Dockerfile 11.7 KB
Newer Older
1
2
3
# SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

4
5
6
7
8
##################################
########## Build Arguments ########
##################################

# Base image configuration
9
10
11
12
13
14
ARG BASE_IMAGE="nvcr.io/nvidia/cuda-dl-base"
# TODO OPS-612: NCCL will hang with 25.03, so use 25.01 for now
# Please check https://github.com/ai-dynamo/dynamo/pull/1065
# for details and reproducer to manually test if the image
# can be updated to later versions.
ARG BASE_IMAGE_TAG="25.01-cuda12.8-devel-ubuntu24.04"
15
16

# Build configuration
17
ARG RELEASE_BUILD=false
18
ARG ENABLE_KVBM=false
19
ARG CARGO_BUILD_JOBS
20
21
22
23
24
25
26
27
28
29
30
31
32
33

# Define general architecture ARGs for supporting both x86 and aarch64 builds.
#   ARCH: Used for package suffixes (e.g., amd64, arm64)
#   ARCH_ALT: Used for Rust targets, manylinux suffix (e.g., x86_64, aarch64)
#
# Default values are for x86/amd64:
#   --build-arg ARCH=amd64 --build-arg ARCH_ALT=x86_64
#
# For arm64/aarch64, build with:
#   --build-arg ARCH=arm64 --build-arg ARCH_ALT=aarch64
#TODO OPS-592: Leverage uname -m to determine ARCH instead of passing it as an arg
ARG ARCH=amd64
ARG ARCH_ALT=x86_64

34
35
36
37
38
39
40
41
42
43
44
# SCCACHE configuration
ARG USE_SCCACHE
ARG SCCACHE_BUCKET=""
ARG SCCACHE_REGION=""

# NIXL configuration
ARG NIXL_UCX_REF=v1.19.0
ARG NIXL_REF=0.4.1

# Python configuration
ARG PYTHON_VERSION=3.12
45
46
47
48
49
50
51

##################################
########## Base Image ############
##################################

FROM ${BASE_IMAGE}:${BASE_IMAGE_TAG} AS base

52
# Redeclare ARGs for this stage
53
54
ARG ARCH
ARG ARCH_ALT
55
56
57
58
59
60
ARG PYTHON_VERSION
ARG USE_SCCACHE
ARG SCCACHE_BUCKET
ARG SCCACHE_REGION
ARG NIXL_UCX_REF
ARG NIXL_REF
61
62

USER root
63
64
65
66
67
WORKDIR /opt/dynamo

##################################
########## Tool Installation #####
##################################
68

69
# Install uv package manager
70
71
COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/

72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
# Install SCCACHE if requested
COPY container/use-sccache.sh /tmp/use-sccache.sh
RUN if [ "$USE_SCCACHE" = "true" ]; then \
        /tmp/use-sccache.sh install; \
    fi

# Set SCCACHE environment variables
ENV SCCACHE_BUCKET=${USE_SCCACHE:+${SCCACHE_BUCKET}} \
    SCCACHE_REGION=${USE_SCCACHE:+${SCCACHE_REGION}} \
    SCCACHE_S3_KEY_PREFIX=${USE_SCCACHE:+${ARCH}} \
    RUSTC_WRAPPER=${USE_SCCACHE:+sccache} \
    CMAKE_C_COMPILER_LAUNCHER=${USE_SCCACHE:+sccache} \
    CMAKE_CXX_COMPILER_LAUNCHER=${USE_SCCACHE:+sccache} \
    CMAKE_CUDA_COMPILER_LAUNCHER=${USE_SCCACHE:+sccache}

##################################
########## Rust Setup ############
##################################

91
92
93
94
# Rust environment setup
ENV RUSTUP_HOME=/usr/local/rustup \
    CARGO_HOME=/usr/local/cargo \
    PATH=/usr/local/cargo/bin:$PATH \
95
    RUST_VERSION=1.89.0
96
97
98
99

# Define Rust target based on ARCH_ALT ARG
ARG RUSTARCH=${ARCH_ALT}-unknown-linux-gnu

100
# Install Rust
101
102
103
104
105
106
RUN wget --tries=3 --waitretry=5 "https://static.rust-lang.org/rustup/archive/1.28.1/${RUSTARCH}/rustup-init" && \
    chmod +x rustup-init && \
    ./rustup-init -y --no-modify-path --profile minimal --default-toolchain $RUST_VERSION --default-host ${RUSTARCH} && \
    rm rustup-init && \
    chmod -R a+w $RUSTUP_HOME $CARGO_HOME

107
108
109
110
111
##################################
########## System Dependencies ###
##################################

# Install system packages
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
RUN apt-get update -y \
    && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
        # NIXL build dependencies
        autoconf \
        automake \
        cmake \
        git \
        libtool \
        meson \
        net-tools \
        ninja-build \
        pybind11-dev \
        # These headers are missing with the hpcx installer, required
        # by UCX to find RDMA devices
        ibverbs-providers \
        ibverbs-utils \
        libibumad-dev \
        libibverbs-dev \
        librdmacm-dev \
        libnuma-dev \
        rdma-core \
        # Rust build dependencies
        clang \
        libclang-dev \
        protobuf-compiler \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

140
141
142
143
144
##################################
########## External Services #####
##################################

# Install NATS server
145
146
147
148
149
ENV NATS_VERSION="v2.10.28"
RUN --mount=type=cache,target=/var/cache/apt \
    wget --tries=3 --waitretry=5 https://github.com/nats-io/nats-server/releases/download/${NATS_VERSION}/nats-server-${NATS_VERSION}-${ARCH}.deb && \
    dpkg -i nats-server-${NATS_VERSION}-${ARCH}.deb && rm nats-server-${NATS_VERSION}-${ARCH}.deb

150
# Install etcd
151
152
153
154
155
156
157
ENV ETCD_VERSION="v3.5.21"
RUN wget --tries=3 --waitretry=5 https://github.com/etcd-io/etcd/releases/download/$ETCD_VERSION/etcd-$ETCD_VERSION-linux-${ARCH}.tar.gz -O /tmp/etcd.tar.gz && \
    mkdir -p /usr/local/bin/etcd && \
    tar -xvf /tmp/etcd.tar.gz -C /usr/local/bin/etcd --strip-components=1 && \
    rm /tmp/etcd.tar.gz
ENV PATH=/usr/local/bin/etcd/:$PATH

158
159
160
161
162
##################################
########## UCX Build #############
##################################

# Build and install UCX
163
164
165
166
RUN rm -rf /opt/hpcx/ucx && \
    rm -rf /usr/local/ucx && \
    echo "Building UCX with reference $NIXL_UCX_REF" && \
    cd /usr/local/src && \
167
168
    git clone https://github.com/openucx/ucx.git && \
    cd ucx && git checkout $NIXL_UCX_REF && \
169
170
171
172
    CC=${USE_SCCACHE:+sccache gcc} && \
    CXX=${USE_SCCACHE:+sccache g++} && \
    export CC=${CC} && \
    export CXX=${CXX} && \
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
    ./autogen.sh && \
    ./configure \
      --prefix=/usr/local/ucx \
      --enable-shared \
      --disable-static \
      --disable-doxygen-doc \
      --enable-optimizations \
      --enable-cma \
      --enable-devel-headers \
      --with-cuda=/usr/local/cuda \
      --with-verbs \
      --with-efa \
      --with-dm \
      --with-gdrcopy=/usr/local \
      --enable-mt && \
    make -j$(nproc) &&                              \
    make -j$(nproc) install-strip &&                \
190
    /tmp/use-sccache.sh show-stats "UCX" && \
191
192
193
194
195
196
197
198
199
200
201
    echo "/usr/local/ucx/lib" > /etc/ld.so.conf.d/ucx.conf && \
    echo "/usr/local/ucx/lib/ucx" >> /etc/ld.so.conf.d/ucx.conf && \
    ldconfig && \
    cd /usr/local/src && \
    rm -rf ucx

# UCX environment variables
ENV CPATH=/usr/include:$CPATH \
    PATH=/usr/bin:$PATH \
    PKG_CONFIG_PATH=/usr/lib/pkgconfig:$PKG_CONFIG_PATH

202
203
204
205
206
207
208
209
210
211
212
##################################
########## NIXL Setup ############
##################################

# NIXL environment setup
ENV NIXL_SRC_DIR=/opt/nixl \
    NIXL_PREFIX=/opt/nvidia/nvda_nixl \
    NIXL_LIB_DIR=/opt/nvidia/nvda_nixl/lib/${ARCH_ALT}-linux-gnu \
    NIXL_PLUGIN_DIR=/opt/nvidia/nvda_nixl/lib/${ARCH_ALT}-linux-gnu/plugins

# Build and install NIXL
213
214
215
216
217
218
219
220
RUN git clone --depth 1 --branch ${NIXL_REF} "https://github.com/ai-dynamo/nixl.git" ${NIXL_SRC_DIR} && \
    cd ${NIXL_SRC_DIR} && \
    if [ "$ARCH" = "arm64" ]; then \
        nixl_build_args="-Ddisable_gds_backend=true"; \
    else \
        nixl_build_args=""; \
    fi && \
    meson setup build/ --buildtype=release --prefix=$NIXL_PREFIX $nixl_build_args && \
221
222
    ninja -C build/ -j$(nproc) && ninja -C build/ install && \
    /tmp/use-sccache.sh show-stats "NIXL" && \
223
224
225
226
    echo "$NIXL_LIB_DIR" > /etc/ld.so.conf.d/nixl.conf && \
    echo "$NIXL_PLUGIN_DIR" >> /etc/ld.so.conf.d/nixl.conf && \
    ldconfig

227
# Build NIXL Python module
228
229
230
231
232
233
234
235
# TODO OPS-590: Move gds_path selection based on arch into NIXL build and re-enable gds backend for arm64
RUN if [ "$ARCH" = "arm64" ]; then \
        cd ${NIXL_SRC_DIR} && uv build . --out-dir /opt/dynamo/wheelhouse/nixl \
        --config-settings=setup-args="-Ddisable_gds_backend=true"; \
    else \
        cd ${NIXL_SRC_DIR} && uv build . --out-dir /opt/dynamo/wheelhouse/nixl; \
    fi

236
237
238
239
240
241
##################################
########## Python Environment ####
##################################

# Create and activate virtual environment
ARG PYTHON_VERSION
242
RUN mkdir -p /opt/dynamo/venv && \
243
    uv venv /opt/dynamo/venv --python $PYTHON_VERSION
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261

ENV VIRTUAL_ENV=/opt/dynamo/venv \
    PATH="/opt/dynamo/venv/bin:${PATH}"

# Install common and test dependencies
RUN --mount=type=bind,source=./container/deps/requirements.txt,target=/tmp/requirements.txt \
    --mount=type=bind,source=./container/deps/requirements.test.txt,target=/tmp/requirements.test.txt \
    uv pip install --requirement /tmp/requirements.txt --requirement /tmp/requirements.test.txt

##################################
##### Wheel Build Image ##########
##################################

# Redeclare ARCH_ALT ARG so it's available for interpolation in the FROM instruction
ARG ARCH_ALT

FROM quay.io/pypa/manylinux_2_28_${ARCH_ALT} AS wheel_builder

262
263
264
# Redeclare ARGs for this stage
ARG ARCH
ARG ARCH_ALT
265
266
ARG CARGO_BUILD_JOBS
ARG RELEASE_BUILD
267
ARG ENABLE_KVBM
268
269
270
ARG USE_SCCACHE
ARG SCCACHE_BUCKET
ARG SCCACHE_REGION
271
272
273

WORKDIR /opt/dynamo

274
275
276
# Set environment variables
ENV CARGO_BUILD_JOBS=${CARGO_BUILD_JOBS:-16} \
    RUSTUP_HOME=/usr/local/rustup \
277
278
279
    CARGO_HOME=/usr/local/cargo \
    CARGO_TARGET_DIR=/opt/dynamo/target \
    VIRTUAL_ENV=/opt/dynamo/venv \
280
281
282
283
284
285
286
287
288
    NIXL_PREFIX=/opt/nvidia/nvda_nixl \
    PATH=/usr/local/cargo/bin:/opt/dynamo/venv/bin:$PATH

# Install system dependencies
ARG PYTHON_VERSION
RUN dnf update -y \
    && dnf install -y llvm-toolset protobuf-compiler python${PYTHON_VERSION}-devel wget \
    && dnf clean all \
    && rm -rf /var/cache/dnf
289

290
# Copy artifacts from base stage
291
292
293
294
295
COPY --from=base $RUSTUP_HOME $RUSTUP_HOME
COPY --from=base $CARGO_HOME $CARGO_HOME
COPY --from=base $NIXL_PREFIX $NIXL_PREFIX
COPY --from=base $VIRTUAL_ENV $VIRTUAL_ENV

296
297
298
299
300
301
302
303
304
305
306
# Install SCCACHE if requested
COPY container/use-sccache.sh /tmp/use-sccache.sh
RUN if [ "$USE_SCCACHE" = "true" ]; then \
        /tmp/use-sccache.sh install; \
    fi

# Set SCCACHE environment variables
ENV SCCACHE_BUCKET=${USE_SCCACHE:+${SCCACHE_BUCKET}} \
    SCCACHE_REGION=${USE_SCCACHE:+${SCCACHE_REGION}} \
    SCCACHE_S3_KEY_PREFIX=${USE_SCCACHE:+${ARCH}} \
    RUSTC_WRAPPER=${USE_SCCACHE:+sccache}
307

308
309
# Copy source code (order matters for layer caching)
COPY pyproject.toml README.md LICENSE Cargo.toml Cargo.lock rust-toolchain.toml hatch_build.py /opt/dynamo/
310
311
312
COPY lib/ /opt/dynamo/lib/
COPY components/ /opt/dynamo/components/

313
# Build wheels
314
315
316
RUN uv build --wheel --out-dir /opt/dynamo/dist && \
    cd /opt/dynamo/lib/bindings/python && \
    uv pip install maturin[patchelf] && \
317
318
319
320
321
    if [ "$ENABLE_KVBM" = "true" ]; then \
        maturin build --release --features block-manager --out /opt/dynamo/dist; \
    else \
        maturin build --release --out /opt/dynamo/dist; \
    fi && \
322
323
324
    if [ "$RELEASE_BUILD" = "true" ]; then \
        uv run --python 3.11 maturin build --release --out /opt/dynamo/dist && \
        uv run --python 3.10 maturin build --release --out /opt/dynamo/dist; \
325
326
    fi && \
    /tmp/use-sccache.sh show-stats "Dynamo"
327
328
329
330

##############################################
########## Dev entrypoint image ##############
##############################################
331

332
333
334
335
336
337
338
339
340
FROM base AS dev

# Application environment variables
ENV DYNAMO_HOME=/opt/dynamo \
    CARGO_TARGET_DIR=/opt/dynamo/target \
    PYTHONPATH=/opt/dynamo:$PYTHONPATH

WORKDIR /opt/dynamo

341
# Copy built artifacts
342
343
344
345
COPY --from=wheel_builder /opt/dynamo/dist/*.whl /opt/dynamo/wheelhouse/
COPY --from=wheel_builder $CARGO_TARGET_DIR $CARGO_TARGET_DIR
COPY --from=wheel_builder $CARGO_HOME $CARGO_HOME

346
# Install Python packages
347
348
349
350
351
352
353
354
COPY benchmarks/ /opt/dynamo/benchmarks/
RUN uv pip install \
    /opt/dynamo/wheelhouse/ai_dynamo_runtime*cp312*.whl \
    /opt/dynamo/wheelhouse/ai_dynamo*any.whl \
    /opt/dynamo/wheelhouse/nixl/nixl*.whl \
    /opt/dynamo/benchmarks && \
    rm -rf /opt/dynamo/benchmarks

355
# Setup launch banner
356
357
358
359
360
RUN --mount=type=bind,source=./container/launch_message.txt,target=/opt/dynamo/launch_message.txt \
    sed '/^#\s/d' /opt/dynamo/launch_message.txt > ~/.launch_screen && \
    echo "cat ~/.launch_screen" >> ~/.bashrc

ENTRYPOINT ["/opt/nvidia/nvidia_entrypoint.sh"]
361
CMD []