Update base for Update on "[wip] Introduce MergedDataMap"

Differential Revision: [D76529405](https://our.internmc.facebook.com/intern/diff/D76529405/)

[ghstack-poisoned]

Author: lucylq

.ci/docker/ci_commit_pins/pytorch.txt

@@ -1 +1 @@
-5616fa4a68718ead203314a3467f7dd9547153ae
+9b498d3bb28b8e3411ce464dd2755c5b96d92c8f

.ci/docker/common/install_conda.sh

@@ -13,7 +13,7 @@ source "$(dirname "${BASH_SOURCE[0]}")/utils.sh"
 install_miniconda() {
   BASE_URL="https://repo.anaconda.com/miniconda"
   CONDA_FILE="Miniconda3-py${PYTHON_VERSION//./}_${MINICONDA_VERSION}-Linux-x86_64.sh"
-  if [[ $(uname -m) == "aarch64" ]]; then 
+  if [[ $(uname -m) == "aarch64" ]]; then
     CONDA_FILE="Miniconda3-py${PYTHON_VERSION//./}_${MINICONDA_VERSION}-Linux-aarch64.sh"
   fi
 
@@ -71,4 +71,8 @@ fix_conda_ubuntu_libstdcxx() {
 install_miniconda
 install_python
 install_pip_dependencies
-fix_conda_ubuntu_libstdcxx
+# Hack breaks the job on aarch64 but is still necessary everywhere
+# else.
+if [ "$(uname -m)" != "aarch64" ]; then
+    fix_conda_ubuntu_libstdcxx
+fi

.ci/docker/conda-env-ci.txt

@@ -1,4 +1,4 @@
-cmake=3.26.4
+cmake=3.31.2
 ninja=1.10.2
 libuv
 llvm-openmp

.ci/scripts/check_c10_sync.sh

@@ -12,3 +12,4 @@ pushd pytorch
 git checkout "$pytorch_pin"
 popd
 "$(dirname "${BASH_SOURCE[0]}")"/compare_dirs.sh runtime/core/portable_type/c10/c10 pytorch/c10
+"$(dirname "${BASH_SOURCE[0]}")"/compare_dirs.sh runtime/core/portable_type/c10/torch/standalone pytorch/torch/standalone

.github/workflows/apple.yml

@@ -39,7 +39,7 @@ jobs:
         id: set_version
         shell: bash
         run: |
-          VERSION="0.7.0.$(TZ='PST8PDT' date +%Y%m%d)"
+          VERSION="0.8.0.$(TZ='PST8PDT' date +%Y%m%d)"
           echo "version=$VERSION" >> "$GITHUB_OUTPUT"
 
   build-demo-ios:

.github/workflows/trunk.yml

@@ -271,6 +271,36 @@ jobs:
           exit 1
         fi
 
+  nxp-build-test:
+    name: nxp-build-test
+    uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
+    permissions:
+      id-token: write
+      contents: read
+    with:
+      runner: linux.2xlarge
+      docker-image: executorch-ubuntu-22.04-arm-sdk
+      submodules: 'recursive'
+      ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
+      timeout: 90
+      script: |
+        # The generic Linux job chooses to use base env, not the one setup by the image
+        CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
+        conda activate "${CONDA_ENV}"
+
+        # Build
+        cmake -DEXECUTORCH_BUILD_NXP_NEUTRON=ON -Bcmake-out .
+        cmake --build cmake-out --target executorch_delegate_neutron --config Release
+
+        # Build check for the neutron backend library
+        lib_neutron="cmake-out/backends/nxp/libexecutorch_delegate_neutron.a"
+        if [ -f $lib_neutron ]; then
+            echo "Neutron backend library built."
+        else
+            echo "Neutron backend library not found!"
+            exit 1
+        fi
+
   test-coreml-delegate:
     name: test-coreml-delegate
     uses: pytorch/test-infra/.github/workflows/macos_job.yml@main

.lintrunner.toml

@@ -76,6 +76,7 @@ exclude_patterns = [
     'examples/demo-apps/apple_ios/**',
     'examples/demo-apps/react-native/rnllama/ios/**',
     'extension/apple/**',
+    'extension/llm/apple/**',
     # File contains @generated
     'extension/llm/custom_ops/spinquant/fast_hadamard_transform_special.h',
     'extension/llm/custom_ops/spinquant/test/fast_hadamard_transform_special_unstrided_cpu.h',

CMakeLists.txt

@@ -496,6 +496,10 @@ if(EXECUTORCH_BUILD_CADENCE)
   add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/backends/cadence)
 endif()
 
+if(EXECUTORCH_BUILD_NXP_NEUTRON)
+  add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/backends/nxp)
+endif()
+
 if(EXECUTORCH_BUILD_COREML)
   add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/backends/apple/coreml)
 endif()
@@ -705,12 +709,8 @@ if(EXECUTORCH_BUILD_EXECUTOR_RUNNER)
   endif()
 
   add_executable(executor_runner ${_executor_runner__srcs})
-  if(CMAKE_BUILD_TYPE STREQUAL "Release")
-    if(APPLE)
-      target_link_options(executor_runner PRIVATE "LINKER:-dead_strip")
-    else()
-      target_link_options(executor_runner PRIVATE "LINKER:--gc-sections")
-    endif()
+  if(NOT CMAKE_BUILD_TYPE STREQUAL "Debug")
+    target_link_options_gc_sections(executor_runner)
   endif()
   target_link_libraries(executor_runner ${_executor_runner_libs})
   target_compile_options(executor_runner PUBLIC ${_common_compile_options})

backends/arm/_passes/__init__.py

@@ -22,6 +22,7 @@
 from .convert_split_to_slice import ConvertSplitToSlicePass  # noqa
 from .convert_squeezes_to_view import ConvertSqueezesToViewPass  # noqa
 from .convert_to_clamp import ConvertToClampPass  # noqa
+from .decompose_acosh_pass import DecomposeAcoshPass  # noqa
 from .decompose_atan_pass import DecomposeAtanPass  # noqa
 from .decompose_avg_pool2d import DecomposeAvgPool2d  # noqa
 from .decompose_batch_norm_no_stats import DecomposeBatchNormNoStatsPass  # noqa

backends/arm/_passes/arm_pass_manager.py

@@ -25,6 +25,7 @@
     ConvertSplitToSlicePass,
     ConvertSqueezesToViewPass,
     ConvertToClampPass,
+    DecomposeAcoshPass,
     DecomposeAtanPass,
     DecomposeAvgPool2d,
     DecomposeBatchNormNoStatsPass,
@@ -151,6 +152,7 @@ def _tosa_080_BI_pipeline(self, exported_program: ExportedProgram) -> GraphModul
 
     def _tosa_080_MI_pipeline(self, exported_program: ExportedProgram) -> GraphModule:
         self.add_pass(DecomposeRoundPass())
+        self.add_pass(DecomposeAcoshPass())
         self.add_pass(DecomposeSqrtPass())
         self.add_pass(DecomposeAtanPass())
         self.add_pass(ConvertIntPowToMuls())