minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[Deps] Retire the CUDA 12 lane (#38404)

Browse Source 
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
This commit is contained in:
Mohammad Miadh Angkad
2026-09-10 16:58:09 -07:00
committed by GitHub
co-authored by Mohammad Angkad
parent fae8cd84cb
commit 52c191da52
38 changed files with 170 additions and 737 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/ci/cuda/ci_install_dependency.sh
+32 -78
View File
@@ -2,7 +2,7 @@
# Install dependencies for CUDA CI jobs.
#
# CU_VERSION (default: cu130) controls PyTorch index URL, FlashInfer JIT cache
# index, and nvrtc variant selection.
# index, and the sglang wheel index. CUDA 13 only.
set -euxo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
@@ -32,10 +32,14 @@ mark_step_done() {
configure_environment() {
# CU_VERSION controls PyTorch index URL, FlashInfer JIT cache index, and
# nvrtc variant selection (cu12 vs cu13).
# the sglang wheel index. Only CUDA 13 lanes exist: PyTorch 2.14 publishes
# no CUDA 12 wheels for the cu129 index the retired cu12 lane used.
CU_VERSION="${CU_VERSION:-cu130}"
CU_STRIP="${CU_VERSION#cu}"
CU_MAJOR="${CU_STRIP:0:2}"
case "${CU_STRIP}" in
13*) ;;
*) echo "FATAL: unsupported CU_VERSION=${CU_VERSION}; only CUDA 13 is supported"; exit 1 ;;
esac
OPTIONAL_DEPS="${1:-}"
@@ -366,11 +370,6 @@ remove_stale_cuda12_nvidia_wheels() {
local -a NVIDIA_WHEELS_TO_RESTORE=()
local -a STALE_CUDA12_NVIDIA_WHEELS=()
if [ "$CU_MAJOR" != "13" ]; then
mark_step_done "${FUNCNAME[0]}"
return
fi
mapfile -t INSTALLED_NVIDIA_WHEELS < <(
python3 -m pip list --format=freeze | sed -n '/^nvidia-.*==/p'
)
@@ -462,30 +461,6 @@ install_pytorch_stack() {
mark_step_done "${FUNCNAME[0]}"
}
install_cuda12_deepep_wheel() {
if [ "$CU_MAJOR" = "13" ]; then
echo "CUDA 13 uses the public sgl-deep-ep wheel declared in python/pyproject.toml"
mark_step_done "${FUNCNAME[0]}"
return
fi
local version
version=$(grep -Po -m1 '"sgl-deep-ep==\K[^"]+' python/pyproject.toml || true)
if [ -z "$version" ]; then
echo "ERROR: python/pyproject.toml must pin sgl-deep-ep"
exit 1
fi
# CUDA 12 wheels intentionally live only on the SGLang wheel index. Their
# local version satisfies the public-version pyproject pin, so the later
# editable SGLang install keeps this CUDA-matched wheel.
$PIP_CMD install "sgl-deep-ep==${version}+${CU_VERSION}" \
--index-url "https://docs.sglang.ai/whl/${CU_VERSION}/" \
--force-reinstall --no-deps $PIP_INSTALL_SUFFIX
mark_step_done "${FUNCNAME[0]}"
}
require_prebuilt_rust_exts() {
# Stages whose download succeeded set this to none. Runs before
# setup_pip_toolchain uninstalls sglang, so clearing it here still reaches
@@ -551,14 +526,10 @@ install_sglang() {
}
install_nccl() {
if [ "$CU_MAJOR" = "13" ]; then
# PyTorch pins 2.29.7, so this override must run after every command
# that resolves Python dependencies (including lmms-eval).
$PIP_CMD install "nvidia-nccl-cu13==2.30.7" \
--force-reinstall --no-deps $PIP_INSTALL_SUFFIX
else
echo "CUDA ${CU_MAJOR} does not require the NCCL Gin wheel"
fi
# PyTorch pins 2.29.7, so this override must run after every command
# that resolves Python dependencies (including lmms-eval).
$PIP_CMD install "nvidia-nccl-cu13==2.30.7" \
--force-reinstall --no-deps $PIP_INSTALL_SUFFIX
mark_step_done "${FUNCNAME[0]}"
}
@@ -628,8 +599,8 @@ install_sglang_kernel() {
if [ "${CUSTOM_BUILD_SGL_KERNEL:-}" != "true" ]; then
# The PyPI default wheel tracks one CUDA version (currently cu130); other
# runners (e.g. h20 / cu129) need the +${CU_VERSION}-tagged wheel from the
# sglang index, linked against the right libnvrtc.
# runners need the +${CU_VERSION}-tagged wheel from the sglang index,
# linked against the right libnvrtc.
SGL_KERNEL_WANTED="${SGL_KERNEL_VERSION_FROM_SRT}+${CU_VERSION}"
if installed_wheel_ok sglang-kernel "${SGL_KERNEL_WANTED}" reject-local; then
echo "sglang-kernel==${SGL_KERNEL_WANTED} already installed, keeping it"
@@ -640,18 +611,11 @@ install_sglang_kernel() {
echo "CUSTOM_BUILD_SGL_KERNEL=true: keeping freshly built sgl-kernel wheel."
fi
SGL_DEEP_GEMM_VERSION=$(grep -Po -m1 '(?<=sgl-deep-gemm==)[0-9A-Za-z\.\-]+' python/pyproject.toml)
if [ "$CU_MAJOR" = "13" ]; then
SGL_DEEP_GEMM_WANTED="${SGL_DEEP_GEMM_VERSION}"
else
SGL_DEEP_GEMM_WANTED="${SGL_DEEP_GEMM_VERSION}+cu129"
fi
# No reject-local: nothing builds sgl-deep-gemm locally.
if installed_wheel_ok sgl-deep-gemm "${SGL_DEEP_GEMM_WANTED}"; then
echo "sgl-deep-gemm==${SGL_DEEP_GEMM_WANTED} already installed, keeping it"
elif [ "$CU_MAJOR" = "13" ]; then
$PIP_CMD install "sgl-deep-gemm==${SGL_DEEP_GEMM_VERSION}" --force-reinstall $PIP_INSTALL_SUFFIX
if installed_wheel_ok sgl-deep-gemm "${SGL_DEEP_GEMM_VERSION}"; then
echo "sgl-deep-gemm==${SGL_DEEP_GEMM_VERSION} already installed, keeping it"
else
$PIP_CMD install "https://github.com/sgl-project/whl/releases/download/v${SGL_DEEP_GEMM_VERSION}/sgl_deep_gemm-${SGL_DEEP_GEMM_VERSION}+cu129-py3-none-manylinux2014_$(uname -m).whl" --force-reinstall $PIP_INSTALL_SUFFIX
$PIP_CMD install "sgl-deep-gemm==${SGL_DEEP_GEMM_VERSION}" --force-reinstall $PIP_INSTALL_SUFFIX
fi
mark_step_done "${FUNCNAME[0]}"
@@ -739,17 +703,10 @@ stabilize_flashinfer_jit_paths() {
install_extra_deps() {
MOONCAKE_VERSION="0.3.13"
NIXL_VERSION="1.3.0"
if [ "$CU_MAJOR" = "13" ]; then
MOONCAKE_PKG="mooncake-transfer-engine-cuda13==${MOONCAKE_VERSION}"
MOONCAKE_STALE_PKG="mooncake-transfer-engine"
NIXL_BIN_NAME="nixl-cu13"
EXTRA_NVIDIA_SPECS="nvidia-cuda-nvrtc"
else
MOONCAKE_PKG="mooncake-transfer-engine==${MOONCAKE_VERSION}"
MOONCAKE_STALE_PKG="mooncake-transfer-engine-cuda13"
NIXL_BIN_NAME="nixl-cu12"
EXTRA_NVIDIA_SPECS="nvidia-cuda-nvrtc-cu12"
fi
MOONCAKE_PKG="mooncake-transfer-engine-cuda13==${MOONCAKE_VERSION}"
MOONCAKE_STALE_PKG="mooncake-transfer-engine"
NIXL_BIN_NAME="nixl-cu13"
EXTRA_NVIDIA_SPECS="nvidia-cuda-nvrtc"
# Both variants own the same mooncake/ package files and bin/ scripts
# (mooncake_master, etc.). Uninstalling the stale variant deletes shared
# files that the live variant's RECORD still references, so we force a
@@ -834,20 +791,18 @@ verify_imports() {
import ctypes
import importlib.metadata
import os
import sys
if sys.argv[1] == "13":
if importlib.metadata.version("nvidia-nccl-cu13") != "2.30.7":
raise SystemExit("nvidia-nccl-cu13 was changed after the final CI override")
nccl = ctypes.CDLL("libnccl.so.2")
nccl_version = ctypes.c_int()
status = nccl.ncclGetVersion(ctypes.byref(nccl_version))
if status != 0 or nccl_version.value != 23007:
raise SystemExit(
f"expected NCCL runtime 2.30.7, got status={status}, "
f"raw_version={nccl_version.value}"
)
print("NCCL package and runtime versions are 2.30.7")
if importlib.metadata.version("nvidia-nccl-cu13") != "2.30.7":
raise SystemExit("nvidia-nccl-cu13 was changed after the final CI override")
nccl = ctypes.CDLL("libnccl.so.2")
nccl_version = ctypes.c_int()
status = nccl.ncclGetVersion(ctypes.byref(nccl_version))
if status != 0 or nccl_version.value != 23007:
raise SystemExit(
f"expected NCCL runtime 2.30.7, got status={status}, "
f"raw_version={nccl_version.value}"
)
print("NCCL package and runtime versions are 2.30.7")
import torch
print(torch.version.cuda)
@@ -882,7 +837,7 @@ for mod in ("server", "grpc", "multimodal"):
except Exception as exc:
raise SystemExit(f"{name} is present but does not load: {exc!r}")
print(f"{name} loads")
' "$CU_MAJOR"
'
mark_step_done "${FUNCNAME[0]}"
}
@@ -904,7 +859,6 @@ main() {
remove_stale_cuda12_nvidia_wheels
uninstall_stale_flashinfer
install_pytorch_stack
install_cuda12_deepep_wheel
setup_cargo_cache
install_sglang
release_cargo_cache_lock
scripts/ci/utils/docker_build_metadata_args.py
+1 -1
View File
@@ -3,7 +3,7 @@ import datetime
import json
import sys
MOVING_TAGS = {"dev", "dev-cu12", "dev-cu13", "latest"}
MOVING_TAGS = {"dev", "dev-cu13", "latest"}
def render_tag_template(tag: str, version: str, date: str, short_sha: str) -> str: