flip use_cxx11_abi naming

Author: zewenli98

.github/workflows/docgen.yml

@@ -42,7 +42,7 @@ jobs:
               run: echo "sha=$(git rev-parse --short HEAD)" >> $GITHUB_OUTPUT
             - name: Build Python Package
               env:
-                  USE_CXX11_ABI: 1
+                  USE_PRE_CXX11_ABI: 0
               run: |
                   python3 -m pip install --pre . --extra-index-url https://download.pytorch.org/whl/nightly/cu126
             - name: Generate New Docs

CONTRIBUTING.md

@@ -75,7 +75,7 @@ Environment variables supported by nox
 ```
 PYT_PATH          - To use different PYTHONPATH than system installed Python packages
 TOP_DIR           - To set the root directory of the noxfile
-USE_CXX11         - To use cxx11_abi (Defaults to 0)
+USE_PRE_CXX11     - To use pre_cxx11_abi (Defaults to 0)
 USE_HOST_DEPS     - To use host dependencies for tests (Defaults to 0)
 ```
 

docker/Dockerfile

@@ -12,8 +12,8 @@ RUN test -n "$TENSORRT_VERSION" || (echo "No tensorrt version specified, please
 ARG PYTHON_VERSION=3.10
 ENV PYTHON_VERSION=${PYTHON_VERSION}
 
-ARG USE_CXX11_ABI
-ENV USE_CXX11=${USE_CXX11_ABI}
+ARG USE_PRE_CXX11_ABI
+ENV USE_PRE_CXX11=${USE_PRE_CXX11_ABI}
 ENV DEBIAN_FRONTEND=noninteractive
 
 # Install basic dependencies

docker/README.md

@@ -1,30 +1,30 @@
 # Building a Torch-TensorRT container
 
-* Use `Dockerfile` to build a container which provides the exact development environment that our master branch is usually tested against.
+* Use `Dockerfile` to build a container which provides the exact development environment that our main branch is usually tested against.
 
 * The `Dockerfile` currently uses <a href="https://github.com/bazelbuild/bazelisk">Bazelisk</a> to select the Bazel version, and uses the exact library versions of Torch and CUDA listed in <a href="https://github.com/pytorch/TensorRT#dependencies">dependencies</a>.
   * The desired versions of TensorRT must be specified as build-args, with major and minor versions as in: `--build-arg TENSORRT_VERSION=a.b`
-  * [**Optional**] The desired base image be changed by explicitly setting a base image, as in `--build-arg BASE_IMG=nvidia/cuda:11.8.0-devel-ubuntu22.04`, though this is optional
+  * [**Optional**] The desired base image be changed by explicitly setting a base image, as in `--build-arg BASE_IMG=nvidia/cuda:11.8.0-devel-ubuntu22.04`, though this is optional.
   * [**Optional**] Additionally, the desired Python version can be changed by explicitly setting a version, as in `--build-arg PYTHON_VERSION=3.10`, though this is optional as well.
 
-* This `Dockerfile` installs `pre-cxx11-abi` versions of Pytorch and builds Torch-TRT using `pre-cxx11-abi` libtorch as well.
+* This `Dockerfile` installs `cxx11-abi` versions of Pytorch and builds Torch-TRT using `cxx11-abi` libtorch as well. As of torch 2.7, torch requires `cxx11-abi` for all CUDA 11.8, 12.4, and 12.6.
 
-Note: By default the container uses the `pre-cxx11-abi` version of Torch + Torch-TRT. If you are using a workflow that requires a build of PyTorch on the CXX11 ABI (e.g. using the PyTorch NGC containers as a base image), add the Docker build argument: `--build-arg USE_CXX11_ABI=1`
+Note: By default the container uses the `cxx11-abi` version of Torch + Torch-TRT. If you are using a workflow that requires a build of PyTorch on the PRE CXX11 ABI, please add the Docker build argument: `--build-arg USE_PRE_CXX11_ABI=1`
 
 ### Dependencies
 
 * Install nvidia-docker by following https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker
 
 ### Instructions
 
-- The example below uses TensorRT 10.6.0.26
+- The example below uses TensorRT 10.7.0.23
 - See <a href="https://github.com/pytorch/TensorRT#dependencies">dependencies</a> for a list of current default dependencies.
 
 > From root of Torch-TensorRT repo
 
 Build:
 ```
-DOCKER_BUILDKIT=1 docker build --build-arg TENSORRT_VERSION=10.6.0 -f docker/Dockerfile -t torch_tensorrt:latest .
+DOCKER_BUILDKIT=1 docker build --build-arg TENSORRT_VERSION=10.7.0 -f docker/Dockerfile -t torch_tensorrt:latest .
 ```
 
 Run:

docker/dist-build.sh

@@ -4,10 +4,10 @@ set -x
 
 TOP_DIR=$(cd $(dirname $0); pwd)/..
 
-if [[ -z "${USE_CXX11}" ]]; then
+if [[ -z "${USE_PRE_CXX11}" ]]; then
     BUILD_CMD="python -m pip wheel .  --extra-index-url https://download.pytorch.org/whl/nightly/cu124 -w dist"
 else
-    BUILD_CMD="python -m pip wheel . --config-setting="--build-option=--use-cxx11-abi" --extra-index-url https://download.pytorch.org/whl/nightly/cu124 -w dist"
+    BUILD_CMD="python -m pip wheel . --config-setting="--build-option=--use-pre-cxx11-abi" --extra-index-url https://download.pytorch.org/whl/nightly/cu124 -w dist"
 fi
 
 # TensorRT restricts our pip version

docsrc/getting_started/installation.rst

@@ -235,13 +235,13 @@ recommended commands:
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
 | libtorch-shared-with-deps-*.zip from PyTorch.org            | python -m pip install .                                  | bazel build //:libtorchtrt -c opt \-\-config pre_cxx11_abi         |
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
-| libtorch-cxx11-abi-shared-with-deps-*.zip from PyTorch.org  | python setup.py bdist_wheel --use-cxx11-abi              | bazel build //:libtorchtrt -c opt                                  |
+| libtorch-cxx11-abi-shared-with-deps-*.zip from PyTorch.org  | python setup.py bdist_wheel                              | bazel build //:libtorchtrt -c opt                                  |
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
-| PyTorch preinstalled in an NGC container                    | python setup.py bdist_wheel --use-cxx11-abi              | bazel build //:libtorchtrt -c opt                                  |
+| PyTorch preinstalled in an NGC container                    | python setup.py bdist_wheel                              | bazel build //:libtorchtrt -c opt                                  |
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
-| PyTorch from the NVIDIA Forums for Jetson                   | python setup.py bdist_wheel --use-cxx11-abi              | bazel build //:libtorchtrt -c opt                                  |
+| PyTorch from the NVIDIA Forums for Jetson                   | python setup.py bdist_wheel                              | bazel build //:libtorchtrt -c opt                                  |
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
-| PyTorch built from Source                                   | python setup.py bdist_wheel --use-cxx11-abi              | bazel build //:libtorchtrt -c opt                                  |
+| PyTorch built from Source                                   | python setup.py bdist_wheel                              | bazel build //:libtorchtrt -c opt                                  |
 +-------------------------------------------------------------+----------------------------------------------------------+--------------------------------------------------------------------+
 
     NOTE: For all of the above cases you must correctly declare the source of PyTorch you intend to use in your WORKSPACE file for both Python and C++ builds. See below for more information
@@ -383,8 +383,8 @@ Compile the Python API using the following command from the ``//py`` directory:
 
 .. code-block:: shell
 
-    python3 setup.py install --use-cxx11-abi
+    python3 setup.py install
 
-If you have a build of PyTorch that uses Pre-CXX11 ABI drop the ``--use-cxx11-abi`` flag
+If you have a build of PyTorch that uses Pre-CXX11 ABI drop the ``--use-pre-cxx11-abi`` flag
 
 If you are building for Jetpack 4.5 add the ``--jetpack-version 5.0`` flag

docsrc/getting_started/jetpack.rst

@@ -115,5 +115,5 @@ Please make sure to build torch_tensorrt wheel file from source release/2.5 bran
     # replace the MODULE.bazel with the jetpack one
     cat toolchains/jp_workspaces/MODULE.bazel.tmpl | envsubst > MODULE.bazel
     # build and install torch_tensorrt wheel file
-    python setup.py --use-cxx11-abi install --user
+    python setup.py install --user
 

noxfile.py

@@ -21,10 +21,10 @@
 )
 print(f"Test root directory {TOP_DIR}")
 
-# Set the USE_CXX11=1 to use cxx11_abi
-USE_CXX11 = 0 if not "USE_CXX11" in os.environ else os.environ["USE_CXX11"]
-if USE_CXX11:
-    print("Using cxx11 abi")
+# Set the USE_PRE_CXX11=1 to use pre_cxx11_abi
+USE_PRE_CXX11 = 0 if not "USE_PRE_CXX11" in os.environ else os.environ["USE_PRE_CXX11"]
+if USE_PRE_CXX11:
+    print("Using pre cxx11 abi")
 
 # Set the USE_HOST_DEPS=1 to use host dependencies for tests
 USE_HOST_DEPS = 0 if not "USE_HOST_DEPS" in os.environ else os.environ["USE_HOST_DEPS"]
@@ -61,8 +61,8 @@ def download_models(session):
 def install_torch_trt(session):
     print("Installing latest torch-tensorrt build")
     session.chdir(os.path.join(TOP_DIR, "py"))
-    if USE_CXX11:
-        session.run("python", "setup.py", "develop", "--use-cxx11-abi")
+    if USE_PRE_CXX11:
+        session.run("python", "setup.py", "develop", "--use-pre-cxx11-abi")
     else:
         session.run("python", "setup.py", "develop")
 

packaging/env_vars.txt

@@ -1,3 +1,2 @@
 export CI_BUILD="1"
-export RELEASE="1"
-export USE_CXX11_ABI="1"
+export RELEASE="1"

py/README.md

@@ -29,16 +29,16 @@ torch.jit.save(trt_ts_module, "trt_torchscript_module.ts") # save the TRT embedd
 
 ## Installation
 
-| ABI / Platform                          | Installation command                                         |
-| --------------------------------------- | ------------------------------------------------------------ |
-| Pre CXX11 ABI (Linux x86_64)            | python3 setup.py install                                     |
-| CXX ABI  (Linux x86_64)                 | python3 setup.py install --use-cxx11-abi                     |
-| Pre CXX11 ABI (Jetson platform aarch64) | python3 setup.py install --jetpack-version 4.6               |
-| CXX11 ABI (Jetson platform aarch64)     | python3 setup.py install --jetpack-version 4.6 --use-cxx11-abi |
+| ABI / Platform                          | Installation command                                              |
+| --------------------------------------- | ----------------------------------------------------------------- |
+| Pre CXX11 ABI (Linux x86_64)            | python3 setup.py install --use-pre-cxx11-abi                      |
+| CXX ABI  (Linux x86_64)                 | python3 setup.py install                                          |
+| Pre CXX11 ABI (Jetson platform aarch64) | python3 setup.py install --jetpack-version 4.6 --use-pre-cxx11-abi|
+| CXX11 ABI (Jetson platform aarch64)     | python3 setup.py install --jetpack-version 4.6                    |
 
 For Linux x86_64 platform, Pytorch libraries default to pre cxx11 abi. So, please use `python3 setup.py install`.
 
-On Jetson platforms, NVIDIA hosts <a href="https://forums.developer.nvidia.com/t/pytorch-for-jetson-version-1-10-now-available/72048">pre-built Pytorch wheel files</a>. These wheel files are built with CXX11 ABI. So on jetson platforms, please use `python3 setup.py install --jetpack-version 4.6 --use-cxx11-abi`
+On Jetson platforms, NVIDIA hosts <a href="https://forums.developer.nvidia.com/t/pytorch-for-jetson-version-1-10-now-available/72048">pre-built Pytorch wheel files</a>. These wheel files are built with CXX11 ABI. So on jetson platforms, please use `python3 setup.py install --jetpack-version 4.6`
 
 ## Under the Hood
 

py/ci/Dockerfile.ci

@@ -10,7 +10,6 @@ RUN wget -q https://developer.nvidia.com/downloads/compute/machine-learning/tens
 
 ENV TENSORRT_DIR=/TensorRT-10.7.0.23
 ENV TENSORRT_VERSION=10.7.0
-ENV USE_CXX11_ABI=1
 
 RUN wget https://github.com/bazelbuild/bazelisk/releases/download/v1.17.0/bazelisk-linux-amd64 \
     && mv bazelisk-linux-amd64 /usr/bin/bazel \

setup.py

@@ -78,7 +78,7 @@ def load_dep_info():
 
 dir_path = os.path.join(str(get_root_dir()), "py")
 
-CXX11_ABI = IS_WINDOWS
+PRE_CXX11_ABI = False
 JETPACK_VERSION = None
 PY_ONLY = False
 NO_TS = False
@@ -136,13 +136,13 @@ def load_dep_info():
     if ci_env_var == "1":
         CI_BUILD = True
 
-if "--use-cxx11-abi" in sys.argv:
-    sys.argv.remove("--use-cxx11-abi")
-    CXX11_ABI = True
+if "--use-pre-cxx11-abi" in sys.argv:
+    sys.argv.remove("--use-pre-cxx11-abi")
+    PRE_CXX11_ABI = True
 
-if (cxx11_abi_env_var := os.environ.get("USE_CXX11_ABI")) is not None:
-    if cxx11_abi_env_var == "1":
-        CXX11_ABI = True
+if (pre_cxx11_abi_env_var := os.environ.get("USE_PRE_CXX11_ABI")) is not None:
+    if pre_cxx11_abi_env_var == "1":
+        PRE_CXX11_ABI = True
 
 if platform.uname().processor == "aarch64":
     if "--jetpack-version" in sys.argv:
@@ -165,9 +165,9 @@ def load_dep_info():
         )
         JETPACK_VERSION = "6.1"
 
-    if not CXX11_ABI:
+    if PRE_CXX11_ABI:
         warnings.warn(
-            "Jetson platform detected but did not see --use-cxx11-abi option, if using a pytorch distribution provided by NVIDIA include this flag"
+            "Jetson platform detected. Please remove --use-pre-cxx11-abi flag if you are using it."
         )
 
 
@@ -182,7 +182,7 @@ def load_dep_info():
 
 
 def build_libtorchtrt_pre_cxx11_abi(
-    develop=True, use_dist_dir=True, cxx11_abi=False, rt_only=False
+    develop=True, use_dist_dir=True, pre_cxx11_abi=True, rt_only=False
 ):
     cmd = [BAZEL_EXE, "build"]
     if rt_only:
@@ -196,7 +196,7 @@ def build_libtorchtrt_pre_cxx11_abi(
         cmd.append("--compilation_mode=opt")
     if use_dist_dir:
         cmd.append("--distdir=third_party/dist_dir/x86_64-linux-gnu")
-    if not cxx11_abi:
+    if pre_cxx11_abi:
         cmd.append("--config=python")
     else:
         print("using CXX11 ABI build")
@@ -292,9 +292,9 @@ def finalize_options(self):
     def run(self):
 
         if not PY_ONLY:
-            global CXX11_ABI
+            global PRE_CXX11_ABI
             build_libtorchtrt_pre_cxx11_abi(
-                develop=True, cxx11_abi=CXX11_ABI, rt_only=NO_TS
+                develop=True, pre_cxx11_abi=PRE_CXX11_ABI, rt_only=NO_TS
             )
             copy_libtorchtrt(rt_only=NO_TS)
 
@@ -316,9 +316,9 @@ def finalize_options(self):
     def run(self):
 
         if not PY_ONLY:
-            global CXX11_ABI
+            global PRE_CXX11_ABI
             build_libtorchtrt_pre_cxx11_abi(
-                develop=False, cxx11_abi=CXX11_ABI, rt_only=NO_TS
+                develop=False, pre_cxx11_abi=PRE_CXX11_ABI, rt_only=NO_TS
             )
             copy_libtorchtrt(rt_only=NO_TS)
 
@@ -339,9 +339,9 @@ def finalize_options(self):
 
     def run(self):
         if not PY_ONLY:
-            global CXX11_ABI
+            global PRE_CXX11_ABI
             build_libtorchtrt_pre_cxx11_abi(
-                develop=False, cxx11_abi=CXX11_ABI, rt_only=NO_TS
+                develop=False, pre_cxx11_abi=PRE_CXX11_ABI, rt_only=NO_TS
             )
             copy_libtorchtrt(rt_only=NO_TS)
 
@@ -365,9 +365,9 @@ def run(self):
             gen_version_file()
             editable_wheel.run(self)
         else:
-            global CXX11_ABI
+            global PRE_CXX11_ABI
             build_libtorchtrt_pre_cxx11_abi(
-                develop=True, cxx11_abi=CXX11_ABI, rt_only=NO_TS
+                develop=True, pre_cxx11_abi=PRE_CXX11_ABI, rt_only=NO_TS
             )
             gen_version_file()
             copy_libtorchtrt(rt_only=NO_TS)
@@ -561,9 +561,9 @@ def run(self):
                     "-Wno-deprecated-declarations",
                 ]
                 + (
-                    ["-D_GLIBCXX_USE_CXX11_ABI=1"]
-                    if CXX11_ABI
-                    else ["-D_GLIBCXX_USE_CXX11_ABI=0"]
+                    ["-D_GLIBCXX_USE_CXX11_ABI=0"]
+                    if PRE_CXX11_ABI
+                    else ["-D_GLIBCXX_USE_CXX11_ABI=1"]
                 )
             ),
             extra_link_args=(
@@ -584,9 +584,9 @@ def run(self):
                     "-export-dynamic",
                 ]
                 + (
-                    ["-D_GLIBCXX_USE_CXX11_ABI=1"]
-                    if CXX11_ABI
-                    else ["-D_GLIBCXX_USE_CXX11_ABI=0"]
+                    ["-D_GLIBCXX_USE_CXX11_ABI=0"]
+                    if PRE_CXX11_ABI
+                    else ["-D_GLIBCXX_USE_CXX11_ABI=1"]
                 )
             ),
             undef_macros=["NDEBUG"],