Add support for Blackwell architecture

[msharara1998]

Feature request

I tried latest cuda image (cuda-latest), but did not work on my RTX 5090. Here are the container logs:

cuda compute cap 120 is not supported

Motivation

Using text embedding inference server on RTX 5090 GPUs and other types having compute capability of 120 and Blackwell architecture

Your contribution

I tried editing the Dockerfile.cuda with by changing nvidia/cuda to the latest image: nvidia/cuda:12.9.1-cudnn-devel-ubuntu22.04
and by editing candle installation to support cuda compute cap of 120. The build was successful, but I got an error with candle upon running the container:

2025-06-24T07:44:11.669953Z  INFO text_embeddings_router: router/src/lib.rs:235: Starting model backend
2025-06-24T07:44:17.135967Z ERROR text_embeddings_backend: backends/src/lib.rs:388: Could not start Candle backend: Could not start backend: Runtime compute cap 120 is not compatible with compile time compute cap 120
Error: Could not create backend

Here is my modified Dockerfile.cuda:

FROM nvidia/cuda:12.9.1-cudnn-devel-ubuntu22.04 AS base-builder

ENV SCCACHE=0.10.0
ENV RUSTC_WRAPPER=/usr/local/bin/sccache
ENV PATH="/root/.cargo/bin:${PATH}"
# aligned with `cargo-chef` version in `lukemathwalker/cargo-chef:latest-rust-1.85-bookworm`
ENV CARGO_CHEF=0.1.71

RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
    curl \
    libssl-dev \
    pkg-config \
    && rm -rf /var/lib/apt/lists/*

# Donwload and configure sccache
RUN curl -fsSL https://github.com/mozilla/sccache/releases/download/v$SCCACHE/sccache-v$SCCACHE-x86_64-unknown-linux-musl.tar.gz | tar -xzv --strip-components=1 -C /usr/local/bin sccache-v$SCCACHE-x86_64-unknown-linux-musl/sccache && \
    chmod +x /usr/local/bin/sccache

RUN curl https://sh.rustup.rs -sSf | bash -s -- -y
RUN cargo install cargo-chef --version $CARGO_CHEF --locked

FROM base-builder AS planner

WORKDIR /usr/src

COPY backends backends
COPY core core
COPY router router
COPY Cargo.toml ./
COPY Cargo.lock ./

RUN cargo chef prepare  --recipe-path recipe.json

FROM base-builder AS builder

ARG CUDA_COMPUTE_CAP=120
ARG GIT_SHA
ARG DOCKER_LABEL

# Limit parallelism
ARG RAYON_NUM_THREADS
ARG CARGO_BUILD_JOBS
ARG CARGO_BUILD_INCREMENTAL

# sccache specific variables
ARG SCCACHE_GHA_ENABLED

WORKDIR /usr/src

RUN --mount=type=secret,id=actions_results_url,env=ACTIONS_RESULTS_URL \
    --mount=type=secret,id=actions_runtime_token,env=ACTIONS_RUNTIME_TOKEN \
    if [ ${CUDA_COMPUTE_CAP} -ge 75 -a ${CUDA_COMPUTE_CAP} -lt 80 ]; \
    then  \
        nvprune --generate-code code=sm_${CUDA_COMPUTE_CAP} /usr/local/cuda/lib64/libcublas_static.a -o /usr/local/cuda/lib64/libcublas_static.a; \
    elif [ ${CUDA_COMPUTE_CAP} -ge 80 -a ${CUDA_COMPUTE_CAP} -lt 90 ]; \
    then  \
        nvprune --generate-code code=sm_80 --generate-code code=sm_${CUDA_COMPUTE_CAP} /usr/local/cuda/lib64/libcublas_static.a -o /usr/local/cuda/lib64/libcublas_static.a; \
    elif [ ${CUDA_COMPUTE_CAP} -eq 90 ]; \
    then  \
        nvprune --generate-code code=sm_90 /usr/local/cuda/lib64/libcublas_static.a -o /usr/local/cuda/lib64/libcublas_static.a; \
    elif [ ${CUDA_COMPUTE_CAP} -eq 120 ]; \
    then  \
        nvprune --generate-code code=sm_${CUDA_COMPUTE_CAP} /usr/local/cuda/lib64/libcublas_static.a -o /usr/local/cuda/lib64/libcublas_static.a; \ 
    else  \
        echo "cuda compute cap ${CUDA_COMPUTE_CAP} is not supported"; exit 1; \
    fi;

COPY --from=planner /usr/src/recipe.json recipe.json

RUN --mount=type=secret,id=actions_results_url,env=ACTIONS_RESULTS_URL \
    --mount=type=secret,id=actions_runtime_token,env=ACTIONS_RUNTIME_TOKEN \
    if [ ${CUDA_COMPUTE_CAP} -ge 75 -a ${CUDA_COMPUTE_CAP} -lt 80 ]; \
    then \
        cargo chef cook --release --features candle-cuda-turing --features static-linking --no-default-features --recipe-path recipe.json && sccache -s; \
    elif [ ${CUDA_COMPUTE_CAP} -eq 120 ]; \
    then \
        cargo chef cook --release --features candle-cuda-turing --features static-linking --no-default-features --recipe-path recipe.json && sccache -s; \
    else \
        cargo chef cook --release --features candle-cuda --features static-linking --no-default-features --recipe-path recipe.json && sccache -s; \
    fi;

COPY backends backends
COPY core core
COPY router router
COPY Cargo.toml ./
COPY Cargo.lock ./

FROM builder AS http-builder

RUN --mount=type=secret,id=actions_results_url,env=ACTIONS_RESULTS_URL \
    --mount=type=secret,id=actions_runtime_token,env=ACTIONS_RUNTIME_TOKEN \
    if [ ${CUDA_COMPUTE_CAP} -ge 75 -a ${CUDA_COMPUTE_CAP} -lt 80 ]; \
    then \
        cargo build --release --bin text-embeddings-router -F candle-cuda-turing -F static-linking -F http --no-default-features && sccache -s; \
    elif [ ${CUDA_COMPUTE_CAP} -eq 120 ]; \
    then \
        cargo build --release --bin text-embeddings-router -F candle-cuda-turing -F static-linking -F http --no-default-features && sccache -s; \
    else \
        cargo build --release --bin text-embeddings-router -F candle-cuda -F static-linking -F http --no-default-features && sccache -s; \
    fi;

FROM builder AS grpc-builder

RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
    unzip \
    && rm -rf /var/lib/apt/lists/*

RUN PROTOC_ZIP=protoc-21.12-linux-x86_64.zip && \
    curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v21.12/$PROTOC_ZIP && \
    unzip -o $PROTOC_ZIP -d /usr/local bin/protoc && \
    unzip -o $PROTOC_ZIP -d /usr/local 'include/*' && \
    rm -f $PROTOC_ZIP

COPY proto proto

RUN --mount=type=secret,id=actions_results_url,env=ACTIONS_RESULTS_URL \
    --mount=type=secret,id=actions_runtime_token,env=ACTIONS_RUNTIME_TOKEN \
    if [ ${CUDA_COMPUTE_CAP} -ge 75 -a ${CUDA_COMPUTE_CAP} -lt 80 ]; \
    then \
        cargo build --release --bin text-embeddings-router -F candle-cuda-turing -F static-linking -F grpc --no-default-features && sccache -s; \
    elif [ ${CUDA_COMPUTE_CAP} -eq 120 ]; \
    then \
        cargo build --release --bin text-embeddings-router -F candle-cuda-turing -F static-linking -F grpc --no-default-features && sccache -s; \
    else \
        cargo build --release --bin text-embeddings-router -F candle-cuda -F static-linking -F grpc --no-default-features && sccache -s; \
    fi;

FROM nvidia/cuda:12.9.1-cudnn-devel-ubuntu22.04 AS base

ARG DEFAULT_USE_FLASH_ATTENTION=True

ENV HUGGINGFACE_HUB_CACHE=/data \
    PORT=80 \
    USE_FLASH_ATTENTION=$DEFAULT_USE_FLASH_ATTENTION

RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
    ca-certificates \
    libssl-dev \
    curl \
    && rm -rf /var/lib/apt/lists/*

FROM base AS grpc

COPY --from=grpc-builder /usr/src/target/release/text-embeddings-router /usr/local/bin/text-embeddings-router

ENTRYPOINT ["text-embeddings-router"]
CMD ["--json-output"]

FROM base

COPY --from=http-builder /usr/src/target/release/text-embeddings-router /usr/local/bin/text-embeddings-router

ENTRYPOINT ["text-embeddings-router"]
CMD ["--json-output"]

[amin3141]

+1 for this

[trillionmonster]

candle not support 120

[toppurls-png]

+1 for this. On RTX 5060 Ti.

[danielealbano]

Check out #735, I should have fixed the support