Merge remote-tracking branch 'origin/master' into feature/online-flow

Author: slaren

.devops/llama-cli-cuda.Dockerfile

@@ -23,15 +23,16 @@ RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
         export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
     fi && \
     cmake -B build -DGGML_CUDA=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
-    cmake --build build --config Release --target llama-cli -j$(nproc)
+    cmake --build build --config Release --target llama-cli -j$(nproc) && \
+    mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
 
 FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime
 
 RUN apt-get update && \
     apt-get install -y libgomp1
 
-COPY --from=build /app/build/ggml/src/libggml.so /libggml.so
-COPY --from=build /app/build/src/libllama.so /libllama.so
-COPY --from=build /app/build/bin/llama-cli /llama-cli
+COPY --from=build /app/lib/ /
+COPY --from=build /app/build/bin/llama-cli /
 
 ENTRYPOINT [ "/llama-cli" ]

.devops/llama-cli-musa.Dockerfile

@@ -16,15 +16,16 @@ WORKDIR /app
 COPY . .
 
 RUN cmake -B build -DGGML_MUSA=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
-    cmake --build build --config Release --target llama-cli -j$(nproc)
+    cmake --build build --config Release --target llama-cli -j$(nproc) && \
+    mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
 
 FROM ${BASE_MUSA_RUN_CONTAINER} AS runtime
 
 RUN apt-get update && \
     apt-get install -y libgomp1
 
-COPY --from=build /app/build/ggml/src/libggml.so /libggml.so
-COPY --from=build /app/build/src/libllama.so /libllama.so
+COPY --from=build /app/lib/ /
 COPY --from=build /app/build/bin/llama-cli /llama-cli
 
 ENTRYPOINT [ "/llama-cli" ]

.devops/llama-server-cuda.Dockerfile

@@ -23,15 +23,16 @@ RUN if [ "${CUDA_DOCKER_ARCH}" != "default" ]; then \
         export CMAKE_ARGS="-DCMAKE_CUDA_ARCHITECTURES=${CUDA_DOCKER_ARCH}"; \
     fi && \
     cmake -B build -DGGML_CUDA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
-    cmake --build build --config Release --target llama-server -j$(nproc)
+    cmake --build build --config Release --target llama-server -j$(nproc) && \
+    mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
 
 FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime
 
 RUN apt-get update && \
     apt-get install -y libcurl4-openssl-dev libgomp1 curl
 
-COPY --from=build /app/build/ggml/src/libggml.so /libggml.so
-COPY --from=build /app/build/src/libllama.so /libllama.so
+COPY --from=build /app/lib/ /
 COPY --from=build /app/build/bin/llama-server /llama-server
 
 # Must be set to 0.0.0.0 so it can listen to requests from host machine

.devops/llama-server-musa.Dockerfile

@@ -16,15 +16,16 @@ WORKDIR /app
 COPY . .
 
 RUN cmake -B build -DGGML_MUSA=ON -DLLAMA_CURL=ON ${CMAKE_ARGS} -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined . && \
-    cmake --build build --config Release --target llama-server -j$(nproc)
+    cmake --build build --config Release --target llama-server -j$(nproc) && \
+    mkdir -p /app/lib && \
+    find build -name "*.so" -exec cp {} /app/lib \;
 
 FROM ${BASE_MUSA_RUN_CONTAINER} AS runtime
 
 RUN apt-get update && \
     apt-get install -y libcurl4-openssl-dev libgomp1 curl
 
-COPY --from=build /app/build/ggml/src/libggml.so /libggml.so
-COPY --from=build /app/build/src/libllama.so /libllama.so
+COPY --from=build /app/lib/ /
 COPY --from=build /app/build/bin/llama-server /llama-server
 
 # Must be set to 0.0.0.0 so it can listen to requests from host machine

.devops/nix/package.nix

@@ -126,9 +126,9 @@ effectiveStdenv.mkDerivation (finalAttrs: {
   };
 
   postPatch = ''
-    substituteInPlace ./ggml/src/ggml-metal.m \
+    substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \
       --replace '[bundle pathForResource:@"ggml-metal" ofType:@"metal"];' "@\"$out/bin/ggml-metal.metal\";"
-    substituteInPlace ./ggml/src/ggml-metal.m \
+    substituteInPlace ./ggml/src/ggml-metal/ggml-metal.m \
       --replace '[bundle pathForResource:@"default" ofType:@"metallib"];' "@\"$out/bin/default.metallib\";"
   '';
 
@@ -173,7 +173,7 @@ effectiveStdenv.mkDerivation (finalAttrs: {
       (cmakeBool "GGML_NATIVE" false)
       (cmakeBool "GGML_BLAS" useBlas)
       (cmakeBool "GGML_CUDA" useCuda)
-      (cmakeBool "GGML_HIPBLAS" useRocm)
+      (cmakeBool "GGML_HIP" useRocm)
       (cmakeBool "GGML_METAL" useMetalKit)
       (cmakeBool "GGML_VULKAN" useVulkan)
       (cmakeBool "GGML_STATIC" enableStatic)

.github/workflows/build.yml

@@ -55,7 +55,13 @@ jobs:
           sysctl -a
           mkdir build
           cd build
-          cmake -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL_EMBED_LIBRARY=ON -DLLAMA_CURL=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF ..
+          cmake .. \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DLLAMA_CURL=ON \
+            -DGGML_METAL_USE_BF16=ON \
+            -DGGML_METAL_EMBED_LIBRARY=ON \
+            -DGGML_RPC=ON \
+            -DBUILD_SHARED_LIBS=OFF
           cmake --build . --config Release -j $(sysctl -n hw.logicalcpu)
 
       - name: Test
@@ -113,7 +119,12 @@ jobs:
           sysctl -a
           # Metal is disabled due to intermittent failures with Github runners not having a GPU:
           # https://github.com/ggerganov/llama.cpp/actions/runs/8635935781/job/23674807267#step:5:2313
-          cmake -B build -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL=OFF -DLLAMA_CURL=ON -DGGML_RPC=ON -DBUILD_SHARED_LIBS=OFF
+          cmake -B build \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DLLAMA_CURL=ON \
+            -DGGML_METAL=OFF \
+            -DGGML_RPC=ON \
+            -DBUILD_SHARED_LIBS=OFF
           cmake --build build --config Release -j $(sysctl -n hw.logicalcpu)
 
       - name: Test
@@ -394,13 +405,13 @@ jobs:
       - name: Build with native CMake HIP support
         id: cmake_build
         run: |
-          cmake -B build -S . -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIPBLAS=ON
+          cmake -B build -S . -DCMAKE_HIP_COMPILER="$(hipconfig -l)/clang" -DGGML_HIP=ON
           cmake --build build --config Release -j $(nproc)
 
       - name: Build with legacy HIP support
         id: cmake_build_legacy_hip
         run: |
-          cmake -B build2 -S . -DCMAKE_C_COMPILER=hipcc -DCMAKE_CXX_COMPILER=hipcc -DGGML_HIPBLAS=ON
+          cmake -B build2 -S . -DCMAKE_C_COMPILER=hipcc -DCMAKE_CXX_COMPILER=hipcc -DGGML_HIP=ON
           cmake --build build2 --config Release -j $(nproc)
 
   ubuntu-22-cmake-sycl:
@@ -569,6 +580,7 @@ jobs:
           mkdir build
           cd build
           cmake -G Xcode .. \
+            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_BUILD_EXAMPLES=OFF \
             -DLLAMA_BUILD_TESTS=OFF \
@@ -599,6 +611,7 @@ jobs:
           mkdir build
           cd build
           cmake -G Xcode .. \
+            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_BUILD_EXAMPLES=OFF \
             -DLLAMA_BUILD_TESTS=OFF \
@@ -734,7 +747,7 @@ jobs:
         id: clone_kompute
         if: ${{ matrix.build == 'kompute-x64' }}
         run: |
-          git submodule update --init ggml/src/kompute
+          git submodule update --init ggml/src/ggml-kompute/kompute
 
       - name: Download OpenBLAS
         id: get_openblas
@@ -1001,7 +1014,7 @@ jobs:
         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" -DGGML_HIPBLAS=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON
+          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" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
 
   windows-latest-cmake-hip-release:
@@ -1037,7 +1050,7 @@ jobs:
         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" -DGGML_HIPBLAS=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON
+          cmake -G "Unix Makefiles" -B build -S . -DCMAKE_C_COMPILER="${env:HIP_PATH}\bin\clang.exe" -DCMAKE_CXX_COMPILER="${env:HIP_PATH}\bin\clang++.exe" -DGGML_HIP=ON -DCMAKE_BUILD_TYPE=Release -DAMDGPU_TARGETS=${{ matrix.gpu_target }} -DGGML_RPC=ON
           cmake --build build -j ${env:NUMBER_OF_PROCESSORS}
           md "build\bin\rocblas\library\"
           cp "${env:HIP_PATH}\bin\hipblas.dll" "build\bin\"

.gitmodules

@@ -1,3 +1,3 @@
 [submodule "kompute"]
-	path = ggml/src/kompute
+	path = ggml/src/ggml-kompute/kompute
 	url = https://github.com/nomic-ai/kompute.git

CMakeLists.txt

@@ -140,7 +140,6 @@ set(LLAMA_INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR} CACHE PATH "Location o
 set(LLAMA_LIB_INSTALL_DIR     ${CMAKE_INSTALL_LIBDIR}     CACHE PATH "Location of library files")
 set(LLAMA_BIN_INSTALL_DIR     ${CMAKE_INSTALL_BINDIR}     CACHE PATH "Location of binary  files")
 
-
 # At the moment some compile definitions are placed within the ggml/src
 # directory but not exported on the `ggml` target. This could be improved by
 # determining _precisely_ which defines are necessary for the llama-config