Merge branch 'main' into bump-cmake-version

Author: GregoryComer

.github/workflows/trunk.yml

@@ -302,6 +302,37 @@ jobs:
           exit 1
         fi
 
+  test-arm-ootb-linux:
+    name: test-arm-ootb-linux
+    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}"
+
+        # Follow the steps required before running the notebooks
+        # Try to mirror these as closely as possible
+        source .ci/scripts/utils.sh
+        install_executorch "--use-pt-pinned-commit"
+
+        .ci/scripts/setup-arm-baremetal-tools.sh
+        source examples/arm/ethos-u-scratch/setup_path.sh
+
+        # Install requirements for converting notebooks
+        pip install notebook
+
+        # Run OOTB tests
+        backends/arm/test/test_arm_ootb.sh
+
   test-coreml-delegate:
     name: test-coreml-delegate
     uses: pytorch/test-infra/.github/workflows/macos_job.yml@main

CMakeLists.txt

@@ -50,8 +50,6 @@
 cmake_minimum_required(VERSION 3.29)
 project(executorch)
 
-# MARK: - Start EXECUTORCH_H12025_BUILD_MIGRATION
-
 include(${PROJECT_SOURCE_DIR}/tools/cmake/common/preset.cmake)
 include(${PROJECT_SOURCE_DIR}/tools/cmake/Utils.cmake)
 include(CMakeDependentOption)
@@ -84,6 +82,7 @@ announce_configured_options(BUCK2)
 
 announce_configured_options(CMAKE_CXX_COMPILER_ID)
 announce_configured_options(CMAKE_TOOLCHAIN_FILE)
+announce_configured_options(BUILD_TESTING)
 
 load_build_preset()
 include(${PROJECT_SOURCE_DIR}/tools/cmake/preset/default.cmake)
@@ -99,11 +98,6 @@ else()
 endif()
 announce_configured_options(CCACHE_PROGRAM)
 
-# Print all the configs that were called with announce_configured_options.
-print_configured_options()
-
-# MARK: - End EXECUTORCH_H12025_BUILD_MIGRATION
-
 set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
 
 # Setup RPATH. See
@@ -752,3 +746,6 @@ if(EXECUTORCH_BUILD_ANDROID_JNI)
 endif()
 
 include(Test.cmake)
+
+# Print all the configs that were called with announce_configured_options.
+print_configured_options()

CMakePresets.json

@@ -8,14 +8,15 @@
     },
     {
       "name": "macos",
-      "displayName": "Build everything buildable on macOS",
+      "displayName": "Build ExecuTorch for macOS",
       "inherits": ["common"],
       "generator": "Xcode",
       "cacheVariables": {
         "CMAKE_TOOLCHAIN_FILE": "${sourceDir}/third-party/ios-cmake/ios.toolchain.cmake",
         "EXECUTORCH_BUILD_PRESET_FILE": "${sourceDir}/tools/cmake/preset/macos.cmake",
         "PLATFORM": "MAC_ARM64",
-        "DEPLOYMENT_TARGET": "12.0"
+        "DEPLOYMENT_TARGET": "12.0",
+        "CMAKE_MACOSX_BUNDLE": "OFF"
       },
       "condition": {
         "lhs": "${hostSystemName}",
@@ -25,7 +26,7 @@
     },
     {
       "name": "ios",
-      "displayName": "Build everything buildable on iOS",
+      "displayName": "Build ExecuTorch for iOS",
       "inherits": ["common"],
       "generator": "Xcode",
       "cacheVariables": {
@@ -42,7 +43,7 @@
     },
     {
       "name": "ios-simulator",
-      "displayName": "Build everything buildable on iOS simulator",
+      "displayName": "Build ExecuTorch for iOS Simulator",
       "inherits": ["common"],
       "generator": "Xcode",
       "cacheVariables": {
@@ -59,7 +60,7 @@
     },
     {
       "name": "linux",
-      "displayName": "Build everything buildable on Linux",
+      "displayName": "Build ExecuTorch for Linux",
       "inherits": ["common"],
       "cacheVariables": {
         "CMAKE_SYSTEM_NAME": "Linux",
@@ -88,29 +89,21 @@
     {
         "name": "llm",
         "displayName": "Build LLM libraries",
-        "inherits": [
-            "common"
-        ],
+        "inherits": ["common"],
         "cacheVariables": {
             "EXECUTORCH_BUILD_PRESET_FILE": "${sourceDir}/tools/cmake/preset/llm.cmake",
             "CMAKE_OSX_DEPLOYMENT_TARGET": "12.0"
         },
         "condition": {
             "type": "inList",
             "string": "${hostSystemName}",
-            "list": [
-                "Darwin",
-                "Linux",
-                "Windows"
-            ]
+            "list": ["Darwin", "Linux", "Windows"]
         }
     },
     {
         "name": "zephyr",
-        "displayName": "Build everything buildable on Zephyr RTOS",
-        "inherits": [
-            "common"
-        ],
+        "displayName": "Build ExecuTorch for Zephyr RTOS",
+        "inherits": ["common"],
         "cacheVariables": {
             "EXECUTORCH_BUILD_PRESET_FILE": "${sourceDir}/tools/cmake/preset/zephyr.cmake",
             "CMAKE_TOOLCHAIN_FILE": "${sourceDir}/examples/zephyr/x86_64-linux-arm-zephyr-eabi-gcc.cmake"

backends/arm/_passes/decompose_asin_pass.py

@@ -85,12 +85,11 @@ def _build_polynomial(
         return result
 
     def call_operator(self, op, args, kwargs, meta):
+        if op not in edge_asin_op:
+            return super().call_operator(op, args, kwargs, meta)
         logging.info(
             f"Approximating asin. This may introduce small numerical errors. For details, see {__file__}."
         )
-        if op not in edge_asin_op:
-            return super().call_operator(op, args, kwargs, meta)
-
         x = args[0]
         half = 0.5
         one = 1.0

backends/arm/test/ops/test_add.py

@@ -205,6 +205,5 @@ def test_add_tensor_vgf_int(test_data: input_t1):
         aten_op,
         exir_op,
         tosa_version="TOSA-1.0+INT",
-        symmetric_io_quantization=True,
     )
     pipeline.run()

backends/arm/test/test_arm_ootb.sh

@@ -0,0 +1,18 @@
+#!/usr/bin/env bash
+
+# Copyright 2025 Arm Limited and/or its affiliates.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+set -e
+
+run_ootb_tests_ethos_u() {
+    echo "$FUNCNAME: Running out-of-the-box tests for Arm Ethos-U"
+    jupyter nbconvert \
+        --to notebook \
+        --execute examples/arm/ethos_u_minimal_example.ipynb
+    echo "${FUNCNAME}: PASS"
+}
+
+run_ootb_tests_ethos_u

backends/arm/test/tester/arm_tester.py

@@ -43,6 +43,7 @@
     EthosUQuantizer,
     get_symmetric_quantization_config,
     TOSAQuantizer,
+    VgfQuantizer,
 )
 from executorch.backends.arm.test.runner_utils import (
     dbg_tosa_fb_to_json,
@@ -332,6 +333,8 @@ def quantize(
                 quantizer = TOSAQuantizer(tosa_spec)
             elif is_ethosu(self.compile_spec):
                 quantizer = EthosUQuantizer(self.compile_spec)
+            elif is_vgf(self.compile_spec):
+                quantizer = VgfQuantizer(self.compile_spec)
             quantize_stage = tester.Quantize(
                 quantizer,
                 get_symmetric_quantization_config(),

backends/arm/test/tester/test_pipeline.py

@@ -861,18 +861,15 @@ def __init__(
         rtol: float = 1e-03,
         qtol: int = 1,
         dynamic_shapes: Optional[Tuple[Any]] = None,
+        transform_passes: Optional[
+            Union[Sequence[PassType], Dict[str, Sequence[PassType]]]
+        ] = None,
     ):
 
         if (
             symmetric_io_quantization or per_channel_quantization
         ) and tosa_version == "TOSA-1.0+FP":
             raise ValueError("Dont configure quantization with FP TOSA profile.")
-        if (
-            symmetric_io_quantization is False
-            and per_channel_quantization is False
-            and tosa_version == "TOSA-1.0+INT"
-        ):
-            raise ValueError("Missing quantization options for INT TOSA profile.")
 
         tosa_profile = TosaSpecification.create_from_string(tosa_version)
         compile_spec = common.get_vgf_compile_spec(
@@ -887,6 +884,7 @@ def __init__(
             exir_op,
             use_to_edge_transform_and_lower,
             dynamic_shapes,
+            transform_passes=transform_passes,
         )
 
         if symmetric_io_quantization or per_channel_quantization:
@@ -900,7 +898,7 @@ def __init__(
         else:
             quant_stage = None
 
-        if quant_stage:
+        if "INT" in tosa_version:
             self.add_stage(self.tester.quantize, quant_stage, pos=0)
 
             self.add_stage_after(

backends/cadence/aot/TARGETS

@@ -199,6 +199,19 @@ python_library(
     ],
 )
 
+python_unittest(
+    name = "test_program_builder",
+    srcs = [
+        "tests/test_program_builder.py",
+    ],
+    typing = True,
+    deps = [
+        ":program_builder",
+        "//caffe2:torch",
+        "//later:lib",
+    ],
+)
+
 python_library(
     name = "fuse_ops",
     srcs = [
@@ -539,8 +552,10 @@ python_unittest(
     ],
     typing = True,
     deps = [
+        ":program_builder",
         "//executorch/backends/cadence/aot:graph_builder",
         "//executorch/backends/cadence/aot:ops_registrations",
+        "//executorch/runtime:runtime",
         "//later:lib",
     ],
 )

backends/cadence/aot/compiler.py

@@ -34,7 +34,7 @@
 )
 from executorch.exir.passes import ToOutVarPass
 from executorch.exir.passes.sym_shape_eval_pass import HintBasedSymShapeEvalPass
-from executorch.exir.program._program import to_edge_with_preserved_ops
+from executorch.exir.program._program import to_edge
 from torch._inductor.decomposition import remove_decompositions
 
 from torch.export.exported_program import ExportedProgram
@@ -219,9 +219,9 @@ def quantize_pt2(
     torch.ops.aten.angle.default,
     torch.ops.aten.rms_norm.default,
 ]
-TO_EDGE_PRESERVE_OPS: tuple[torch._ops.OpOverload, ...] = (
+TO_EDGE_PRESERVE_OPS: list[torch._ops.OpOverload, ...] = [
     torch.ops.aten.rms_norm.default,
-)
+]
 
 
 def _lower_ep_to_edge(
@@ -233,18 +233,18 @@ def _lower_ep_to_edge(
     """
     Lower an ExportedProgram to an EdgeProgramManager (in edge IR).
     """
-    # Call to_edge_with_preserved_ops to convert the graph to edge IR.
+    # Call to_edge to convert the graph to edge IR.
     # Note: dim_order is skipped (https://github.com/pytorch/executorch/issues/3704)
-    edge_prog_manager = to_edge_with_preserved_ops(
+    edge_prog_manager = to_edge(
         expo_program,
         compile_config=EdgeCompileConfig(
             _skip_dim_order=True,
             # Allow specific non-core aten ops in the IR.
             _core_aten_ops_exception_list=TO_EDGE_OP_EXCEPTION_LIST
             + (core_aten_exceptions or []),
+            preserve_ops=TO_EDGE_PRESERVE_OPS,
         ),
         constant_methods=constant_methods,
-        preserve_ops=TO_EDGE_PRESERVE_OPS,
     )
 
     if dump_graphs: