: verify tools
+      - name: Install msys2 tools
         run: |
-          get-command gcc
-          gcc --version
-          get-command make
-          make --version
+          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang") -NoNewWindow -Wait
       - uses: actions/setup-go@v5
         with:
           go-version-file: go.mod
           cache: true
-      - name: 'Install ROCm'
+      # ROCM installation steps
+      - name: 'Cache ROCm installer'
+        id: cache-rocm
+        uses: actions/cache@v4
+        with:
+          path: rocm-install.exe
+          key: ${{ env.ROCM_WINDOWS_URL }}
+      - name: 'Conditionally Download ROCm'
+        if: steps.cache-rocm.outputs.cache-hit != 'true'
         run: |
           $ErrorActionPreference = "Stop"
-          write-host "downloading AMD HIP Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP"
-          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
-          write-host "Completed AMD HIP"
+          Invoke-WebRequest -Uri "${env:ROCM_WINDOWS_URL}" -OutFile "rocm-install.exe"
+      - name: 'Install ROCm'
+        run: |
+          Start-Process "rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
       - name: 'Verify ROCm'
         run: |
           & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
+          echo "HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path | select -first 1)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+      - name: make rocm runner
+        run: |
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          $cores = (Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }
-          make -j $cores
-        name: make
+          make -C llama print-HIP_PATH print-HIP_LIB_DIR
+          make rocm
       - uses: actions/upload-artifact@v4
         with:
           name: generate-windows-rocm
@@ -219,39 +163,20 @@ jobs:
     strategy:
       matrix:
         cuda:
-          - version: "11"
-            url: 'https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe'
-          - version: "12"
-            url: 'https://developer.download.nvidia.com/compute/cuda/12.4.0/local_installers/cuda_12.4.0_551.61_windows.exe'
+          - version: "${{ env.CUDA_11_WINDOWS_VER }}"
+            url: ${{ env.CUDA_11_WINDOWS_URL }}
+          - version: "${{ env.CUDA_12_WINDOWS_VER }}"
+            url: ${{ env.CUDA_12_WINDOWS_URL }}
     env:
       KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
     steps:
       - uses: actions/checkout@v4
+      - name: Set make jobs default
+        run: |
+          echo "MAKEFLAGS=--jobs=$((Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
       - name: Set Version
         shell: bash
         run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
       - name: Install msys2
         run: |
           $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
@@ -274,35 +199,38 @@ jobs:
         with:
           go-version-file: go.mod
           cache: true
-      - name: 'Install CUDA ${{ matrix.cuda.version }}'
+      # CUDA installation steps
+      - name: 'Cache CUDA installer'
+        id: cache-cuda
+        uses: actions/cache@v4
+        with:
+          path: cuda-install.exe
+          key: ${{ matrix.cuda.url }}
+      - name: 'Conditionally Download CUDA'
+        if: steps.cache-cuda.outputs.cache-hit != 'true'
         run: |
           $ErrorActionPreference = "Stop"
-          write-host "downloading CUDA Installer"
-          Invoke-WebRequest -Uri "${{ matrix.cuda.url }}" -OutFile "${env:RUNNER_TEMP}\cuda-install.exe"
-          write-host "Installing CUDA"
-          Start-Process "${env:RUNNER_TEMP}\cuda-install.exe" -ArgumentList '-s' -NoNewWindow -Wait
-          write-host "Completed CUDA"
+          Invoke-WebRequest -Uri "${{ matrix.cuda.url }}" -OutFile "cuda-install.exe"
+      - name: 'Install CUDA'
+        run: |
+          $subpackages = @("cudart", "nvcc", "cublas", "cublas_dev") | foreach-object {"${_}_${{ matrix.cuda.version }}"}
+          Start-Process "cuda-install.exe" -ArgumentList (@("-s") + $subpackages) -NoNewWindow -Wait
+      - name: 'Verify CUDA'
+        run: |
+          & (resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0] --version
           $cudaPath=((resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0].path | split-path | split-path)
           $cudaVer=($cudaPath | split-path -leaf ) -replace 'v(\d+).(\d+)', '$1_$2' 
-          echo "$cudaPath\bin" >> $env:GITHUB_PATH
-          echo "CUDA_PATH=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_V${cudaVer}=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" >> $env:GITHUB_ENV
-      - name: 'Verify CUDA'
-        run: nvcc -V
-      - run: go get ./...
-      - name: make
+          echo "$cudaPath\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
+          echo "CUDA_PATH=$cudaPath" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          echo "CUDA_PATH_V${cudaVer}=$cudaPath" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      - name: make cuda runner
         run: |
-          $gopath=(get-command go).source | split-path -parent
-          $cudabin=(get-command nvcc).source | split-path
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$cudabin;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $cores = (Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }
-          make -j $cores
+          make cuda_v$(($env:CUDA_PATH | split-path -leaf) -replace 'v(\d+).*', '$1')
       - uses: actions/upload-artifact@v4
         with:
           name: generate-windows-cuda-${{ matrix.cuda.version }}
@@ -528,10 +456,10 @@ jobs:
           name: generate-windows-cpu
       - uses: actions/download-artifact@v4
         with:
-          name: generate-windows-cuda-11
+          name: generate-windows-cuda-${{ env.CUDA_11_WINDOWS_VER }}
       - uses: actions/download-artifact@v4
         with:
-          name: generate-windows-cuda-12
+          name: generate-windows-cuda-${{ env.CUDA_12_WINDOWS_VER }}
       - uses: actions/download-artifact@v4
         with:
           name: generate-windows-rocm
@@ -541,11 +469,8 @@ jobs:
           path: dist
       - run: dir build
       - run: |
-          $gopath=(get-command go).source | split-path -parent
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
           $env:OLLAMA_SKIP_GENERATE="1"
           $env:ARCH="amd64"
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }

.github/workflows/test.yaml

@@ -1,5 +1,11 @@
 name: test
 
+env:
+  ROCM_WINDOWS_URL: https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe
+  MSYS2_URL: https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe
+  CUDA_12_WINDOWS_URL: https://developer.download.nvidia.com/compute/cuda/12.4.0/local_installers/cuda_12.4.0_551.61_windows.exe
+  CUDA_12_WINDOWS_VER: 12.4
+
 concurrency:
   # For PRs, later CI runs preempt previous ones. e.g. a force push on a PR
   # cancels running CI jobs and starts all new ones.
@@ -99,49 +105,45 @@ jobs:
         with:
           go-version-file: go.mod
           cache: true
-      - name: 'Install ROCm'
+      - name: Set make jobs default
+        run: |
+          echo "MAKEFLAGS=--jobs=$((Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      # ROCM installation steps
+      - name: 'Cache ROCm installer'
+        id: cache-rocm
+        uses: actions/cache@v4
+        with:
+          path: rocm-install.exe
+          key: ${{ env.ROCM_WINDOWS_URL }}
+      - name: 'Conditionally Download ROCm'
+        if: steps.cache-rocm.outputs.cache-hit != 'true'
         run: |
           $ErrorActionPreference = "Stop"
-          write-host "downloading AMD HIP Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP"
-          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
-          write-host "Completed AMD HIP"
+          Invoke-WebRequest -Uri "${env:ROCM_WINDOWS_URL}" -OutFile "rocm-install.exe"
+      - name: 'Install ROCm'
+        run: |
+          Start-Process "rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
       - name: 'Verify ROCm'
         run: |
           & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version
-      - name: Install msys2
+          echo "HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path | select -first 1)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      - name: Add msys paths
         run: |
-          $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
-          write-host "Downloading msys2"
-          Invoke-WebRequest -Uri "${msys2_url}" -OutFile "${env:RUNNER_TEMP}\msys2.exe"
-          write-host "Installing msys2"
-          Start-Process "${env:RUNNER_TEMP}\msys2.exe" -ArgumentList @("in", "--confirm-command", "--accept-messages", "--root", "C:/msys64") -NoNewWindow -Wait
           echo "c:\msys64\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
+          echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
       - name: Install msys2 tools
         run: |
-          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang", "make") -NoNewWindow -Wait
-          echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: verify tools
+          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang") -NoNewWindow -Wait
+
+      - name: make rocm runner
         run: |
-          get-command gcc
-          gcc --version
-          get-command make
-          make --version
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:PATH="$gopath;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          $cores = (Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores
-          write-host $env:HIP_PATH
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }
           make -C llama print-HIP_PATH print-HIP_LIB_DIR
-          make -j $cores rocm
-        name: make
+          make rocm
 
   # CUDA generation step
   runners-windows-cuda:
@@ -154,55 +156,49 @@ jobs:
         with:
           go-version-file: go.mod
           cache: true
-      - name: 'Install CUDA'
+      - name: Set make jobs default
+        run: |
+          echo "MAKEFLAGS=--jobs=$((Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores)" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      # CUDA installation steps
+      - name: 'Cache CUDA installer'
+        id: cache-cuda
+        uses: actions/cache@v4
+        with:
+          path: cuda-install.exe
+          key: ${{ env.CUDA_12_WINDOWS_URL }}
+      - name: 'Conditionally Download CUDA'
+        if: steps.cache-cuda.outputs.cache-hit != 'true'
         run: |
           $ErrorActionPreference = "Stop"
-          write-host "downloading CUDA Installer"
-          Invoke-WebRequest -Uri "https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe" -OutFile "${env:RUNNER_TEMP}\cuda-install.exe"
-          write-host "Installing CUDA"
-          Start-Process "${env:RUNNER_TEMP}\cuda-install.exe" -ArgumentList '-s' -NoNewWindow -Wait
-          write-host "Completed CUDA"
+          Invoke-WebRequest -Uri "${env:CUDA_12_WINDOWS_URL}" -OutFile "cuda-install.exe"
+      - name: 'Install CUDA'
+        run: |
+          $subpackages = @("cudart", "nvcc", "cublas", "cublas_dev") | foreach-object {"${_}_${{ env.CUDA_12_WINDOWS_VER }}"}
+          Start-Process "cuda-install.exe" -ArgumentList (@("-s") + $subpackages) -NoNewWindow -Wait
+      - name: 'Verify CUDA'
+        run: |
+          & (resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0] --version
           $cudaPath=((resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0].path | split-path | split-path)
           $cudaVer=($cudaPath | split-path -leaf ) -replace 'v(\d+).(\d+)', '$1_$2' 
-          echo "$cudaPath\bin" >> $env:GITHUB_PATH
-          echo "CUDA_PATH=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_V${cudaVer}=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" >> $env:GITHUB_ENV
-      - name: 'Verify CUDA'
-        run: nvcc -V
-      - name: Install msys2
+          echo "$cudaPath\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
+          echo "CUDA_PATH=$cudaPath" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          echo "CUDA_PATH_V${cudaVer}=$cudaPath" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+
+      - name: Add msys paths
         run: |
-          $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
-          write-host "Downloading msys2"
-          Invoke-WebRequest -Uri "${msys2_url}" -OutFile "${env:RUNNER_TEMP}\msys2.exe"
-          write-host "Installing msys2"
-          Start-Process "${env:RUNNER_TEMP}\msys2.exe" -ArgumentList @("in", "--confirm-command", "--accept-messages", "--root", "C:/msys64") -NoNewWindow -Wait
           echo "c:\msys64\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: Install msys2 tools
-        run: |
-          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang", "make") -NoNewWindow -Wait
           echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: verify tools
+      - name: Install msys2 tools
         run: |
-          get-command gcc
-          gcc --version
-          get-command make
-          make --version
-      - run: go get ./...
-      - name: make
+          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang") -NoNewWindow -Wait
+      - name: make cuda runner
         run: |
-          $gopath=(get-command go).source | split-path -parent
-          $cudabin=(get-command nvcc).source | split-path
           import-module 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Microsoft.VisualStudio.DevShell.dll'
           Enter-VsDevShell -vsinstallpath 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise' -skipautomaticlocation -DevCmdArguments '-arch=x64 -no_logo'
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$cudabin;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $cores = (Get-ComputerInfo -Property CsProcessors).CsProcessors.NumberOfCores
           if (!(gcc --version | select-string -quiet clang)) { throw "wrong gcc compiler detected - must be clang" }
-          make -j $cores cuda_v11
-        env:
-          OLLAMA_SKIP_CPU_GENERATE: '1'
+          make cuda_v$(($env:CUDA_PATH | split-path -leaf) -replace 'v(\d+).*', '$1')
 
   runners-cpu:
     needs: [changes]
@@ -227,28 +223,15 @@ jobs:
         with:
           go-version-file: go.mod
           cache: true
-      - run: go get ./...
-      - name: Install msys2
+      - name: Add msys paths
         if: ${{ startsWith(matrix.os, 'windows-') }}
         run: |
-          $msys2_url="https://github.com/msys2/msys2-installer/releases/download/2024-07-27/msys2-x86_64-20240727.exe"
-          write-host "Downloading msys2"
-          Invoke-WebRequest -Uri "${msys2_url}" -OutFile "${env:RUNNER_TEMP}\msys2.exe"
-          write-host "Installing msys2"
-          Start-Process "${env:RUNNER_TEMP}\msys2.exe" -ArgumentList @("in", "--confirm-command", "--accept-messages", "--root", "C:/msys64") -NoNewWindow -Wait
           echo "c:\msys64\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: Install msys2 tools
-        if: ${{ startsWith(matrix.os, 'windows-') }}
-        run: |
-          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang", "make") -NoNewWindow -Wait
           echo "C:\msys64\clang64\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
-      - name: verify tools
+      - name: Install msys2 tools
         if: ${{ startsWith(matrix.os, 'windows-') }}
         run: |
-          get-command gcc
-          gcc --version
-          get-command make
-          make --version
+          Start-Process "c:\msys64\usr\bin\pacman.exe" -ArgumentList @("-S", "--noconfirm", "mingw-w64-clang-x86_64-gcc-compat", "mingw-w64-clang-x86_64-clang") -NoNewWindow -Wait
       - name: 'Build Windows Go Runners'
         if: ${{ startsWith(matrix.os, 'windows-') }}
         run: |

scripts/build_windows.ps1

@@ -75,7 +75,6 @@ function checkEnv() {
     } else {
         write-host "Code signing disabled - please set KEY_CONTAINERS to sign and copy ollama_inc.crt to the top of the source tree"
     }
-
 }
 
 
@@ -91,11 +90,6 @@ function buildOllama() {
     write-host "Building ollama CLI"
     & go build -trimpath -ldflags "-s -w -X=github.com/ollama/ollama/version.Version=$script:VERSION -X=github.com/ollama/ollama/server.mode=release" .
     if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
-    if ("${env:KEY_CONTAINER}") {
-        & "${script:SignTool}" sign /v /fd sha256 /t http://timestamp.digicert.com /f "${script:OLLAMA_CERT}" `
-            /csp "Google Cloud KMS Provider" /kc ${env:KEY_CONTAINER} ollama.exe
-        if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
-    }
     New-Item -ItemType Directory -Path .\dist\windows-${script:TARGET_ARCH}\ -Force
     cp .\ollama.exe .\dist\windows-${script:TARGET_ARCH}\
 }
@@ -106,11 +100,6 @@ function buildApp() {
     & windres -l 0 -o ollama.syso ollama.rc
     & go build -trimpath -ldflags "-s -w -H windowsgui -X=github.com/ollama/ollama/version.Version=$script:VERSION -X=github.com/ollama/ollama/server.mode=release" -o "${script:SRC_DIR}\dist\windows-${script:TARGET_ARCH}-app.exe" .
     if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
-    if ("${env:KEY_CONTAINER}") {
-        & "${script:SignTool}" sign /v /fd sha256 /t http://timestamp.digicert.com /f "${script:OLLAMA_CERT}" `
-            /csp "Google Cloud KMS Provider" /kc ${env:KEY_CONTAINER} "${script:SRC_DIR}\dist\windows-${script:TARGET_ARCH}-app.exe"
-        if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
-    }
 }
 
 function gatherDependencies() {
@@ -143,16 +132,19 @@ function gatherDependencies() {
         copy-item -path "${env:VCToolsRedistDir}\vc_redist.arm64.exe" -destination "${script:DIST_DIR}" -verbose
     }
 
-
     cp "${script:SRC_DIR}\app\ollama_welcome.ps1" "${script:SRC_DIR}\dist\"
+}
+
+function sign() {
     if ("${env:KEY_CONTAINER}") {
-        write-host "about to sign"
-        foreach ($file in (get-childitem "${script:DIST_DIR}\lib\ollama\cu*.dll") + @("${script:SRC_DIR}\dist\ollama_welcome.ps1")){
-            write-host "signing $file"
-            & "${script:SignTool}" sign /v /fd sha256 /t http://timestamp.digicert.com /f "${script:OLLAMA_CERT}" `
-                /csp "Google Cloud KMS Provider" /kc ${env:KEY_CONTAINER} $file
-            if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
-        }
+        write-host "Signing Ollama executables, scripts and libraries"
+        & "${script:SignTool}" sign /v /fd sha256 /t http://timestamp.digicert.com /f "${script:OLLAMA_CERT}" `
+            /csp "Google Cloud KMS Provider" /kc ${env:KEY_CONTAINER} `
+            $(get-childitem -path "${script:SRC_DIR}\dist" -r -include @('ollama_welcome.ps1')) `
+            $(get-childitem -path "${script:SRC_DIR}\dist\windows-*" -r -include @('*.exe', '*.dll'))
+        if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
+    } else {
+        write-host "Signing not enabled"
     }
 }
 
@@ -183,6 +175,7 @@ try {
         buildOllama
         buildApp
         gatherDependencies
+        sign
         buildInstaller
         distZip
     } else {