mirror of
https://github.com/mudler/LocalAI.git
synced 2026-09-12 22:33:54 -04:00
Claudio Maradonna
9e831d7709
* fix(ds4): build CUDA kernels for the target architecture The ds4 backend compiled its CUDA objects with no -arch. Upstream's Makefile leaves CUDA_ARCH empty and its `cuda` target refuses to build without one, offering `cuda-spark` (sm_121) and `cuda-generic` (native) instead. We invoke its object targets directly, which bypasses that guard, so nvcc fell back to its default architecture and the kernels ran as JIT'd PTX on the real GPU. On GB10 (sm_121) that silently corrupted inference: any prompt over roughly 128 tokens produced text unrelated to the input and never closed its thinking block, so content came back empty and the chat showed only reasoning; longer prompts failed with "cuda decode failed". It also cost close to two orders of magnitude of prefill throughput. Measured on one box, same model, same prompt, same GPU, upstream ds4 at the pinned commit, differing only in the nvcc flags: make -B ds4 (archless, as we build it) garbage output 4.21 t/s make cuda-spark (compute_121a/sm_121a) correct output 325.70 t/s Select an architecture list from CUDA_MAJOR_VERSION, which the backend matrix already declares for both ds4 cublas entries but Dockerfile.ds4 never forwarded. Upstream's CUDA_ARCH takes a single value, so it cannot express the fat binary these images need; NVCC_ARCH_FLAGS is overridden instead, since a command-line assignment wins over its `:=`. The lists are copied from vllm-cpp rather than invented so the two CUDA images cover the same GPUs, with l4t/arm64 covering Orin, Thor and GB10. An empty CUDA_MAJOR_VERSION keeps upstream's `native` behaviour for local developer builds, and no CI runner has a GPU to enumerate. DS4_CUDA_HAVE_MXF4 is deliberately left unset: upstream defines it only for single-arch sm_120/sm_121 builds and guards it with a plain #ifdef rather than __CUDA_ARCH__, so it cannot be combined with older archs. It gates an optional MXFP4 indexer fast path whose #ifndef branch returns 0 and falls back cleanly, so omitting it costs speed on GB10, not correctness. Assisted-by: Claude Code:claude-opus-5 Signed-off-by: Claudio Maradonna <git@codeshifter.xyz> * test(ds4): cover the multi-batch prefill regression The architecture fix has no automated guard: every existing e2e spec uses a short prompt, and the miscompiled backend answered short prompts correctly. The corruption only appears once a prompt spans more than one prefill batch, so the whole suite passed against a backend that produced garbage in normal use. Add an opt-in "long_prefill" capability to the backend e2e suite that sends a prompt well past one batch with a known needle and asserts the answer still reflects it, and document in the ds4 guide why the build must never omit an nvcc architecture, how to check which flags a configuration resolves to without compiling, and how to run the new spec. Assisted-by: Claude Code:claude-opus-5 Signed-off-by: Claudio Maradonna <git@codeshifter.xyz> --------- Signed-off-by: Claudio Maradonna <git@codeshifter.xyz>
44 lines
1.7 KiB
Docker
44 lines
1.7 KiB
Docker
ARG BASE_IMAGE=ubuntu:24.04
|
|
ARG APT_MIRROR=""
|
|
ARG APT_PORTS_MIRROR=""
|
|
|
|
# BASE_IMAGE is either ubuntu:24.04 (for cpu builds) or nvidia/cuda:13.0.0-devel-ubuntu24.04
|
|
# (for cublas builds). Both ship apt + Ubuntu Noble packages; the nvidia/cuda base
|
|
# additionally provides /usr/local/cuda. Darwin (Metal) builds bypass this Dockerfile
|
|
# entirely via scripts/build/ds4-darwin.sh.
|
|
FROM ${BASE_IMAGE} AS builder
|
|
ARG BUILD_TYPE
|
|
ARG TARGETARCH
|
|
ARG TARGETVARIANT
|
|
ARG CUDA_MAJOR_VERSION
|
|
|
|
ENV BUILD_TYPE=${BUILD_TYPE} \
|
|
DEBIAN_FRONTEND=noninteractive \
|
|
PATH=/usr/local/cuda/bin:${PATH}
|
|
|
|
WORKDIR /build
|
|
|
|
# Install build-time deps via plain apt - install-base-deps.sh's full pipeline
|
|
# (CUDA keyring + from-source gRPC) is unnecessary here:
|
|
# - CUDA: when BASE_IMAGE=nvidia/cuda:*, /usr/local/cuda is already populated;
|
|
# for the cpu build we don't need CUDA at all.
|
|
# - gRPC/Protobuf: system apt packages are sufficient; ds4's wrapper only links
|
|
# against them, it doesn't ship the gRPC source tree.
|
|
# - nlohmann-json: dsml_renderer's only third-party dep.
|
|
RUN apt-get update && \
|
|
apt-get install -y --no-install-recommends \
|
|
git cmake build-essential pkg-config ca-certificates \
|
|
libgrpc++-dev libprotobuf-dev protobuf-compiler protobuf-compiler-grpc \
|
|
nlohmann-json3-dev && \
|
|
apt-get clean && \
|
|
rm -rf /var/lib/apt/lists/*
|
|
|
|
COPY . /LocalAI
|
|
|
|
RUN --mount=type=cache,target=/root/.ccache,id=ds4-ccache-${TARGETARCH}-${BUILD_TYPE},sharing=locked \
|
|
make -C /LocalAI/backend/cpp/ds4 BUILD_TYPE=${BUILD_TYPE} \
|
|
CUDA_MAJOR_VERSION=${CUDA_MAJOR_VERSION} NATIVE=false grpc-server package
|
|
|
|
FROM scratch
|
|
COPY --from=builder /LocalAI/backend/cpp/ds4/package/. ./
|