mirror of
https://github.com/tensorflow/tensorflow.git
synced 2024-11-21 21:05:19 +00:00
6411af23a9
This updates CUDA for both TF and XLA. It also enables the CUDA driver forward compatibility mode for XLA since XLA's CUDA graph integration needs a newer driver version. PiperOrigin-RevId: 673974335
921 lines
51 KiB
Plaintext
921 lines
51 KiB
Plaintext
# TensorFlow Bazel configuration file.
|
|
# This file tries to group and simplify build options for TensorFlow
|
|
#
|
|
# ----CONFIG OPTIONS----
|
|
# Android options:
|
|
# android:
|
|
# android_arm:
|
|
# android_arm64:
|
|
# android_x86:
|
|
# android_x86_64:
|
|
#
|
|
# iOS options:
|
|
# ios:
|
|
# ios_armv7:
|
|
# ios_arm64:
|
|
# ios_x86_64:
|
|
# ios_fat:
|
|
#
|
|
# Macosx options
|
|
# darwin_arm64:
|
|
#
|
|
# Compiler options:
|
|
# cuda_clang: Use Clang when building CUDA code.
|
|
# avx_linux: Build with avx instruction set on linux.
|
|
# avx_win: Build with avx instruction set on windows
|
|
#
|
|
# Other build options:
|
|
# short_logs: Only log errors during build, skip warnings.
|
|
# verbose_logs: Show all compiler warnings during build.
|
|
# monolithic: Build all TF C++ code into a single shared object.
|
|
# dynamic_kernels: Try to link all kernels dynamically (experimental).
|
|
# dbg: Build with debug info
|
|
#
|
|
# TF version options;
|
|
# v2: Build TF v2
|
|
#
|
|
# Feature and Third party library support options:
|
|
# xla: Build TF with XLA
|
|
# tpu: Build TF with TPU support
|
|
# cuda: Build with CUDA support.
|
|
# cuda_clang Build with CUDA Clang support.
|
|
# rocm: Build with AMD GPU support (rocm)
|
|
# mkl: Enable full mkl support.
|
|
# nogcp: Disable GCS support.
|
|
# nonccl: Disable nccl support.
|
|
#
|
|
#
|
|
# Remote build execution options (only configured to work with TF team projects for now.)
|
|
# rbe_base: General RBE options shared by all flavors.
|
|
# rbe_linux: General RBE options used on all linux builds.
|
|
# rbe_win_base: General RBE options used on all Windows builds. Not to be used standalone.
|
|
# rbe_win_clang: Options specific to compiling using Clang.
|
|
#
|
|
# rbe_linux_cpu: RBE options to build with only CPU support.
|
|
# rbe_linux_cuda: RBE options to build with GPU support using clang.
|
|
# rbe_linux_cuda_nvcc: RBE options to build with GPU support using nvcc.
|
|
#
|
|
# Embedded Linux options (experimental and only tested with TFLite build yet)
|
|
# elinux: General Embedded Linux options shared by all flavors.
|
|
# elinux_aarch64: Embedded Linux options for aarch64 (ARM64) CPU support.
|
|
# elinux_armhf: Embedded Linux options for armhf (ARMv7) CPU support.
|
|
#
|
|
# Release build options (for all operating systems)
|
|
# release_base: Common options for all builds on all operating systems.
|
|
# release_cpu_linux: Toolchain and CUDA options for Linux CPU builds.
|
|
# release_gpu_linux: Toolchain and CUDA options for Linux GPU builds.
|
|
# release_cpu_macos: Toolchain and CUDA options for MacOS CPU builds.
|
|
# release_cpu_windows: Toolchain and CUDA options for Windows CPU builds.
|
|
|
|
# Default build options. These are applied first and unconditionally.
|
|
|
|
# For projects which use TensorFlow as part of a Bazel build process, putting
|
|
# nothing in a bazelrc will default to a monolithic build. The following line
|
|
# opts in to modular op registration support by default.
|
|
build --define framework_shared_object=true
|
|
build --define tsl_protobuf_header_only=true
|
|
|
|
build --define=use_fast_cpp_protos=true
|
|
build --define=allow_oversize_protos=true
|
|
|
|
build --spawn_strategy=standalone
|
|
build -c opt
|
|
|
|
# Make Bazel print out all options from rc files.
|
|
build --announce_rc
|
|
|
|
# TODO(mihaimaruseac): Document this option or remove if no longer needed
|
|
build --define=grpc_no_ares=true
|
|
|
|
# See https://github.com/bazelbuild/bazel/issues/7362 for information on what
|
|
# --incompatible_remove_legacy_whole_archive flag does.
|
|
# This flag is set to true in Bazel 1.0 and newer versions. We tried to migrate
|
|
# Tensorflow to the default, however test coverage wasn't enough to catch the
|
|
# errors.
|
|
# There is ongoing work on Bazel team's side to provide support for transitive
|
|
# shared libraries. As part of migrating to transitive shared libraries, we
|
|
# hope to provide a better mechanism for control over symbol exporting, and
|
|
# then tackle this issue again.
|
|
#
|
|
# TODO: Remove the following two lines once TF doesn't depend on Bazel wrapping
|
|
# all library archives in -whole_archive -no_whole_archive.
|
|
build --noincompatible_remove_legacy_whole_archive
|
|
build --features=-force_no_whole_archive
|
|
|
|
# TODO(mihaimaruseac): Document this option or remove if no longer needed
|
|
build --enable_platform_specific_config
|
|
|
|
# Enable XLA support by default.
|
|
build --define=with_xla_support=true
|
|
|
|
# TODO(mihaimaruseac): Document this option or remove if no longer needed
|
|
build --config=short_logs
|
|
|
|
# TODO(mihaimaruseac): Document this option or remove if no longer needed
|
|
build --config=v2
|
|
|
|
# TF now has `cc_shared_library` targets, so it needs the experimental flag
|
|
# TODO(rostam): Remove when `cc_shared_library` is enabled by default
|
|
build --experimental_cc_shared_library
|
|
|
|
# cc_shared_library ensures no library is linked statically more than once.
|
|
build --experimental_link_static_libraries_once=false
|
|
|
|
# Prevent regressions on those two incompatible changes
|
|
# TODO: remove those flags when they are flipped in the default Bazel version TF uses.
|
|
build --incompatible_enforce_config_setting_visibility
|
|
# TODO: also enable this flag after fixing the visibility violations
|
|
# build --incompatible_config_setting_private_default_visibility
|
|
|
|
# Default options should come above this line.
|
|
|
|
# Android configs. Bazel needs to have --cpu and --fat_apk_cpu both set to the
|
|
# target CPU to build transient dependencies correctly. See
|
|
# https://docs.bazel.build/versions/master/user-manual.html#flag--fat_apk_cpu
|
|
build:android --crosstool_top=//external:android/crosstool
|
|
build:android --host_crosstool_top=@bazel_tools//tools/cpp:toolchain
|
|
build:android_arm --config=android
|
|
build:android_arm --cpu=armeabi-v7a
|
|
build:android_arm --fat_apk_cpu=armeabi-v7a
|
|
build:android_arm64 --config=android
|
|
build:android_arm64 --cpu=arm64-v8a
|
|
build:android_arm64 --fat_apk_cpu=arm64-v8a
|
|
build:android_x86 --config=android
|
|
build:android_x86 --cpu=x86
|
|
build:android_x86 --fat_apk_cpu=x86
|
|
build:android_x86_64 --config=android
|
|
build:android_x86_64 --cpu=x86_64
|
|
build:android_x86_64 --fat_apk_cpu=x86_64
|
|
|
|
# Build everything statically for Android since all static libs are later
|
|
# bundled together into a single .so for deployment.
|
|
build:android --dynamic_mode=off
|
|
|
|
# Sets the default Apple platform to macOS.
|
|
build:macos --apple_platform_type=macos
|
|
|
|
# gRPC on MacOS requires this #define
|
|
build:macos --copt=-DGRPC_BAZEL_BUILD
|
|
|
|
# Avoid hitting command line argument limit
|
|
build:macos --features=archive_param_file
|
|
|
|
# Settings for MacOS on ARM CPUs.
|
|
build:macos_arm64 --cpu=darwin_arm64
|
|
build:macos_arm64 --macos_minimum_os=11.0
|
|
|
|
# iOS configs for each architecture and the fat binary builds.
|
|
build:ios --apple_platform_type=ios
|
|
build:ios --apple_bitcode=embedded --copt=-fembed-bitcode
|
|
build:ios --copt=-Wno-c++11-narrowing
|
|
build:ios_armv7 --config=ios
|
|
build:ios_armv7 --cpu=ios_armv7
|
|
build:ios_arm64 --config=ios
|
|
build:ios_arm64 --cpu=ios_arm64
|
|
build:ios_arm64e --config=ios
|
|
build:ios_arm64e --cpu=ios_arm64e
|
|
build:ios_sim_arm64 --config=ios
|
|
build:ios_sim_arm64 --cpu=ios_sim_arm64
|
|
build:ios_x86_64 --config=ios
|
|
build:ios_x86_64 --cpu=ios_x86_64
|
|
build:ios_fat --config=ios
|
|
build:ios_fat --ios_multi_cpus=armv7,arm64,i386,x86_64
|
|
|
|
# Config to use a mostly-static build and disable modular op registration
|
|
# support (this will revert to loading TensorFlow with RTLD_GLOBAL in Python).
|
|
# By default, TensorFlow will build with a dependence on
|
|
# //tensorflow:libtensorflow_framework.so.
|
|
build:monolithic --define framework_shared_object=false
|
|
build:monolithic --define tsl_protobuf_header_only=false
|
|
build:monolithic --experimental_link_static_libraries_once=false # b/229868128
|
|
|
|
# Please note that MKL on MacOS is still not supported.
|
|
# If you would like to use a local MKL instead of downloading, please set the
|
|
# environment variable "TF_MKL_ROOT" every time before build.
|
|
build:mkl --define=build_with_mkl=true --define=enable_mkl=true
|
|
build:mkl --define=tensorflow_mkldnn_contraction_kernel=0
|
|
build:mkl --define=build_with_openmp=true
|
|
build:mkl -c opt
|
|
|
|
# config to build OneDNN backend with a user specified threadpool.
|
|
build:mkl_threadpool --define=build_with_mkl=true --define=enable_mkl=true
|
|
build:mkl_threadpool --define=tensorflow_mkldnn_contraction_kernel=0
|
|
build:mkl_threadpool --define=build_with_mkl_opensource=true
|
|
build:mkl_threadpool -c opt
|
|
|
|
# Config setting to build oneDNN with Compute Library for the Arm Architecture (ACL).
|
|
build:mkl_aarch64 --define=build_with_mkl_aarch64=true
|
|
build:mkl_aarch64 --define=build_with_openmp=true
|
|
build:mkl_aarch64 --define=build_with_acl=true
|
|
build:mkl_aarch64 -c opt
|
|
|
|
# Config setting to build oneDNN with Compute Library for the Arm Architecture (ACL).
|
|
# with Eigen threadpool support
|
|
build:mkl_aarch64_threadpool --define=build_with_mkl_aarch64=true
|
|
build:mkl_aarch64_threadpool -c opt
|
|
|
|
# CUDA: This config refers to building CUDA op kernels with nvcc.
|
|
build:cuda --repo_env TF_NEED_CUDA=1
|
|
build:cuda --crosstool_top=@local_config_cuda//crosstool:toolchain
|
|
build:cuda --@local_config_cuda//:enable_cuda
|
|
# Default CUDA and CUDNN versions.
|
|
build:cuda --repo_env=HERMETIC_CUDA_VERSION="12.5.1"
|
|
build:cuda --repo_env=HERMETIC_CUDNN_VERSION="9.3.0"
|
|
# This flag is needed to include CUDA libraries.
|
|
build:cuda --@local_config_cuda//cuda:include_cuda_libs=true
|
|
|
|
# This configuration is used for building the wheels.
|
|
build:cuda_wheel --@local_config_cuda//cuda:include_cuda_libs=false
|
|
|
|
# CUDA: This config refers to building CUDA op kernels with clang.
|
|
build:cuda_clang --config=cuda
|
|
build:cuda_clang --@local_config_cuda//:cuda_compiler=clang
|
|
build:cuda_clang --copt=-Qunused-arguments
|
|
# Select supported compute capabilities (supported graphics cards).
|
|
# This is the same as the official TensorFlow builds.
|
|
# See https://developer.nvidia.com/cuda-gpus#compute
|
|
# `compute_XY` enables PTX embedding in addition to SASS. PTX
|
|
# is forward compatible beyond the current compute capability major
|
|
# release while SASS is only forward compatible inside the current
|
|
# major release. Example: sm_80 kernels can run on sm_89 GPUs but
|
|
# not on sm_90 GPUs. compute_80 kernels though can also run on sm_90 GPUs.
|
|
build:cuda_clang --repo_env=HERMETIC_CUDA_COMPUTE_CAPABILITIES="sm_60,sm_70,sm_80,sm_89,compute_90"
|
|
# Set lld as the linker.
|
|
build:cuda_clang --host_linkopt="-fuse-ld=lld"
|
|
build:cuda_clang --host_linkopt="-lm"
|
|
build:cuda_clang --linkopt="-fuse-ld=lld"
|
|
build:cuda_clang --linkopt="-lm"
|
|
|
|
# Set up compilation CUDA version and paths and use the CUDA Clang toolchain.
|
|
build:cuda_clang_official --config=cuda_clang
|
|
build:cuda_clang_official --repo_env=HERMETIC_CUDA_VERSION="12.3.2"
|
|
build:cuda_clang_official --repo_env=HERMETIC_CUDNN_VERSION="9.3.0"
|
|
build:cuda_clang_official --action_env=CLANG_CUDA_COMPILER_PATH="/usr/lib/llvm-18/bin/clang"
|
|
build:cuda_clang_official --crosstool_top="@local_config_cuda//crosstool:toolchain"
|
|
|
|
# Build with nvcc for CUDA and clang for host
|
|
build:nvcc_clang --config=cuda
|
|
build:nvcc_clang --action_env=TF_NVCC_CLANG="1"
|
|
build:nvcc_clang --@local_config_cuda//:cuda_compiler=nvcc
|
|
|
|
|
|
# Debug config
|
|
build:dbg -c dbg
|
|
# Only include debug info for files under tensorflow/, excluding kernels, to
|
|
# reduce the size of the debug info in the binary. This is because if the debug
|
|
# sections in the ELF binary are too large, errors can occur. See
|
|
# https://github.com/tensorflow/tensorflow/issues/48919.
|
|
# Users can still include debug info for a specific kernel, e.g. with:
|
|
# --config=dbg --per_file_copt=+tensorflow/core/kernels/identity_op.*@-g
|
|
# Since this .bazelrc file is synced between the tensorflow/tensorflow repo and
|
|
# the openxla/xla repo, also include debug info for files under xla/.
|
|
build:dbg --per_file_copt=+.*,-tensorflow.*,-xla.*@-g0
|
|
build:dbg --per_file_copt=+tensorflow/core/kernels.*@-g0
|
|
# for now, disable arm_neon. see: https://github.com/tensorflow/tensorflow/issues/33360
|
|
build:dbg --cxxopt -DTF_LITE_DISABLE_X86_NEON
|
|
# AWS SDK must be compiled in release mode. see: https://github.com/tensorflow/tensorflow/issues/37498
|
|
build:dbg --copt -DDEBUG_BUILD
|
|
|
|
# Config to build TF TPU
|
|
build:tpu --define=with_tpu_support=true
|
|
build:tpu --define=framework_shared_object=true
|
|
build:tpu --copt=-DLIBTPU_ON_GCE
|
|
build:tpu --define=enable_mlir_bridge=true
|
|
|
|
build:rocm --crosstool_top=@local_config_rocm//crosstool:toolchain
|
|
build:rocm --define=using_rocm_hipcc=true
|
|
build:rocm --define=tensorflow_mkldnn_contraction_kernel=0
|
|
build:rocm --repo_env TF_NEED_ROCM=1
|
|
|
|
build:sycl --crosstool_top=@local_config_sycl//crosstool:toolchain
|
|
build:sycl --define=using_sycl=true
|
|
build:sycl --define=tensorflow_mkldnn_contraction_kernel=0
|
|
build:sycl --repo_env TF_NEED_SYCL=1
|
|
|
|
# Options to disable default on features
|
|
build:nogcp --define=no_gcp_support=true
|
|
build:nonccl --define=no_nccl_support=true
|
|
|
|
# Modular TF build options
|
|
build:dynamic_kernels --define=dynamic_loaded_kernels=true
|
|
build:dynamic_kernels --copt=-DAUTOLOAD_DYNAMIC_KERNELS
|
|
|
|
# Don't trigger --config=<host platform> when cross-compiling.
|
|
build:android --noenable_platform_specific_config
|
|
build:ios --noenable_platform_specific_config
|
|
|
|
# Suppress all C++ compiler warnings, otherwise build logs become 10s of MBs.
|
|
build:android --copt=-w
|
|
build:ios --copt=-w
|
|
build:linux --host_copt=-w
|
|
build:macos --copt=-w
|
|
build:windows --copt=/W0
|
|
build:windows --host_copt=/W0
|
|
|
|
# Suppress most C++ compiler warnings to reduce log size but allow
|
|
# for specific warnings to still be present.
|
|
build:linux --copt="-Wno-all"
|
|
build:linux --copt="-Wno-extra"
|
|
build:linux --copt="-Wno-deprecated"
|
|
build:linux --copt="-Wno-deprecated-declarations"
|
|
build:linux --copt="-Wno-ignored-attributes"
|
|
build:linux --copt="-Wno-array-bounds"
|
|
|
|
# Add unused-result as an error on Linux.
|
|
build:linux --copt="-Wunused-result"
|
|
build:linux --copt="-Werror=unused-result"
|
|
# Add switch as an error on Linux.
|
|
build:linux --copt="-Wswitch"
|
|
build:linux --copt="-Werror=switch"
|
|
# Required for building with clang
|
|
build:linux --copt="-Wno-error=unused-but-set-variable"
|
|
|
|
# Linux ARM64 specific options
|
|
build:linux_arm64 --copt="-mtune=generic" --copt="-march=armv8-a" --copt="-O3"
|
|
|
|
|
|
# On Windows, `__cplusplus` is wrongly defined without this switch
|
|
# See https://devblogs.microsoft.com/cppblog/msvc-now-correctly-reports-__cplusplus/
|
|
build:windows --copt=/Zc:__cplusplus
|
|
build:windows --host_copt=/Zc:__cplusplus
|
|
|
|
# Tensorflow uses M_* math constants that only get defined by MSVC headers if
|
|
# _USE_MATH_DEFINES is defined.
|
|
build:windows --copt=/D_USE_MATH_DEFINES
|
|
build:windows --host_copt=/D_USE_MATH_DEFINES
|
|
|
|
# Windows has a relatively short command line limit, which TF has begun to hit.
|
|
# See https://docs.bazel.build/versions/main/windows.html
|
|
build:windows --features=compiler_param_file
|
|
build:windows --features=archive_param_file
|
|
|
|
# Speed Windows compile times. Available in VS 16.4 (we are on 16.11). See
|
|
# https://groups.google.com/a/tensorflow.org/d/topic/build/SsW98Eo7l3o/discussion
|
|
build:windows --copt=/d2ReducedOptimizeHugeFunctions
|
|
build:windows --host_copt=/d2ReducedOptimizeHugeFunctions
|
|
|
|
# Before VS 2017 15.8, the member "type" would non-conformingly have an
|
|
# alignment of only alignof(max_align_t). VS 2017 15.8 was fixed to handle this
|
|
# correctly, but the fix inherently changes layout and breaks binary
|
|
# compatibility (*only* for uses of aligned_storage with extended alignments).
|
|
build:windows --copt=-D_ENABLE_EXTENDED_ALIGNED_STORAGE
|
|
build:windows --host_copt=-D_ENABLE_EXTENDED_ALIGNED_STORAGE
|
|
|
|
# Enable the runfiles symlink tree on Windows. This makes it possible to build
|
|
# the pip package on Windows without an intermediate data-file archive, as the
|
|
# build_pip_package script in its current form (as of Aug 2023) uses the
|
|
# runfiles symlink tree to decide what to put into the Python wheel.
|
|
startup --windows_enable_symlinks
|
|
build:windows --enable_runfiles
|
|
|
|
# Default paths for TF_SYSTEM_LIBS
|
|
build:linux --define=PREFIX=/usr
|
|
build:linux --define=LIBDIR=$(PREFIX)/lib
|
|
build:linux --define=INCLUDEDIR=$(PREFIX)/include
|
|
build:linux --define=PROTOBUF_INCLUDE_PATH=$(PREFIX)/include
|
|
build:macos --define=PREFIX=/usr
|
|
build:macos --define=LIBDIR=$(PREFIX)/lib
|
|
build:macos --define=INCLUDEDIR=$(PREFIX)/include
|
|
build:macos --define=PROTOBUF_INCLUDE_PATH=$(PREFIX)/include
|
|
# TF_SYSTEM_LIBS do not work on windows.
|
|
|
|
# By default, build TF in C++ 17 mode.
|
|
build:android --cxxopt=-std=c++17
|
|
build:android --host_cxxopt=-std=c++17
|
|
build:ios --cxxopt=-std=c++17
|
|
build:ios --host_cxxopt=-std=c++17
|
|
build:linux --cxxopt=-std=c++17
|
|
build:linux --host_cxxopt=-std=c++17
|
|
build:macos --cxxopt=-std=c++17
|
|
build:macos --host_cxxopt=-std=c++17
|
|
build:windows --cxxopt=/std:c++17
|
|
build:windows --host_cxxopt=/std:c++17
|
|
|
|
# On windows, we still link everything into a single DLL.
|
|
build:windows --config=monolithic
|
|
|
|
# On linux, we dynamically link small amount of kernels
|
|
build:linux --config=dynamic_kernels
|
|
|
|
# Make sure to include as little of windows.h as possible
|
|
build:windows --copt=-DWIN32_LEAN_AND_MEAN
|
|
build:windows --host_copt=-DWIN32_LEAN_AND_MEAN
|
|
build:windows --copt=-DNOGDI
|
|
build:windows --host_copt=-DNOGDI
|
|
|
|
# MSVC (Windows): Standards-conformant preprocessor mode
|
|
# See https://docs.microsoft.com/en-us/cpp/preprocessor/preprocessor-experimental-overview
|
|
build:windows --copt=/Zc:preprocessor
|
|
build:windows --host_copt=/Zc:preprocessor
|
|
|
|
# Misc build options we need for windows.
|
|
build:windows --linkopt=/DEBUG
|
|
build:windows --host_linkopt=/DEBUG
|
|
build:windows --linkopt=/OPT:REF
|
|
build:windows --host_linkopt=/OPT:REF
|
|
build:windows --linkopt=/OPT:ICF
|
|
build:windows --host_linkopt=/OPT:ICF
|
|
|
|
# Verbose failure logs when something goes wrong
|
|
build:windows --verbose_failures
|
|
|
|
# Work around potential issues with large command lines on windows.
|
|
# See: https://github.com/bazelbuild/bazel/issues/5163
|
|
build:windows --features=compiler_param_file
|
|
|
|
# Do not risk cache corruption. See:
|
|
# https://github.com/bazelbuild/bazel/issues/3360
|
|
build:linux --experimental_guard_against_concurrent_changes
|
|
|
|
# Configure short or long logs
|
|
build:short_logs --output_filter=DONT_MATCH_ANYTHING
|
|
build:verbose_logs --output_filter=
|
|
|
|
# Instruction set optimizations
|
|
# TODO(gunan): Create a feature in toolchains for avx/avx2 to
|
|
# avoid having to define linux/win separately.
|
|
build:avx_linux --copt=-mavx
|
|
build:avx_linux --host_copt=-mavx
|
|
build:avx_win --copt=/arch:AVX
|
|
|
|
# Use Clang-cl compiler on Windows
|
|
build:win_clang --copt=/clang:-Weverything
|
|
build:win_clang --extra_toolchains=@local_config_cc//:cc-toolchain-x64_windows-clang-cl
|
|
build:win_clang --extra_execution_platforms=//tensorflow/tools/toolchains/win:x64_windows-clang-cl
|
|
build:win_clang --host_platform=//tensorflow/tools/toolchains/win:x64_windows-clang-cl
|
|
build:win_clang --compiler=clang-cl
|
|
build:win_clang --linkopt=/FORCE:MULTIPLE
|
|
build:win_clang --host_linkopt=/FORCE:MULTIPLE
|
|
test:win_clang --linkopt=/FORCE:MULTIPLE
|
|
test:win_clang --host_linkopt=/FORCE:MULTIPLE
|
|
|
|
# Same config as above but for XLA, which has different toolchain paths
|
|
build:win_clang_xla --copt=/clang:-Weverything
|
|
build:win_clang_xla --extra_toolchains=@local_config_cc//:cc-toolchain-x64_windows-clang-cl
|
|
build:win_clang_xla --extra_execution_platforms=//tools/toolchains/win:x64_windows-clang-cl
|
|
build:win_clang_xla --host_platform=//tools/toolchains/win:x64_windows-clang-cl
|
|
build:win_clang_xla --compiler=clang-cl
|
|
build:win_clang_xla --linkopt=/FORCE:MULTIPLE
|
|
build:win_clang_xla --host_linkopt=/FORCE:MULTIPLE
|
|
test:win_clang_xla --action_env=PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC;.PY;.PYW
|
|
test:win_clang_xla --linkopt=/FORCE:MULTIPLE
|
|
test:win_clang_xla --host_linkopt=/FORCE:MULTIPLE
|
|
|
|
# Options to build TensorFlow 1.x or 2.x.
|
|
# TODO(kanglan): Change v2's define to default behavior
|
|
build:v2 --define=tf_api_version=2 --action_env=TF2_BEHAVIOR=1
|
|
|
|
# Enable all targets in XLA
|
|
build:cpu_cross --define=with_cross_compiler_support=true
|
|
|
|
# Disable XLA on mobile.
|
|
build:xla --define=with_xla_support=true # TODO: remove, it's on by default.
|
|
build:android --define=with_xla_support=false
|
|
build:ios --define=with_xla_support=false
|
|
|
|
# BEGIN TF REMOTE BUILD EXECUTION OPTIONS
|
|
# Options when using remote execution
|
|
# WARNING: THESE OPTIONS WONT WORK IF YOU DO NOT HAVE PROPER AUTHENTICATION AND PERMISSIONS
|
|
|
|
# Allow creation of resultstore URLs for any bazel invocation
|
|
build:resultstore --google_default_credentials
|
|
build:resultstore --bes_backend=buildeventservice.googleapis.com
|
|
build:resultstore --bes_instance_name="tensorflow-testing"
|
|
build:resultstore --bes_results_url="https://source.cloud.google.com/results/invocations"
|
|
build:resultstore --bes_timeout=600s
|
|
|
|
# Flag to enable remote config
|
|
common --experimental_repo_remote_exec
|
|
|
|
# Make Bazel not try to probe the host system for a C++ toolchain.
|
|
build:rbe_base --config=resultstore
|
|
build:rbe_base --repo_env=BAZEL_DO_NOT_DETECT_CPP_TOOLCHAIN=1
|
|
build:rbe_base --define=EXECUTOR=remote
|
|
build:rbe_base --jobs=800
|
|
build:rbe_base --remote_executor=grpcs://remotebuildexecution.googleapis.com
|
|
build:rbe_base --remote_timeout=3600
|
|
build:rbe_base --spawn_strategy=remote,worker,standalone,local
|
|
# Attempt to minimize the amount of data transfer between bazel and the remote
|
|
# workers:
|
|
build:rbe_base --remote_download_toplevel
|
|
test:rbe_base --test_env=USER=anon
|
|
|
|
# TODO(kanglan): Check if we want to merge rbe_linux into rbe_linux_cpu.
|
|
build:rbe_linux --config=rbe_base
|
|
build:rbe_linux --action_env=PATH="/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/go/bin"
|
|
# Non-rbe settings we should include because we do not run configure
|
|
build:rbe_linux --config=avx_linux
|
|
# TODO(gunan): Check why we need this specified in rbe, but not in other builds.
|
|
build:rbe_linux --linkopt=-lrt
|
|
build:rbe_linux --host_linkopt=-lrt
|
|
build:rbe_linux --linkopt=-lm
|
|
build:rbe_linux --host_linkopt=-lm
|
|
|
|
build:rbe_linux_cpu --config=rbe_linux
|
|
# Linux cpu and cuda builds share the same toolchain now.
|
|
build:rbe_linux_cpu --host_crosstool_top="@local_config_cuda//crosstool:toolchain"
|
|
build:rbe_linux_cpu --crosstool_top="@local_config_cuda//crosstool:toolchain"
|
|
build:rbe_linux_cpu --extra_toolchains="@local_config_cuda//crosstool:toolchain-linux-x86_64"
|
|
build:rbe_linux_cpu --repo_env=CC="/usr/lib/llvm-18/bin/clang"
|
|
build:rbe_linux_cpu --repo_env=TF_SYSROOT="/dt9"
|
|
build:rbe_linux_cpu --extra_execution_platforms="@sigbuild-r2.17-clang_config_platform//:platform"
|
|
build:rbe_linux_cpu --host_platform="@sigbuild-r2.17-clang_config_platform//:platform"
|
|
build:rbe_linux_cpu --platforms="@sigbuild-r2.17-clang_config_platform//:platform"
|
|
# This is needed for all Clang17 builds but must not be present in GCC builds.
|
|
build:rbe_linux_cpu --copt=-Wno-error=unused-command-line-argument
|
|
# This was added in clang-16 by https://reviews.llvm.org/D133574.
|
|
# Can be removed once upb is updated, since a type definition is used within
|
|
# offset of in the current version of ubp.
|
|
# See https://github.com/protocolbuffers/upb/blob/9effcbcb27f0a665f9f345030188c0b291e32482/upb/upb.c#L183.
|
|
build:rbe_linux_cpu --copt=-Wno-gnu-offsetof-extensions
|
|
# Python config is the same across all containers because the binary is the same
|
|
build:rbe_linux_cpu --python_path="/usr/bin/python3"
|
|
# These you may need to change for your own GCP project.
|
|
common:rbe_linux_cpu --remote_instance_name=projects/tensorflow-testing/instances/default_instance
|
|
|
|
# TODO(kanglan): Remove it after toolchain update is complete.
|
|
build:rbe_linux_cpu_old --config=rbe_linux
|
|
build:rbe_linux_cpu_old --host_crosstool_top="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_cuda//crosstool:toolchain"
|
|
build:rbe_linux_cpu_old --crosstool_top="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_cuda//crosstool:toolchain"
|
|
build:rbe_linux_cpu_old --extra_toolchains="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_cuda//crosstool:toolchain-linux-x86_64"
|
|
build:rbe_linux_cpu_old --extra_execution_platforms="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_platform//:platform"
|
|
build:rbe_linux_cpu_old --host_platform="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_platform//:platform"
|
|
build:rbe_linux_cpu_old --platforms="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_platform//:platform"
|
|
build:rbe_linux_cpu_old --python_path="/usr/local/bin/python3.9"
|
|
common:rbe_linux_cpu_old --remote_instance_name=projects/tensorflow-testing/instances/default_instance
|
|
|
|
build:rbe_linux_cuda --config=cuda_clang_official
|
|
build:rbe_linux_cuda --config=rbe_linux_cpu
|
|
# For Remote build execution -- GPU configuration
|
|
build:rbe_linux_cuda --repo_env=REMOTE_GPU_TESTING=1
|
|
|
|
build:rbe_linux_cuda_nvcc --config=rbe_linux_cuda
|
|
build:rbe_linux_cuda_nvcc --config=nvcc_clang
|
|
build:rbe_linux_cuda_nvcc --repo_env TF_NCCL_USE_STUB=1
|
|
|
|
build:rbe_win_base --config=rbe_base
|
|
build:rbe_win_base --shell_executable=C:\\tools\\msys64\\usr\\bin\\bash.exe
|
|
build:rbe_win_base --remote_instance_name=projects/tensorflow-testing/instances/windows
|
|
# Don't build the python zip archive in the RBE build.
|
|
build:rbe_win_base --remote_download_minimal
|
|
build:rbe_win_base --enable_runfiles
|
|
build:rbe_win_base --nobuild_python_zip
|
|
build:rbe_win_base --define=override_eigen_strong_inline=true
|
|
|
|
build:rbe_win_clang --config=rbe_win_base
|
|
build:rbe_win_clang --crosstool_top="//tensorflow/tools/toolchains/win/20240424:toolchain"
|
|
build:rbe_win_clang --extra_toolchains="//tensorflow/tools/toolchains/win/20240424:cc-toolchain-x64_windows-clang-cl"
|
|
build:rbe_win_clang --extra_execution_platforms="//tensorflow/tools/toolchains/win:x64_windows-clang-cl"
|
|
build:rbe_win_clang --host_platform="//tensorflow/tools/toolchains/win:x64_windows-clang-cl"
|
|
build:rbe_win_clang --platforms="//tensorflow/tools/toolchains/win:x64_windows-clang-cl"
|
|
build:rbe_win_clang --compiler=clang-cl
|
|
build:rbe_win_clang --linkopt=/FORCE:MULTIPLE
|
|
build:rbe_win_clang --host_linkopt=/FORCE:MULTIPLE
|
|
|
|
# TODO(belitskiy): Rename `rbe_win_clang` to this, once done switching presubmits.
|
|
build:rbe_windows_x86_cpu --config=rbe_win_clang
|
|
|
|
# END TF REMOTE BUILD EXECUTION OPTIONS
|
|
|
|
# TFLite build configs for generic embedded Linux
|
|
build:elinux --crosstool_top=@local_config_embedded_arm//:toolchain
|
|
build:elinux --host_crosstool_top=@bazel_tools//tools/cpp:toolchain
|
|
build:elinux_aarch64 --config=elinux
|
|
build:elinux_aarch64 --cpu=aarch64
|
|
build:elinux_armhf --config=elinux
|
|
build:elinux_armhf --cpu=armhf
|
|
build:elinux_armhf --copt -mfp16-format=ieee
|
|
|
|
# Config-specific options should come above this line.
|
|
|
|
# Load rc file written by ./configure.
|
|
try-import %workspace%/.tf_configure.bazelrc
|
|
try-import %workspace%/xla_configure.bazelrc
|
|
|
|
# Load rc file with user-specific options.
|
|
try-import %workspace%/.bazelrc.user
|
|
|
|
# Here are bazelrc configs for release builds
|
|
# Build TensorFlow v2.
|
|
test:release_base --test_size_filters=small,medium
|
|
|
|
# Enable support for all targets
|
|
build:release_base --config=cpu_cross
|
|
|
|
# Ensure release_base is set on linux
|
|
build:release_linux_base --config=release_base
|
|
|
|
# Disable clang extension that rejects type definitions within offsetof.
|
|
# This was added in clang-16 by https://reviews.llvm.org/D133574.
|
|
# Can be removed once upb is updated, since a type definition is used within
|
|
# offset of in the current version of ubp.
|
|
# See https://github.com/protocolbuffers/upb/blob/9effcbcb27f0a665f9f345030188c0b291e32482/upb/upb.c#L183.
|
|
build:release_linux_base --copt=-Wno-gnu-offsetof-extensions
|
|
build:release_linux_base --copt=-Wno-error=array-parameter
|
|
build:release_linux_base --copt=-Wno-error=unused-command-line-argument
|
|
# Set lld as the linker.
|
|
build:release_linux_base --linkopt="-fuse-ld=lld"
|
|
build:release_linux_base --linkopt="-lm"
|
|
|
|
# We have some invalid linker scripts in the build,
|
|
# so we need to disable this check
|
|
build:release_linux_base --linkopt=-Wl,--undefined-version
|
|
|
|
# Container environment settings below this point.
|
|
# Use Python 3.X as installed in container image
|
|
build:release_linux_base --action_env PYTHON_BIN_PATH="/usr/bin/python3"
|
|
build:release_linux_base --action_env PYTHON_LIB_PATH="/usr/lib/tf_python"
|
|
build:release_linux_base --python_path="/usr/bin/python3"
|
|
# Set Clang as compiler. Use the actual path to clang installed in container.
|
|
build:release_linux_base --repo_env=CC="/usr/lib/llvm-18/bin/clang"
|
|
build:release_linux_base --repo_env=BAZEL_COMPILER="/usr/lib/llvm-18/bin/clang"
|
|
# Test-related settings below this point.
|
|
test:release_linux_base --build_tests_only --keep_going --test_output=errors --verbose_failures=true
|
|
test:release_linux_base --local_test_jobs=HOST_CPUS
|
|
# Give only the list of failed tests at the end of the log
|
|
test:release_linux_base --test_summary=short
|
|
|
|
# Use the Clang toolchain to compile
|
|
build:release_cpu_linux --config=release_linux_base
|
|
build:release_cpu_linux --crosstool_top="@local_config_cuda//crosstool:toolchain"
|
|
build:release_cpu_linux --repo_env=TF_SYSROOT="/dt9"
|
|
# Target the AVX instruction set
|
|
build:release_cpu_linux --config=avx_linux
|
|
|
|
build:release_gpu_linux --config=release_cpu_linux
|
|
# Set up compilation CUDA version and paths and use the CUDA Clang toolchain.
|
|
# Note that linux cpu and cuda builds share the same toolchain now.
|
|
build:release_gpu_linux --config=cuda_clang_official
|
|
# Local test jobs has to be 4 because parallel_gpu_execute is fragile, I think
|
|
test:release_gpu_linux --test_timeout=300,450,1200,3600 --local_test_jobs=4 --run_under=//tensorflow/tools/ci_build/gpu_build:parallel_gpu_execute
|
|
|
|
build:release_arm64_linux --config=release_linux_base
|
|
build:release_arm64_linux --config=linux_arm64
|
|
build:release_arm64_linux --crosstool_top="@ml2014_clang_aarch64_config_aarch64//crosstool:toolchain"
|
|
build:release_arm64_linux --config=mkl_aarch64_threadpool
|
|
build:release_arm64_linux --copt=-flax-vector-conversions
|
|
test:release_arm64_linux --flaky_test_attempts=3
|
|
|
|
build:release_cpu_macos --config=avx_linux
|
|
|
|
# Base build configs for macOS
|
|
build:release_macos_base --action_env DEVELOPER_DIR=/Applications/Xcode.app/Contents/Developer
|
|
build:release_macos_base --define=no_nccl_support=true --output_filter=^$
|
|
|
|
# Ensure release_base is set on mac
|
|
build:release_macos_base --config=release_base
|
|
|
|
# Build configs for macOS x86
|
|
build:release_macos_x86 --config=release_macos_base
|
|
# Build with the AVX instruction set when on macOS x86
|
|
build:release_macos_x86 --config=avx_linux
|
|
build:release_macos_x86 --cpu=darwin
|
|
# Target Catalina as the minimum compatible OS version
|
|
build:release_macos_x86 --macos_minimum_os=10.15
|
|
build:release_macos_x86 --action_env MACOSX_DEPLOYMENT_TARGET=10.15
|
|
|
|
# Build configs for macOS Arm64
|
|
build:release_macos_arm64 --config=release_macos_base
|
|
build:release_macos_arm64 --cpu=darwin_arm64
|
|
build:release_macos_arm64 --define=tensorflow_mkldnn_contraction_kernel=0
|
|
# Target Moneterey as the minimum compatible OS version
|
|
build:release_macos_arm64 --macos_minimum_os=12.0
|
|
build:release_macos_arm64 --action_env MACOSX_DEPLOYMENT_TARGET=12.0
|
|
|
|
# Base test configs for macOS
|
|
test:release_macos_base --verbose_failures=true --local_test_jobs=HOST_CPUS
|
|
test:release_macos_base --test_timeout=300,450,1200,3600 --test_output=errors
|
|
test:release_macos_base --build_tests_only --keep_going
|
|
test:release_macos_base --flaky_test_attempts=3
|
|
|
|
# Test configs for macOS x86
|
|
test:release_macos_x86 --config=release_macos_base
|
|
|
|
# Test configs for macOS Arm64
|
|
test:release_macos_arm64 --config=release_macos_base
|
|
|
|
# Ensure release_base is set on windows
|
|
build:release_cpu_windows --config=release_base
|
|
|
|
# TODO(kanglan): Update windows configs after b/289091160 is fixed
|
|
build:release_cpu_windows --config=avx_win
|
|
build:release_cpu_windows --define=no_tensorflow_py_deps=true
|
|
|
|
# Exclude TFRT integration for anything but Linux.
|
|
build:android --config=no_tfrt
|
|
build:macos --config=no_tfrt
|
|
build:windows --config=no_tfrt
|
|
build:rocm --config=no_tfrt
|
|
build:no_tfrt --deleted_packages=tensorflow/compiler/mlir/tfrt,tensorflow/compiler/mlir/tfrt/benchmarks,tensorflow/compiler/mlir/tfrt/ir,tensorflow/compiler/mlir/tfrt/ir/mlrt,tensorflow/compiler/mlir/tfrt/jit/python_binding,tensorflow/compiler/mlir/tfrt/jit/transforms,tensorflow/compiler/mlir/tfrt/python_tests,tensorflow/compiler/mlir/tfrt/tests,tensorflow/compiler/mlir/tfrt/tests/ifrt,tensorflow/compiler/mlir/tfrt/tests/mlrt,tensorflow/compiler/mlir/tfrt/tests/ir,tensorflow/compiler/mlir/tfrt/tests/analysis,tensorflow/compiler/mlir/tfrt/tests/jit,tensorflow/compiler/mlir/tfrt/tests/lhlo_to_tfrt,tensorflow/compiler/mlir/tfrt/tests/lhlo_to_jitrt,tensorflow/compiler/mlir/tfrt/tests/tf_to_corert,tensorflow/compiler/mlir/tfrt/tests/tf_to_tfrt_data,tensorflow/compiler/mlir/tfrt/tests/saved_model,tensorflow/compiler/mlir/tfrt/transforms/lhlo_gpu_to_tfrt_gpu,tensorflow/compiler/mlir/tfrt/transforms/mlrt,tensorflow/core/runtime_fallback,tensorflow/core/runtime_fallback/conversion,tensorflow/core/runtime_fallback/kernel,tensorflow/core/runtime_fallback/opdefs,tensorflow/core/runtime_fallback/runtime,tensorflow/core/runtime_fallback/util,tensorflow/core/runtime_fallback/test,tensorflow/core/runtime_fallback/test/gpu,tensorflow/core/runtime_fallback/test/saved_model,tensorflow/core/runtime_fallback/test/testdata,tensorflow/core/tfrt/stubs,tensorflow/core/tfrt/tfrt_session,tensorflow/core/tfrt/mlrt,tensorflow/core/tfrt/mlrt/attribute,tensorflow/core/tfrt/mlrt/kernel,tensorflow/core/tfrt/mlrt/bytecode,tensorflow/core/tfrt/mlrt/interpreter,tensorflow/compiler/mlir/tfrt/translate/mlrt,tensorflow/compiler/mlir/tfrt/translate/mlrt/testdata,tensorflow/core/tfrt/gpu,tensorflow/core/tfrt/run_handler_thread_pool,tensorflow/core/tfrt/runtime,tensorflow/core/tfrt/saved_model,tensorflow/core/tfrt/graph_executor,tensorflow/core/tfrt/saved_model/tests,tensorflow/core/tfrt/tpu,tensorflow/core/tfrt/utils,tensorflow/core/tfrt/utils/debug,tensorflow/core/tfrt/saved_model/python,tensorflow/core/tfrt/graph_executor/python,tensorflow/core/tfrt/saved_model/utils
|
|
|
|
# BEGIN TF CACHE HELPER OPTIONS
|
|
# Options when using remote execution
|
|
# WARNING: THESE OPTIONS WONT WORK IF YOU DO NOT HAVE PROPER AUTHENTICATION AND PERMISSIONS
|
|
|
|
# Use --config=tf_public_cache to try and use the TensorFlow public build cache
|
|
# to build TensorFlow. Look at ci/official/envs to find which types of jobs
|
|
# push to the cache. For macOS, use --config=tf_public_macos_cache
|
|
build:tf_public_cache --remote_cache="https://storage.googleapis.com/tensorflow-devinfra-bazel-cache/january2024" --remote_upload_local_results=false
|
|
# Cache pushes are limited to TF's CI system.
|
|
build:tf_public_cache_push --config=tf_public_cache --remote_upload_local_results=true --google_default_credentials
|
|
# Public cache for macOS builds
|
|
build:tf_public_macos_cache --remote_cache="https://storage.googleapis.com/tensorflow-macos-bazel-cache/oct2023" --remote_upload_local_results=false
|
|
# Cache pushes are limited to TF's CI system.
|
|
build:tf_public_macos_cache_push --config=tf_public_macos_cache --remote_upload_local_results=true --google_default_credentials
|
|
|
|
# END TF CACHE HELPER OPTIONS
|
|
# BEGIN TF TEST SUITE OPTIONS
|
|
# These are convenience config options that effectively declare TF's CI test suites. Look
|
|
# at the scripts of ci/official/ to see how TF's CI uses them.
|
|
|
|
# LIBTENSORFLOW TESTS are for building Libtensorflow archives. These are CUDA/CPU-agnostic.
|
|
test:linux_libtensorflow_test --config=cuda_wheel -- //tensorflow/tools/lib_package:libtensorflow_test //tensorflow/tools/lib_package:libtensorflow_java_test
|
|
build:linux_libtensorflow_build --config=cuda_wheel -- //tensorflow/tools/lib_package:libtensorflow.tar.gz //tensorflow/tools/lib_package:libtensorflow_jni.tar.gz //tensorflow/java:libtensorflow.jar //tensorflow/java:libtensorflow-src.jar //tensorflow/tools/lib_package:libtensorflow_proto.zip
|
|
|
|
# PYTHON TESTS run a suite of Python tests intended for verifying that the Python wheel
|
|
# will work properly. These are usually run Nightly or upon Release.
|
|
# CPU WHEEL
|
|
test:linux_cpu_wheel_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_cpu_wheel_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_cpu_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
|
|
test:linux_cpu_wheel_test --config=linux_cpu_wheel_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
|
|
# CUDA WHEEL
|
|
test:linux_cuda_wheel_test_filters --test_tag_filters=gpu,requires-gpu,-no_gpu,-no_oss,-oss_excluded,-oss_serial,-benchmark-test,-no_cuda11,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_cuda_wheel_test_filters --build_tag_filters=gpu,requires-gpu,-no_gpu,-no_oss,-oss_excluded,-oss_serial,-benchmark-test,-no_cuda11,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_cuda_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
|
|
test:linux_cuda_wheel_test --config=linux_cuda_wheel_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
|
|
# ARM64 WHEEL
|
|
test:linux_arm64_wheel_test_filters --test_tag_filters=-no_oss,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_arm64_wheel_test_filters --build_tag_filters=-no_oss,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310
|
|
test:linux_arm64_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
|
|
test:linux_arm64_wheel_test --config=linux_arm64_wheel_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/core/grappler/optimizers:auto_mixed_precision_test_cpu -//tensorflow/core/grappler/optimizers:remapper_test_cpu -//tensorflow/core/kernels/image:resize_bicubic_op_test -//tensorflow/compiler/mlir/tfr/examples/customization:test_ops_test -//tensorflow/compiler/mlir/tfr/examples/mnist:mnist_ops_test -//tensorflow/compiler/mlir/tfr/examples/pad:pad_ops_test
|
|
# MACOS ARM64 WHEEL
|
|
test:macos_arm64_wheel_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test,-no_mac_arm64,-no_aarch64
|
|
test:macos_arm64_wheel_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test,-no_mac_arm64,-no_aarch64
|
|
test:macos_arm64_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
|
|
test:macos_arm64_wheel_test --config=macos_arm64_wheel_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/compiler/aot/...
|
|
# MACOS X86 WHEEL
|
|
test:macos_x86_wheel_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py38,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test
|
|
test:macos_x86_wheel_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py38,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test
|
|
test:macos_x86_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
|
|
test:macos_x86_wheel_test --config=macos_x86_wheel_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/compiler/aot/...
|
|
|
|
# PYCPP TESTS run a suite of Python and C++ tests to verify general correctness over
|
|
# the whole TF code base. These are usually run continuously or upon presubmit.
|
|
# LINUX CPU PYCPP:
|
|
test:linux_cpu_pycpp_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only
|
|
test:linux_cpu_pycpp_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only
|
|
test:linux_cpu_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium
|
|
test:linux_cpu_pycpp_test --config=linux_cpu_pycpp_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
|
|
|
|
# LINUX CUDA PYCPP:
|
|
test:linux_cuda_pycpp_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-benchmark-test,-v1only,gpu,-no_gpu,-no_gpu_presubmit,-no_cuda11
|
|
test:linux_cuda_pycpp_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-benchmark-test,-v1only,gpu,-no_gpu,-no_gpu_presubmit,-no_cuda11
|
|
test:linux_cuda_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium
|
|
test:linux_cuda_pycpp_test --config=linux_cuda_pycpp_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
|
|
|
|
# LINUX ARM64 PYCPP
|
|
# In Linux Arm64 presubmit/continuous build, we cross-compile the binaries on
|
|
# Linux x86 so that we can use RBE. Since tests still need to run on the single
|
|
# host Arm64 machine, the build becomes too slow (~30 min) to be a presubmit.
|
|
# For testing purposes, we want to see the runtime performance of an
|
|
# experimental job that is build-only, i.e, we only build the test targets and
|
|
# do not run them. By prefixing the configs with "build", we can run both
|
|
# `bazel build` and `bazel test` commands with the same config as test configs
|
|
# inherit from build.
|
|
build:linux_arm64_pycpp_test_filters --test_tag_filters=-no_oss,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only
|
|
build:linux_arm64_pycpp_test_filters --build_tag_filters=-no_oss,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only
|
|
build:linux_arm64_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium --flaky_test_attempts=3
|
|
# TODO(michaelhudgins): Why do we need to specifically omit go and java here?
|
|
build:linux_arm64_pycpp_test --config=linux_arm64_pycpp_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/core/grappler/optimizers:auto_mixed_precision_test_cpu -//tensorflow/core/grappler/optimizers:remapper_test_cpu -//tensorflow/core/kernels/image:resize_bicubic_op_test -//tensorflow/compiler/mlir/tfr/examples/customization:test_ops_test -//tensorflow/compiler/mlir/tfr/examples/mnist:mnist_ops_test -//tensorflow/compiler/mlir/tfr/examples/pad:pad_ops_test -//tensorflow/python/tools:aot_compiled_test
|
|
# CROSS-COMPILE ARM64 PYCPP
|
|
build:cross_compile_linux_arm64_pycpp_test --config=linux_arm64_pycpp_test
|
|
# Tests that fail only when cross-compiled
|
|
build:cross_compile_linux_arm64_pycpp_test -//tensorflow/compiler/mlir/quantization/stablehlo:convert_tf_quant_to_mhlo_int_test
|
|
# MACOS ARM64 PYCPP
|
|
test:macos_arm64_pycpp_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test,-no_mac_arm64,-no_aarch64
|
|
test:macos_arm64_pycpp_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test,-no_mac_arm64,-no_aarch64
|
|
test:macos_arm64_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium
|
|
test:macos_arm64_pycpp_test --config=macos_arm64_pycpp_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/compiler/aot/... -//tensorflow/core/kernels/image:resize_bicubic_op_test
|
|
# MACOS X86 PYCPP
|
|
# These are defined as build configs so that we can run a build only job. See
|
|
# the note under "ARM64 PYCPP" for more details.
|
|
build:macos_x86_pycpp_test_filters --test_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py38,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test
|
|
build:macos_x86_pycpp_test_filters --build_tag_filters=-no_oss,-oss_excluded,-oss_serial,-no_oss_py38,-no_oss_py39,-no_oss_py310,-nomac,-no_mac,-mac_excluded,-v1only,-gpu,-tpu,-benchmark-test
|
|
build:macos_x86_pycpp_test_filters --keep_going --test_lang_filters=cc,py --test_size_filters=small,medium
|
|
build:macos_x86_pycpp_test --config=macos_x86_pycpp_test_filters -- //tensorflow/... -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/go/... -//tensorflow/java/... -//tensorflow/tools/toolchains/... -//tensorflow/lite/... -//tensorflow/compiler/aot/...
|
|
# CROSS-COMPILE MACOS X86 PYCPP
|
|
build:cross_compile_macos_x86_pycpp_test --config=macos_x86_pycpp_test
|
|
build:cross_compile_macos_x86_pycpp_test -//tensorflow/core/kernels:quantized_conv_ops_test -//tensorflow/core/kernels:quantized_matmul_op_test -//tensorflow/python/ops:quantized_conv_ops_test -//tensorflow/tools/graph_transforms:transforms_test -//tensorflow/python/tools:aot_compiled_test
|
|
# WINDOWS X86-64 CPU PYCPP
|
|
test:windows_x86_cpu_pycpp_test_filters --test_tag_filters=-no_windows,-windows_excluded,-no_oss,-oss_excluded,-gpu,-tpu,-benchmark-test
|
|
test:windows_x86_cpu_pycpp_test_filters --build_tag_filters=-no_windows,-windows_excluded,-no_oss,-oss_excluded,-benchmark-test
|
|
test:windows_x86_cpu_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium --test_timeout="300,450,1200,3600"
|
|
test:windows_x86_cpu_pycpp_test_opts --copt=/d2ReducedOptimizeHugeFunctions --host_copt=/d2ReducedOptimizeHugeFunctions --dynamic_mode=off --build_tests_only
|
|
test:windows_x86_cpu_pycpp_test --config=windows_x86_cpu_pycpp_test_opts --config=windows_x86_cpu_pycpp_test_filters -- //tensorflow/... -//tensorflow/java/... -//tensorflow/lite/... -//tensorflow/compiler/...
|
|
|
|
# END TF TEST SUITE OPTIONS
|
|
|
|
# START CROSS-COMPILE CONFIGS
|
|
# Set execution platform to Linux x86
|
|
# Note: Lot of the "host_" flags such as "host_cpu" and "host_crosstool_top"
|
|
# flags seem to be actually used to specify the execution platform details. It
|
|
# seems it is this way because these flags are old and predate the distinction
|
|
# between host and execution platform.
|
|
build:cross_compile_base --host_cpu=k8
|
|
build:cross_compile_base --host_crosstool_top=//tensorflow/tools/toolchains/cross_compile/cc:cross_compile_toolchain_suite
|
|
build:cross_compile_base --extra_execution_platforms=//tensorflow/tools/toolchains/cross_compile/config:linux_x86_64
|
|
|
|
# XLA related settings for cross-compiled build. Certain paths are
|
|
# different in the XLA repo.
|
|
build:cross_compile_base_xla --host_cpu=k8
|
|
build:cross_compile_base_xla --host_crosstool_top=//tools/toolchains/cross_compile/cc:cross_compile_toolchain_suite
|
|
build:cross_compile_base_xla --extra_execution_platforms=//tools/toolchains/cross_compile/config:linux_x86_64
|
|
|
|
build:rbe_cross_compile_base --config=rbe_base
|
|
build:rbe_cross_compile_base --remote_instance_name=projects/tensorflow-testing/instances/default_instance
|
|
|
|
# XLA depends on some local Python headers that are configured as Genrule. They
|
|
# are present on the local host machine but not on the remote execution machine,
|
|
# leading to build failures. To resolve the issue, the following line is added
|
|
# to make sure all Genrule targets are excuted locally.
|
|
build:rbe_cross_compile_base_xla --config=rbe_cross_compile_base
|
|
build:rbe_cross_compile_base_xla --strategy=Genrule=standalone
|
|
|
|
# Due to the above strategy, all Genrule commands are executed locally, but the
|
|
# following actions invoke tools (E.g `flatc`, `llvm-tblgen`, etc.) that are
|
|
# only executabe on the RBE (x86) machine, so the strategy_regexp options are
|
|
# added to override and run the actions using remote strategy.
|
|
build:rbe_cross_compile_base_xla --strategy_regexp='Generating code from table.*=remote'
|
|
build:rbe_cross_compile_base_xla --strategy_regexp='Generating flatbuffer files.*=remote'
|
|
build:rbe_cross_compile_base_xla --strategy_regexp='Executing genrule @llvm-project.*=remote'
|
|
|
|
# Test-related settings below this point
|
|
# We cannot run cross-compiled tests on the remote Linux x86 VMs so we need to
|
|
# force all tests to run locally on the Aarch64 host.
|
|
test:rbe_cross_compile_base --strategy=TestRunner=local --build_tests_only
|
|
test:rbe_cross_compile_base --verbose_failures=true --local_test_jobs=HOST_CPUS --test_output=errors
|
|
|
|
test:rbe_cross_compile_base_xla --config=rbe_cross_compile_base
|
|
|
|
# START LINUX AARCH64 CROSS-COMPILE CONFIGS
|
|
build:cross_compile_linux_arm64 --config=cross_compile_base
|
|
|
|
# Set the target CPU to Aarch64
|
|
build:cross_compile_linux_arm64 --platforms=//tensorflow/tools/toolchains/cross_compile/config:linux_aarch64
|
|
build:cross_compile_linux_arm64 --cpu=aarch64
|
|
build:cross_compile_linux_arm64 --crosstool_top=//tensorflow/tools/toolchains/cross_compile/cc:cross_compile_toolchain_suite
|
|
|
|
# XLA uses different paths for platforms and crosstool_top.
|
|
build:cross_compile_linux_arm64_xla --config=cross_compile_base_xla
|
|
build:cross_compile_linux_arm64_xla --platforms=//tools/toolchains/cross_compile/config:linux_aarch64
|
|
build:cross_compile_linux_arm64_xla --crosstool_top=//tools/toolchains/cross_compile/cc:cross_compile_toolchain_suite
|
|
|
|
# RBE cross-compile configs for Linux Aarch64
|
|
build:rbe_cross_compile_linux_arm64 --config=cross_compile_linux_arm64
|
|
build:rbe_cross_compile_linux_arm64 --config=rbe_cross_compile_base
|
|
test:rbe_cross_compile_linux_arm64 --config=rbe_cross_compile_base
|
|
|
|
# RBE cross-compile configs for XLA Linux Aarch64
|
|
build:rbe_cross_compile_linux_arm64_xla --config=cross_compile_linux_arm64_xla
|
|
build:rbe_cross_compile_linux_arm64_xla --config=rbe_cross_compile_base_xla
|
|
test:rbe_cross_compile_linux_arm64_xla --config=rbe_cross_compile_base_xla
|
|
|
|
# END LINUX AARCH64 CROSS-COMPILE CONFIGS
|
|
|
|
# START MACOS CROSS-COMPILE CONFIGS
|
|
build:cross_compile_macos_x86 --config=cross_compile_base
|
|
build:cross_compile_macos_x86 --config=nonccl
|
|
# Target Catalina (10.15) as the minimum supported OS
|
|
build:cross_compile_macos_x86 --action_env MACOSX_DEPLOYMENT_TARGET=10.15
|
|
|
|
# Set the target CPU to Darwin x86
|
|
build:cross_compile_macos_x86 --platforms=//tensorflow/tools/toolchains/cross_compile/config:darwin_x86_64
|
|
build:cross_compile_macos_x86 --cpu=darwin
|
|
build:cross_compile_macos_x86 --crosstool_top=//tensorflow/tools/toolchains/cross_compile/cc:cross_compile_toolchain_suite
|
|
# When RBE cross-compiling for macOS, we need to explicitly register the
|
|
# toolchain. Otherwise, oddly, RBE complains that a "docker container must be
|
|
# specified".
|
|
build:cross_compile_macos_x86 --extra_toolchains=//tensorflow/tools/toolchains/cross_compile/config:macos-x86-cross-compile-cc-toolchain
|
|
# Map --platforms=darwin_x86_64 to --cpu=darwin and vice-versa to make selects()
|
|
# and transistions that use these flags work.
|
|
build:cross_compile_macos_x86 --platform_mappings=tensorflow/tools/toolchains/cross_compile/config/platform_mappings
|
|
|
|
# RBE cross-compile configs for Darwin x86
|
|
build:rbe_cross_compile_macos_x86 --config=cross_compile_macos_x86 --remote_download_minimal
|
|
build:rbe_cross_compile_macos_x86 --bes_backend="" --bes_results_url="" --bes_timeout="0s"
|
|
build:rbe_cross_compile_macos_x86 --experimental_remote_build_event_upload="minimal"
|
|
build:rbe_cross_compile_macos_x86 --config=rbe_cross_compile_base
|
|
build:rbe_cross_compile_macos_x86 --bes_upload_mode=nowait_for_upload_complete
|
|
test:rbe_cross_compile_macos_x86 --config=rbe_cross_compile_base
|
|
# Increase the test timeout as tests often take longer on mac.
|
|
test:rbe_cross_compile_macos_x86 --test_timeout=300,450,1200,3600
|
|
# Limit jobs to 100 to avoid running into "out of memory" issues (b/316266643)
|
|
build:rbe_cross_compile_macos_x86 --jobs=100
|
|
test:rbe_cross_compile_macos_x86 --jobs=100
|
|
# END MACOS CROSS-COMPILE CONFIGS
|
|
# END CROSS-COMPILE CONFIGS
|
|
|
|
# Try to load the XLA warnings config if available
|
|
try-import %workspace%/warnings.bazelrc
|