pytorch/.github/scripts/generate_binary_build_matrix.py

#!/usr/bin/env python3

"""Generates a matrix to be utilized through github actions

Will output a condensed version of the matrix if on a pull request that only
includes the latest version of python we support built on three different
architectures:
    * CPU
    * Latest CUDA
    * Latest ROCM
"""

from typing import Dict, List, Optional, Tuple


CUDA_ARCHES = ["11.8", "12.1"]


ROCM_ARCHES = ["5.6", "5.7"]


CPU_CXX11_ABI_ARCH = ["cpu-cxx11-abi"]


CPU_AARCH64_ARCH = ["cpu-aarch64"]

PYTORCH_EXTRA_INSTALL_REQUIREMENTS = (
    "nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "  # noqa: B950
    "nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-nccl-cu12==2.19.3; platform_system == 'Linux' and platform_machine == 'x86_64' | "
    "nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'"
)


def get_nccl_submodule_version() -> str:
    from pathlib import Path

    nccl_version_mk = (
        Path(__file__).absolute().parent.parent.parent
        / "third_party"
        / "nccl"
        / "nccl"
        / "makefiles"
        / "version.mk"
    )
    if not nccl_version_mk.exists():
        raise RuntimeError(
            "Please make sure that nccl submodule is checked out when importing this script"
        )
    with nccl_version_mk.open("r") as f:
        content = f.read()
    d = {}
    for l in content.split("\n"):
        if not l.startswith("NCCL_"):
            continue
        (k, v) = l.split(":=")
        d[k.strip()] = v.strip()
    return f"{d['NCCL_MAJOR']}.{d['NCCL_MINOR']}.{d['NCCL_PATCH']}"


def get_nccl_wheel_version() -> str:
    import re

    requrements = map(str.strip, re.split("[;|]", PYTORCH_EXTRA_INSTALL_REQUIREMENTS))
    return [x for x in requrements if x.startswith("nvidia-nccl-cu")][0].split("==")[1]


def validate_nccl_dep_consistency() -> None:
    wheel_ver = get_nccl_wheel_version()
    submodule_ver = get_nccl_submodule_version()
    if wheel_ver != submodule_ver:
        raise RuntimeError(
            f"NCCL submodule version {submodule_ver} differs from wheel version {wheel_ver}"
        )


def arch_type(arch_version: str) -> str:
    if arch_version in CUDA_ARCHES:
        return "cuda"
    elif arch_version in ROCM_ARCHES:
        return "rocm"
    elif arch_version in CPU_CXX11_ABI_ARCH:
        return "cpu-cxx11-abi"
    elif arch_version in CPU_AARCH64_ARCH:
        return "cpu-aarch64"
    else:  # arch_version should always be "cpu" in this case
        return "cpu"


# This can be updated to the release version when cutting release branch, i.e. 2.1
DEFAULT_TAG = "main"

WHEEL_CONTAINER_IMAGES = {
    **{
        gpu_arch: f"pytorch/manylinux-builder:cuda{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in CUDA_ARCHES
    },
    **{
        gpu_arch: f"pytorch/manylinux-builder:rocm{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in ROCM_ARCHES
    },
    "cpu": f"pytorch/manylinux-builder:cpu-{DEFAULT_TAG}",
    "cpu-cxx11-abi": f"pytorch/manylinuxcxx11-abi-builder:cpu-cxx11-abi-{DEFAULT_TAG}",
    "cpu-aarch64": f"pytorch/manylinuxaarch64-builder:cpu-aarch64-{DEFAULT_TAG}",
}

CONDA_CONTAINER_IMAGES = {
    **{
        gpu_arch: f"pytorch/conda-builder:cuda{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in CUDA_ARCHES
    },
    "cpu": f"pytorch/conda-builder:cpu-{DEFAULT_TAG}",
}

PRE_CXX11_ABI = "pre-cxx11"
CXX11_ABI = "cxx11-abi"
RELEASE = "release"
DEBUG = "debug"

LIBTORCH_CONTAINER_IMAGES: Dict[Tuple[str, str], str] = {
    **{
        (
            gpu_arch,
            PRE_CXX11_ABI,
        ): f"pytorch/manylinux-builder:cuda{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in CUDA_ARCHES
    },
    **{
        (
            gpu_arch,
            CXX11_ABI,
        ): f"pytorch/libtorch-cxx11-builder:cuda{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in CUDA_ARCHES
    },
    **{
        (
            gpu_arch,
            PRE_CXX11_ABI,
        ): f"pytorch/manylinux-builder:rocm{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in ROCM_ARCHES
    },
    **{
        (
            gpu_arch,
            CXX11_ABI,
        ): f"pytorch/libtorch-cxx11-builder:rocm{gpu_arch}-{DEFAULT_TAG}"
        for gpu_arch in ROCM_ARCHES
    },
    ("cpu", PRE_CXX11_ABI): f"pytorch/manylinux-builder:cpu-{DEFAULT_TAG}",
    ("cpu", CXX11_ABI): f"pytorch/libtorch-cxx11-builder:cpu-{DEFAULT_TAG}",
}

FULL_PYTHON_VERSIONS = ["3.8", "3.9", "3.10", "3.11"]


def translate_desired_cuda(gpu_arch_type: str, gpu_arch_version: str) -> str:
    return {
        "cpu": "cpu",
        "cpu-aarch64": "cpu",
        "cpu-cxx11-abi": "cpu-cxx11-abi",
        "cuda": f"cu{gpu_arch_version.replace('.', '')}",
        "rocm": f"rocm{gpu_arch_version}",
    }.get(gpu_arch_type, gpu_arch_version)


def list_without(in_list: List[str], without: List[str]) -> List[str]:
    return [item for item in in_list if item not in without]


def generate_conda_matrix(os: str) -> List[Dict[str, str]]:
    ret: List[Dict[str, str]] = []
    arches = ["cpu"]
    python_versions = FULL_PYTHON_VERSIONS
    if os == "linux" or os == "windows":
        arches += CUDA_ARCHES
    for python_version in python_versions:
        # We don't currently build conda packages for rocm
        for arch_version in arches:
            gpu_arch_type = arch_type(arch_version)
            gpu_arch_version = "" if arch_version == "cpu" else arch_version
            ret.append(
                {
                    "python_version": python_version,
                    "gpu_arch_type": gpu_arch_type,
                    "gpu_arch_version": gpu_arch_version,
                    "desired_cuda": translate_desired_cuda(
                        gpu_arch_type, gpu_arch_version
                    ),
                    "container_image": CONDA_CONTAINER_IMAGES[arch_version],
                    "package_type": "conda",
                    "build_name": f"conda-py{python_version}-{gpu_arch_type}{gpu_arch_version}".replace(
                        ".", "_"
                    ),
                }
            )
    return ret


def generate_libtorch_matrix(
    os: str,
    abi_version: str,
    arches: Optional[List[str]] = None,
    libtorch_variants: Optional[List[str]] = None,
) -> List[Dict[str, str]]:
    if arches is None:
        arches = ["cpu"]
        if os == "linux":
            arches += CUDA_ARCHES
            arches += ROCM_ARCHES
        elif os == "windows":
            arches += CUDA_ARCHES

    if libtorch_variants is None:
        libtorch_variants = [
            "shared-with-deps",
            "shared-without-deps",
            "static-with-deps",
            "static-without-deps",
        ]

    ret: List[Dict[str, str]] = []
    for arch_version in arches:
        for libtorch_variant in libtorch_variants:
            # one of the values in the following list must be exactly
            # CXX11_ABI, but the precise value of the other one doesn't
            # matter
            gpu_arch_type = arch_type(arch_version)
            gpu_arch_version = "" if arch_version == "cpu" else arch_version
            # ROCm builds without-deps failed even in ROCm runners; skip for now
            if gpu_arch_type == "rocm" and "without-deps" in libtorch_variant:
                continue
            ret.append(
                {
                    "gpu_arch_type": gpu_arch_type,
                    "gpu_arch_version": gpu_arch_version,
                    "desired_cuda": translate_desired_cuda(
                        gpu_arch_type, gpu_arch_version
                    ),
                    "libtorch_variant": libtorch_variant,
                    "libtorch_config": abi_version if os == "windows" else "",
                    "devtoolset": abi_version if os != "windows" else "",
                    "container_image": LIBTORCH_CONTAINER_IMAGES[
                        (arch_version, abi_version)
                    ]
                    if os != "windows"
                    else "",
                    "package_type": "libtorch",
                    "build_name": f"libtorch-{gpu_arch_type}{gpu_arch_version}-{libtorch_variant}-{abi_version}".replace(
                        ".", "_"
                    ),
                }
            )
    return ret


def generate_wheels_matrix(
    os: str,
    arches: Optional[List[str]] = None,
    python_versions: Optional[List[str]] = None,
) -> List[Dict[str, str]]:
    package_type = "wheel"
    if os == "linux" or os == "linux-aarch64":
        # NOTE: We only build manywheel packages for x86_64 and aarch64 linux
        package_type = "manywheel"

    if python_versions is None:
        python_versions = FULL_PYTHON_VERSIONS + ["3.12"]

    if arches is None:
        # Define default compute archivectures
        arches = ["cpu"]
        if os == "linux":
            arches += CPU_CXX11_ABI_ARCH + CUDA_ARCHES + ROCM_ARCHES
        elif os == "windows":
            arches += CUDA_ARCHES
        elif os == "linux-aarch64":
            # Only want the one arch as the CPU type is different and
            # uses different build/test scripts
            arches = ["cpu-aarch64"]

    ret: List[Dict[str, str]] = []
    for python_version in python_versions:
        for arch_version in arches:
            gpu_arch_type = arch_type(arch_version)
            gpu_arch_version = (
                ""
                if arch_version == "cpu"
                or arch_version == "cpu-cxx11-abi"
                or arch_version == "cpu-aarch64"
                else arch_version
            )

            # 12.1 linux wheels require PYTORCH_EXTRA_INSTALL_REQUIREMENTS to install
            if arch_version == "12.1" and os == "linux":
                ret.append(
                    {
                        "python_version": python_version,
                        "gpu_arch_type": gpu_arch_type,
                        "gpu_arch_version": gpu_arch_version,
                        "desired_cuda": translate_desired_cuda(
                            gpu_arch_type, gpu_arch_version
                        ),
                        "devtoolset": "",
                        "container_image": WHEEL_CONTAINER_IMAGES[arch_version],
                        "package_type": package_type,
                        "pytorch_extra_install_requirements": PYTORCH_EXTRA_INSTALL_REQUIREMENTS,
                        "build_name": f"{package_type}-py{python_version}-{gpu_arch_type}{gpu_arch_version}".replace(  # noqa: B950
                            ".", "_"
                        ),
                    }
                )
            else:
                ret.append(
                    {
                        "python_version": python_version,
                        "gpu_arch_type": gpu_arch_type,
                        "gpu_arch_version": gpu_arch_version,
                        "desired_cuda": translate_desired_cuda(
                            gpu_arch_type, gpu_arch_version
                        ),
                        "devtoolset": "cxx11-abi"
                        if arch_version == "cpu-cxx11-abi"
                        else "",
                        "container_image": WHEEL_CONTAINER_IMAGES[arch_version],
                        "package_type": package_type,
                        "build_name": f"{package_type}-py{python_version}-{gpu_arch_type}{gpu_arch_version}".replace(
                            ".", "_"
                        ),
                        "pytorch_extra_install_requirements": PYTORCH_EXTRA_INSTALL_REQUIREMENTS
                        if os != "linux"
                        else "",
                    }
                )
    return ret


validate_nccl_dep_consistency()