fix(cuda/msvc): add CUDA 13.4 CUB fixes, Win32 CreateFileW, and graceful tensor name fallback

.github/workflows/release-prism.yml

@@ -0,0 +1,255 @@
+name: Release (Prism)
+
+on:
+  workflow_dispatch:
+    inputs:
+      create_release:
+        description: 'Create new release'
+        required: true
+        type: boolean
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
+  cancel-in-progress: true
+
+env:
+  BRANCH_NAME: ${{ github.head_ref || github.ref_name }}
+  CMAKE_ARGS: "-DLLAMA_BUILD_EXAMPLES=OFF -DLLAMA_BUILD_TESTS=OFF -DLLAMA_BUILD_TOOLS=ON -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON"
+
+jobs:
+  macOS-arm64:
+    runs-on: macos-14
+
+    steps:
+      - name: Clone
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.16
+        with:
+          key: macOS-latest-cmake-arm64
+          evict-old-files: 1d
+
+      - name: Build
+        run: |
+          cmake -B build \
+            -DCMAKE_INSTALL_RPATH='@loader_path' \
+            -DCMAKE_BUILD_WITH_INSTALL_RPATH=ON \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DGGML_METAL_USE_BF16=ON \
+            -DGGML_METAL_EMBED_LIBRARY=ON \
+            -DGGML_RPC=ON \
+            ${{ env.CMAKE_ARGS }}
+          cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Pack artifacts
+        run: |
+          cp LICENSE ./build/bin/
+          tar -czvf llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.tar.gz -s ",./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v6
+        with:
+          path: llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.tar.gz
+          name: llama-bin-macos-arm64.tar.gz
+
+  linux-cuda:
+    runs-on: ubuntu-22.04
+
+    strategy:
+      matrix:
+        include:
+          - cuda: '12.4'
+            cuda_pkg: '12-4'
+          - cuda: '12.8'
+            cuda_pkg: '12-8'
+          - cuda: '13.1'
+            cuda_pkg: '13-1'
+
+    steps:
+      - name: Clone
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.16
+        with:
+          key: ubuntu-22-cmake-cuda-${{ matrix.cuda }}
+          evict-old-files: 1d
+
+      - name: Install CUDA toolkit
+        run: |
+          wget -q https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
+          sudo dpkg -i cuda-keyring_1.1-1_all.deb
+          sudo apt-get update
+          sudo apt-get -y install cuda-toolkit-${{ matrix.cuda_pkg }}
+          echo "/usr/local/cuda-${{ matrix.cuda }}/bin" >> $GITHUB_PATH
+          echo "CUDA_PATH=/usr/local/cuda-${{ matrix.cuda }}" >> $GITHUB_ENV
+          echo "LD_LIBRARY_PATH=/usr/local/cuda-${{ matrix.cuda }}/lib64:$LD_LIBRARY_PATH" >> $GITHUB_ENV
+
+      - name: Build
+        run: |
+          cmake -B build \
+            -DCMAKE_INSTALL_RPATH='$ORIGIN' \
+            -DCMAKE_BUILD_WITH_INSTALL_RPATH=ON \
+            -DGGML_NATIVE=OFF \
+            -DGGML_CUDA=ON \
+            ${{ env.CMAKE_ARGS }}
+          cmake --build build --config Release -j $(nproc) 2>&1 | grep -v "^nvcc warning"
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Pack artifacts
+        run: |
+          cp LICENSE ./build/bin/
+          tar -czvf llama-${{ steps.tag.outputs.name }}-bin-linux-cuda-${{ matrix.cuda }}-x64.tar.gz --transform "s,./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v6
+        with:
+          path: llama-${{ steps.tag.outputs.name }}-bin-linux-cuda-${{ matrix.cuda }}-x64.tar.gz
+          name: llama-bin-linux-cuda-${{ matrix.cuda }}-x64.tar.gz
+
+  windows-cuda:
+    runs-on: windows-2022
+
+    strategy:
+      matrix:
+        cuda: ['12.4', '13.1']
+
+    steps:
+      - name: Clone
+        uses: actions/checkout@v6
+
+      - name: Install ccache
+        uses: ggml-org/ccache-action@v1.2.16
+        with:
+          key: windows-cuda-${{ 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
+        run: choco install ninja
+
+      - name: Build
+        shell: cmd
+        run: |
+          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" x64
+          cmake -S . -B build -G "Ninja Multi-Config" ^
+            -DGGML_NATIVE=OFF ^
+            -DGGML_CUDA=ON ^
+            -DLLAMA_BUILD_BORINGSSL=ON ^
+            -DCMAKE_CUDA_FLAGS="-diag-suppress=221" ^
+            ${{ env.CMAKE_ARGS }}
+          set /A NINJA_JOBS=%NUMBER_OF_PROCESSORS%-1
+          cmake --build build --config Release -j %NINJA_JOBS%
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Pack artifacts
+        run: |
+          7z a -snl llama-${{ steps.tag.outputs.name }}-bin-win-cuda-${{ matrix.cuda }}-x64.zip .\build\bin\Release\*
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v6
+        with:
+          path: llama-${{ steps.tag.outputs.name }}-bin-win-cuda-${{ matrix.cuda }}-x64.zip
+          name: llama-bin-win-cuda-${{ matrix.cuda }}-x64.zip
+
+      - name: Copy and pack Cuda runtime
+        run: |
+          echo "Cuda install location: ${{ env.CUDA_PATH }}"
+          $dst='.\build\bin\cudart\'
+          robocopy "${{env.CUDA_PATH}}\bin" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll
+          robocopy "${{env.CUDA_PATH}}\lib" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll
+          robocopy "${{env.CUDA_PATH}}\bin\x64" $dst cudart64_*.dll cublas64_*.dll cublasLt64_*.dll
+          7z a cudart-llama-bin-win-cuda-${{ matrix.cuda }}-x64.zip $dst\*
+
+      - name: Upload Cuda runtime
+        uses: actions/upload-artifact@v6
+        with:
+          path: cudart-llama-bin-win-cuda-${{ matrix.cuda }}-x64.zip
+          name: cudart-llama-bin-win-cuda-${{ matrix.cuda }}-x64.zip
+
+  release:
+    if: ${{ github.event.inputs.create_release == 'true' }}
+
+    permissions:
+      contents: write
+
+    runs-on: ubuntu-latest
+
+    needs:
+      - macOS-arm64
+      - linux-cuda
+      - windows-cuda
+
+    steps:
+      - name: Clone
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Download artifacts
+        uses: actions/download-artifact@v7
+        with:
+          path: ./artifact
+          merge-multiple: true
+
+      - name: Move artifacts
+        run: |
+          mkdir -p release
+          mv -v artifact/*.tar.gz release/ 2>/dev/null || true
+          mv -v artifact/*.zip release/ 2>/dev/null || true
+          ls -lh release/
+
+      - name: Create release
+        id: create_release
+        uses: ggml-org/action-create-release@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          tag_name: ${{ steps.tag.outputs.name }}
+          body: |
+            Pre-built binaries (PrismML fork with Q1_0 1-bit quantization support).
+
+            **macOS:**
+            - [macOS Apple Silicon (arm64)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-macos-arm64.tar.gz)
+
+            **Linux:**
+            - [Linux x64 (CUDA 12.4)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-linux-cuda-12.4-x64.tar.gz)
+            - [Linux x64 (CUDA 12.8)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-linux-cuda-12.8-x64.tar.gz)
+            - [Linux x64 (CUDA 13.1)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-linux-cuda-13.1-x64.tar.gz)
+
+            **Windows:**
+            - [Windows x64 (CUDA 12.4)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-win-cuda-12.4-x64.zip) - [CUDA 12.4 DLLs](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/cudart-llama-bin-win-cuda-12.4-x64.zip)
+            - [Windows x64 (CUDA 13.1)](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-win-cuda-13.1-x64.zip) - [CUDA 13.1 DLLs](https://github.com/${{ github.repository }}/releases/download/${{ steps.tag.outputs.name }}/cudart-llama-bin-win-cuda-13.1-x64.zip)
+
+      - name: Upload release
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          for file in release/*; do
+            echo "Uploading $(basename $file)..."
+            gh release upload ${{ steps.tag.outputs.name }} "$file" --clobber
+          done

README.md

@@ -1,3 +1,49 @@
+> [!IMPORTANT]
+> This is a fork of the [Prism-ML fork of llama.cpp](https://github.com/PrismML-Eng/llama.cpp) that is synced to the main llama.cpp repo
+> It is not yet ready for production use and should be considered experimental
+> 
+> The primary benefit of this fork is an up-to-date version of llama.cpp that also merges the capabiilities of the Prism-ML fork
+> to support the [Bonsai 1-bit models](https://huggingface.co/prism-ml/Bonsai-8B-gguf)
+>
+> Note: this is not an official fork and is not supported by the Prism-ML team - this is just a personal fork to demo Bonsai until official support is added
+
+
+## How to use this fork
+
+```bash
+# On MacOS
+git clone https://github.com/Mintplex-Labs/prism-ml-llama.cpp
+cd prism-ml-llama.cpp
+cmake -B build && cmake --build build -j
+```
+
+### You __must__ recode the public Bonsai 1-bit models to work with this fork
+```bash
+# Download the public Bonsai 1-bit models
+wget https://huggingface.co/prism-ml/Bonsai-8B-gguf/resolve/main/Bonsai-8B.gguf -O Bonsai-8B.gguf
+```
+
+### Run the model
+```bash
+# Llama cli
+./build/bin/llama-cli \
+    -m Bonsai-8B.gguf \
+    -p "Explain quantum computing in simple terms." \
+    -n 256 \
+    --temp 0.5 \
+    --top-p 0.85 \
+    --top-k 20 \
+    -ngl 99
+
+# llama server
+./build/bin/llama-server \
+    -m Bonsai-8B.gguf \
+    --host 0.0.0.0 \
+    --port 8080 \
+    -ngl 99
+    --ctx-size 65536
+```
+
 # llama.cpp
 
 ![llama](https://user-images.githubusercontent.com/1991296/230134379-7181e485-c521-4d23-a0d6-f7b3b61ba524.png)