baalajimaestro/ollama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Merge pull request #3708 from remy415/arm64static

Browse source 
move Ollama static build to its own flag
This commit is contained in:
Daniel Hiltgen 2024-04-18 16:04:12 -07:00 committed by GitHub
commit 8d1995c625
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 19 additions and 19 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
Dockerfile
View file

@ -18,7 +18,7 @@ ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
COPY --from=llm-code / /go/src/github.com/ollama/ollama/ COPY --from=llm-code / /go/src/github.com/ollama/ollama/
WORKDIR /go/src/github.com/ollama/ollama/llm/generate WORKDIR /go/src/github.com/ollama/ollama/llm/generate
ARG CGO_CFLAGS ARG CGO_CFLAGS
RUN OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION-devel-rockylinux8 AS cuda-build-arm64 FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION-devel-rockylinux8 AS cuda-build-arm64
ARG CMAKE_VERSION ARG CMAKE_VERSION
@ -28,7 +28,7 @@ ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
COPY --from=llm-code / /go/src/github.com/ollama/ollama/ COPY --from=llm-code / /go/src/github.com/ollama/ollama/
WORKDIR /go/src/github.com/ollama/ollama/llm/generate WORKDIR /go/src/github.com/ollama/ollama/llm/generate
ARG CGO_CFLAGS ARG CGO_CFLAGS
RUN OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete AS rocm-build-amd64 FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete AS rocm-build-amd64
ARG CMAKE_VERSION ARG CMAKE_VERSION
@ -40,7 +40,7 @@ COPY --from=llm-code / /go/src/github.com/ollama/ollama/
WORKDIR /go/src/github.com/ollama/ollama/llm/generate WORKDIR /go/src/github.com/ollama/ollama/llm/generate
ARG CGO_CFLAGS ARG CGO_CFLAGS
ARG AMDGPU_TARGETS ARG AMDGPU_TARGETS
RUN OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
RUN mkdir /tmp/scratch && \ RUN mkdir /tmp/scratch && \
for dep in $(zcat /go/src/github.com/ollama/ollama/llm/build/linux/x86_64/rocm*/bin/deps.txt.gz) ; do \ for dep in $(zcat /go/src/github.com/ollama/ollama/llm/build/linux/x86_64/rocm*/bin/deps.txt.gz) ; do \
cp ${dep} /tmp/scratch/ || exit 1 ; \ cp ${dep} /tmp/scratch/ || exit 1 ; \
@ -64,11 +64,11 @@ WORKDIR /go/src/github.com/ollama/ollama/llm/generate
FROM --platform=linux/amd64 cpu-builder-amd64 AS static-build-amd64 FROM --platform=linux/amd64 cpu-builder-amd64 AS static-build-amd64
RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh
FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu-build-amd64 FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu-build-amd64
RUN OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh
FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx-build-amd64 FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx-build-amd64
RUN OLLAMA_CPU_TARGET="cpu_avx" sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx" sh gen_linux.sh
FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx2-build-amd64 FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx2-build-amd64
RUN OLLAMA_CPU_TARGET="cpu_avx2" sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx2" sh gen_linux.sh
FROM --platform=linux/arm64 centos:7 AS cpu-builder-arm64 FROM --platform=linux/arm64 centos:7 AS cpu-builder-arm64
ARG CMAKE_VERSION ARG CMAKE_VERSION
@ -84,7 +84,7 @@ WORKDIR /go/src/github.com/ollama/ollama/llm/generate
FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64 FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64
RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh
FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64 FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64
RUN OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh
# Intermediate stage used for ./scripts/build_linux.sh # Intermediate stage used for ./scripts/build_linux.sh

24
llm/generate/gen_linux.sh
View file

@ -57,21 +57,21 @@ init_vars
git_module_setup git_module_setup
apply_patches apply_patches
init_vars
if [ -z "${OLLAMA_SKIP_STATIC_GENERATE}" -o "${OLLAMA_CPU_TARGET}" = "static" ]; then
# Builds by default, allows skipping, forces build if OLLAMA_CPU_TARGET="static"
# Enables optimized Dockerfile builds using a blanket skip and targeted overrides
# Static build for linking into the Go binary
init_vars
CMAKE_TARGETS="--target llama --target ggml"
CMAKE_DEFS="-DBUILD_SHARED_LIBS=off -DLLAMA_NATIVE=off -DLLAMA_AVX=off -DLLAMA_AVX2=off -DLLAMA_AVX512=off -DLLAMA_FMA=off -DLLAMA_F16C=off ${CMAKE_DEFS}"
BUILD_DIR="../build/linux/${ARCH}_static"
echo "Building static library"
build
fi
init_vars init_vars
if [ -z "${OLLAMA_SKIP_CPU_GENERATE}" ]; then if [ -z "${OLLAMA_SKIP_CPU_GENERATE}" ]; then
if [ -z "${OLLAMA_CPU_TARGET}" -o "${OLLAMA_CPU_TARGET}" = "static" ]; then
# Static build for linking into the Go binary
init_vars
CMAKE_TARGETS="--target llama --target ggml"
CMAKE_DEFS="-DBUILD_SHARED_LIBS=off -DLLAMA_NATIVE=off -DLLAMA_AVX=off -DLLAMA_AVX2=off -DLLAMA_AVX512=off -DLLAMA_FMA=off -DLLAMA_F16C=off ${CMAKE_DEFS}"
BUILD_DIR="../build/linux/${ARCH}_static"
echo "Building static library"
build
fi
# Users building from source can tune the exact flags we pass to cmake for configuring # Users building from source can tune the exact flags we pass to cmake for configuring
# llama.cpp, and we'll build only 1 CPU variant in that case as the default. # llama.cpp, and we'll build only 1 CPU variant in that case as the default.
if [ -n "${OLLAMA_CUSTOM_CPU_DEFS}" ]; then if [ -n "${OLLAMA_CUSTOM_CPU_DEFS}" ]; then