Changed to bild for rx580 GPU, this uses 5.7.1 rocm libraries
This commit is contained in:
Matt
parent
56318fb365
commit
3449201ce4
138
Dockerfile
138
Dockerfile
|
|
@ -4,7 +4,7 @@ ARG CUDA_VERSION_11=11.3.1
|
|||
ARG CUDA_V11_ARCHITECTURES="50;52;53;60;61;62;70;72;75;80;86"
|
||||
ARG CUDA_VERSION_12=12.4.0
|
||||
ARG CUDA_V12_ARCHITECTURES="60;61;62;70;72;75;80;86;87;89;90;90a"
|
||||
ARG ROCM_VERSION=6.1.2
|
||||
ARG ROCM_VERSION=5.7.1
|
||||
|
||||
# Copy the minimal context we need to run the generate scripts
|
||||
FROM scratch AS llm-code
|
||||
|
|
@ -47,39 +47,39 @@ RUN --mount=type=cache,target=/root/.ccache \
|
|||
OLLAMA_CUSTOM_CUDA_DEFS="-DGGML_CUDA_USE_GRAPHS=on" \
|
||||
bash gen_linux.sh
|
||||
|
||||
FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_11-devel-rockylinux8 AS cuda-11-build-server-arm64
|
||||
ARG CMAKE_VERSION
|
||||
COPY ./scripts/rh_linux_deps.sh /
|
||||
RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
|
||||
ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
ARG CGO_CFLAGS
|
||||
ARG CUDA_V11_ARCHITECTURES
|
||||
ENV GOARCH arm64
|
||||
RUN OLLAMA_SKIP_STATIC_GENERATE=1 \
|
||||
OLLAMA_SKIP_CPU_GENERATE=1 \
|
||||
CMAKE_CUDA_ARCHITECTURES="${CUDA_V11_ARCHITECTURES}" \
|
||||
CUDA_VARIANT="_v11" \
|
||||
bash gen_linux.sh
|
||||
#FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_11-devel-rockylinux8 AS cuda-11-build-server-arm64
|
||||
#ARG CMAKE_VERSION
|
||||
#COPY ./scripts/rh_linux_deps.sh /
|
||||
#RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
|
||||
#ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
#COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
#WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
#ARG CGO_CFLAGS
|
||||
#ARG CUDA_V11_ARCHITECTURES
|
||||
#ENV GOARCH arm64
|
||||
#RUN OLLAMA_SKIP_STATIC_GENERATE=1 \
|
||||
# OLLAMA_SKIP_CPU_GENERATE=1 \
|
||||
# CMAKE_CUDA_ARCHITECTURES="${CUDA_V11_ARCHITECTURES}" \
|
||||
# CUDA_VARIANT="_v11" \
|
||||
# bash gen_linux.sh
|
||||
|
||||
FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_12-devel-rockylinux8 AS cuda-12-build-server-arm64
|
||||
ARG CMAKE_VERSION
|
||||
COPY ./scripts/rh_linux_deps.sh /
|
||||
RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
|
||||
ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
ARG CGO_CFLAGS
|
||||
ARG CUDA_V12_ARCHITECTURES
|
||||
ENV GOARCH arm64
|
||||
RUN --mount=type=cache,target=/root/.ccache \
|
||||
OLLAMA_SKIP_STATIC_GENERATE=1 \
|
||||
OLLAMA_SKIP_CPU_GENERATE=1 \
|
||||
CMAKE_CUDA_ARCHITECTURES="${CUDA_V12_ARCHITECTURES}" \
|
||||
CUDA_VARIANT="_v12" \
|
||||
OLLAMA_CUSTOM_CUDA_DEFS="-DGGML_CUDA_USE_GRAPHS=on" \
|
||||
bash gen_linux.sh
|
||||
#FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_12-devel-rockylinux8 AS cuda-12-build-server-arm64
|
||||
#ARG CMAKE_VERSION
|
||||
#COPY ./scripts/rh_linux_deps.sh /
|
||||
#RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
|
||||
#ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
#COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
#WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
#ARG CGO_CFLAGS
|
||||
#ARG CUDA_V12_ARCHITECTURES
|
||||
#ENV GOARCH arm64
|
||||
#RUN --mount=type=cache,target=/root/.ccache \
|
||||
# OLLAMA_SKIP_STATIC_GENERATE=1 \
|
||||
# OLLAMA_SKIP_CPU_GENERATE=1 \
|
||||
# CMAKE_CUDA_ARCHITECTURES="${CUDA_V12_ARCHITECTURES}" \
|
||||
# CUDA_VARIANT="_v12" \
|
||||
# OLLAMA_CUSTOM_CUDA_DEFS="-DGGML_CUDA_USE_GRAPHS=on" \
|
||||
# bash gen_linux.sh
|
||||
|
||||
|
||||
FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete AS rocm-build-amd64
|
||||
|
|
@ -123,24 +123,24 @@ FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx2-build-amd64
|
|||
RUN --mount=type=cache,target=/root/.ccache \
|
||||
OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx2" bash gen_linux.sh
|
||||
|
||||
FROM --platform=linux/arm64 rockylinux:8 AS cpu-builder-arm64
|
||||
ARG CMAKE_VERSION
|
||||
ARG GOLANG_VERSION
|
||||
COPY ./scripts/rh_linux_deps.sh /
|
||||
RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
|
||||
ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
ARG OLLAMA_CUSTOM_CPU_DEFS
|
||||
ARG CGO_CFLAGS
|
||||
ENV GOARCH arm64
|
||||
WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
#FROM --platform=linux/arm64 rockylinux:8 AS cpu-builder-arm64
|
||||
#ARG CMAKE_VERSION
|
||||
#ARG GOLANG_VERSION
|
||||
#COPY ./scripts/rh_linux_deps.sh /
|
||||
#RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
|
||||
#ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
|
||||
#COPY --from=llm-code / /go/src/github.com/ollama/ollama/
|
||||
#ARG OLLAMA_CUSTOM_CPU_DEFS
|
||||
#ARG CGO_CFLAGS
|
||||
#ENV GOARCH arm64
|
||||
#WORKDIR /go/src/github.com/ollama/ollama/llm/generate
|
||||
|
||||
FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64
|
||||
RUN --mount=type=cache,target=/root/.ccache \
|
||||
OLLAMA_CPU_TARGET="static" bash gen_linux.sh
|
||||
FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64
|
||||
RUN --mount=type=cache,target=/root/.ccache \
|
||||
OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" bash gen_linux.sh
|
||||
#FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64
|
||||
#RUN --mount=type=cache,target=/root/.ccache \
|
||||
# OLLAMA_CPU_TARGET="static" bash gen_linux.sh
|
||||
#FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64
|
||||
#RUN --mount=type=cache,target=/root/.ccache \
|
||||
# OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" bash gen_linux.sh
|
||||
|
||||
|
||||
# Intermediate stage used for ./scripts/build_linux.sh
|
||||
|
|
@ -163,20 +163,20 @@ RUN --mount=type=cache,target=/root/.ccache \
|
|||
go build -trimpath -o dist/linux-amd64/bin/ollama .
|
||||
|
||||
# Intermediate stage used for ./scripts/build_linux.sh
|
||||
FROM --platform=linux/arm64 cpu-build-arm64 AS build-arm64
|
||||
ENV CGO_ENABLED 1
|
||||
ARG GOLANG_VERSION
|
||||
WORKDIR /go/src/github.com/ollama/ollama
|
||||
COPY . .
|
||||
COPY --from=static-build-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
|
||||
COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
|
||||
COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
ARG GOFLAGS
|
||||
ARG CGO_CFLAGS
|
||||
RUN --mount=type=cache,target=/root/.ccache \
|
||||
go build -trimpath -o dist/linux-arm64/bin/ollama .
|
||||
#FROM --platform=linux/arm64 cpu-build-arm64 AS build-arm64
|
||||
#ENV CGO_ENABLED 1
|
||||
#ARG GOLANG_VERSION
|
||||
#WORKDIR /go/src/github.com/ollama/ollama
|
||||
#COPY . .
|
||||
#COPY --from=static-build-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
#COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
|
||||
#COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
#COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
|
||||
#COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
|
||||
#ARG GOFLAGS
|
||||
#ARG CGO_CFLAGS
|
||||
#RUN --mount=type=cache,target=/root/.ccache \
|
||||
# go build -trimpath -o dist/linux-arm64/bin/ollama .
|
||||
|
||||
# Strip out ROCm dependencies to keep the primary image lean
|
||||
FROM --platform=linux/amd64 ubuntu:22.04 as amd64-libs-without-rocm
|
||||
|
|
@ -190,11 +190,11 @@ RUN apt-get update && apt-get install -y ca-certificates && \
|
|||
apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
COPY --from=build-amd64 /go/src/github.com/ollama/ollama/dist/linux-amd64/bin/ /bin/
|
||||
|
||||
FROM --platform=linux/arm64 ubuntu:22.04 as runtime-arm64
|
||||
COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/lib/ /lib/
|
||||
RUN apt-get update && apt-get install -y ca-certificates && \
|
||||
apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/bin/ /bin/
|
||||
#FROM --platform=linux/arm64 ubuntu:22.04 as runtime-arm64
|
||||
#COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/lib/ /lib/
|
||||
#RUN apt-get update && apt-get install -y ca-certificates && \
|
||||
# apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
#COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/bin/ /bin/
|
||||
|
||||
# Radeon images are much larger so we keep it distinct from the CPU/CUDA image
|
||||
FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete as runtime-rocm
|
||||
|
|
|
|||
|
|
@ -41,7 +41,7 @@ const (
|
|||
|
||||
var (
|
||||
// Used to validate if the given ROCm lib is usable
|
||||
ROCmLibGlobs = []string{"libhipblas.so.2*", "rocblas"} // TODO - probably include more coverage of files here...
|
||||
ROCmLibGlobs = []string{"libhipblas.so.1*", "rocblas"} // TODO - probably include more coverage of files here...
|
||||
RocmStandardLocations = []string{"/opt/rocm/lib", "/usr/lib64"}
|
||||
)
|
||||
|
||||
|
|
@ -386,7 +386,7 @@ func AMDValidateLibDir() (string, error) {
|
|||
}
|
||||
|
||||
// Well known ollama installer path
|
||||
installedRocmDir := "/usr/share/ollama/lib/rocm"
|
||||
installedRocmDir := "/opt/rocm-5.7.1"
|
||||
if rocmLibUsable(installedRocmDir) {
|
||||
return installedRocmDir, nil
|
||||
}
|
||||
|
|
|
|||
|
|
@ -59,7 +59,7 @@ var (
|
|||
// With our current CUDA compile flags, older than 5.0 will not work properly
|
||||
var CudaComputeMin = [2]C.int{5, 0}
|
||||
|
||||
var RocmComputeMin = 9
|
||||
var RocmComputeMin = 8
|
||||
|
||||
// TODO find a better way to detect iGPU instead of minimum memory
|
||||
const IGPUMemLimit = 1 * format.GibiByte // 512G is what they typically report, so anything less than 1G must be iGPU
|
||||
|
|
|
|||
|
|
@ -22,6 +22,7 @@ amdGPUs() {
|
|||
return
|
||||
fi
|
||||
GPU_LIST=(
|
||||
"gfx803"
|
||||
"gfx900"
|
||||
"gfx906:xnack-"
|
||||
"gfx908:xnack-"
|
||||
|
|
|
|||
|
|
@ -13,7 +13,7 @@ DOCKER_ORG=${DOCKER_ORG:-"ollama"}
|
|||
RELEASE_IMAGE_REPO=${RELEASE_IMAGE_REPO:-"${DOCKER_ORG}/release"}
|
||||
FINAL_IMAGE_REPO=${FINAL_IMAGE_REPO:-"${DOCKER_ORG}/ollama"}
|
||||
|
||||
BUILD_ARCH=${BUILD_ARCH:-"amd64 arm64"}
|
||||
BUILD_ARCH=${BUILD_ARCH:-"amd64"}
|
||||
|
||||
# Set PUSH to a non-empty string to trigger push instead of load
|
||||
PUSH=${PUSH:-""}
|
||||
|
|
|
|||
|
|
@ -6,7 +6,8 @@ export VERSION=${VERSION:-$(git describe --tags --first-parent --abbrev=7 --long
|
|||
export GOFLAGS="'-ldflags=-w -s \"-X=github.com/ollama/ollama/version.Version=$VERSION\" \"-X=github.com/ollama/ollama/server.mode=release\"'"
|
||||
GZIP=$(which pigz 2>/dev/null || echo "gzip")
|
||||
|
||||
BUILD_ARCH=${BUILD_ARCH:-"amd64 arm64"}
|
||||
#BUILD_ARCH=${BUILD_ARCH:-"amd64 arm64"}
|
||||
BUILD_ARCH=${BUILD_ARCH:-"amd64"}
|
||||
export AMDGPU_TARGETS=${AMDGPU_TARGETS:=""}
|
||||
mkdir -p dist
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue