Merge branch 'pytorch:main' into size_adjust_pass_revisit

Author: Sebastian-Larsson

.ci/scripts/build-qnn-sdk.sh

@@ -1,5 +1,6 @@
 #!/bin/bash
 # Copyright (c) Meta Platforms, Inc. and affiliates.
+# Copyright 2025 Arm Limited and/or its affiliates.
 # All rights reserved.
 #
 # This source code is licensed under the BSD-style license found in the
@@ -11,10 +12,16 @@ set -o xtrace
 build_qnn_backend() {
   echo "Start building qnn backend."
   export ANDROID_NDK_ROOT=/opt/ndk
-  export QNN_SDK_ROOT=/tmp/qnn/2.25.0.240728
+  export QNN_SDK_ROOT=/tmp/qnn/2.28.0.241029
   export EXECUTORCH_ROOT="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")/../.." && pwd)"
 
-  bash backends/qualcomm/scripts/build.sh --skip_aarch64 --job_number 2 --release
+  # Workaround to avoid issues around missing flatccrt library (depending on the
+  # number of jobs used), see issue #7300:
+  # Build twice (second time with `--no_clean`) to make sure libflatccrt.a is
+  # available.
+  # TODO: Remove this workaround once the underlying issue is fixed.
+  bash backends/qualcomm/scripts/build.sh --skip_aarch64 --job_number 2 --release || \
+  bash backends/qualcomm/scripts/build.sh --skip_aarch64 --job_number 2 --release --no_clean
 }
 
 set_up_aot() {

.ci/scripts/setup-qnn-deps.sh

@@ -16,9 +16,9 @@ install_qnn() {
   QNN_INSTALLATION_DIR=/tmp/qnn
   mkdir -p "${QNN_INSTALLATION_DIR}"
 
-  curl -Lo /tmp/v2.25.0.24.07.28.zip "https://softwarecenter.qualcomm.com/api/download/software/qualcomm_neural_processing_sdk/v2.25.0.240728.zip"
+  curl -Lo /tmp/v2.28.0.24.10.29.zip "https://softwarecenter.qualcomm.com/api/download/software/qualcomm_neural_processing_sdk/v2.28.0.241029.zip"
   echo "Finishing downloading qnn sdk."
-  unzip -qo /tmp/v2.25.0.24.07.28.zip -d /tmp
+  unzip -qo /tmp/v2.28.0.24.10.29.zip -d /tmp
   echo "Finishing unzip qnn sdk."
 
 

.ci/scripts/test_llama.sh

@@ -121,7 +121,7 @@ echo "COREML option ${COREML}"
 if [[ "${MODE}" =~ .*qnn.* ]]; then
   QNN=ON
   export EXECUTORCH_ROOT="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." && pwd)"
-  export QNN_SDK_ROOT=/tmp/qnn/2.25.0.240728
+  export QNN_SDK_ROOT=/tmp/qnn/2.28.0.241029
   export LD_LIBRARY_PATH="${QNN_SDK_ROOT}/lib/x86_64-linux-clang"
   export PYTHONPATH=".."
   cp schema/program.fbs exir/_serialize/program.fbs

.gitmodules

@@ -66,7 +66,7 @@
 	url = https://github.com/pybind/pybind11.git
 [submodule "backends/cadence/fusion_g3/third-party/nnlib/nnlib-FusionG3"]
 	path = backends/cadence/fusion_g3/third-party/nnlib/nnlib-FusionG3
-	url = https://github.com/foss-xtensa/nnlib-FusionG3/
+	url = https://github.com/foss-xtensa/nnlib-FusionG3.git
 [submodule "third-party/ao"]
 	path = third-party/ao
 	url = https://github.com/pytorch/ao.git

backends/apple/mps/mps_preprocess.py

@@ -32,6 +32,9 @@
     CompileSpec,
     PreprocessResult,
 )
+
+from executorch.exir.passes.memory_format_ops_pass import DimOrderOpsRevertPass
+from executorch.exir.program._program import _transform
 from torch.export.exported_program import ExportedProgram
 
 FORMAT = "[%(levelname)s %(asctime)s %(filename)s:%(lineno)s] %(message)s"
@@ -83,6 +86,9 @@ def preprocess(
         #    FlatBuffer graph, process the `output` nodes and add their id to
         #    the `output_ids` array in the schema.
 
+        # TODO: Remove this once we have a better support for the dim-order ops.
+        edge_program = _transform(edge_program, DimOrderOpsRevertPass())
+
         mps_graph = MPSGraph(
             version="0",
             mps_nodes=[],

backends/apple/mps/operators/constant_ops.py

@@ -79,6 +79,25 @@ def define_node(
         )
 
 
+@register_node_visitor
+class ToDimOrderEmptyVisitor(NodeVisitor):
+    target = ["dim_order_ops._empty_dim_order.default"]
+
+    def __init__(self, *args) -> None:
+        super().__init__(*args)
+
+    def define_node(
+        self,
+        node: torch.fx.Node,
+        mps_graph: MPSGraph,
+    ) -> None:
+        # We should never get here, because DimOrderOpsRevertPass replaces this with an aten.empty.memory_format op
+        # But if we do, we can't handle it ATM, so raise an exception
+        raise NotImplementedError(
+            "dim_order_ops._empty_dim_order.default is not supported yet"
+        )
+
+
 @register_node_visitor
 class FullLikeVisitor(NodeVisitor):
     target = "aten.full_like.default"

backends/apple/mps/operators/op_clone.py

@@ -33,3 +33,22 @@ def define_node(
                 )
         input_id = self.define_tensor(get_input_node(node, 0), mps_graph)
         self.tensor_to_id[node] = input_id
+
+
+@register_node_visitor
+class ToDimOrderCopyVisitor(NodeVisitor):
+    target = ["dim_order_ops._to_dim_order_copy.default"]
+
+    def __init__(self, *args) -> None:
+        super().__init__(*args)
+
+    def define_node(
+        self,
+        node: torch.fx.Node,
+        mps_graph: MPSGraph,
+    ) -> None:
+        # We should never get here, because DimOrderOpsRevertPass replaces this with an aten._to_copy op
+        # But if we do, we can't handle it ATM, so raise an exception
+        raise NotImplementedError(
+            "dim_order_ops._to_dim_order_copy.default is not supported yet"
+        )

backends/apple/mps/test/test_mps.py

@@ -1829,6 +1829,21 @@ def forward(self, x):
             Clone(), model_inputs, func_name=inspect.stack()[0].function[5:]
         )
 
+    def test_mps_backend_to_copy(self):
+        class Copy(torch.nn.Module):
+            def forward(self, x):
+                return (
+                    torch.ops.aten._to_copy.default(
+                        x + 2, memory_format=torch.contiguous_format
+                    )
+                    + x
+                )
+
+        model_inputs = (torch.randn(1, 3, 3),)
+        self.lower_and_test_with_partitioner(
+            Copy(), model_inputs, func_name=inspect.stack()[0].function[5:]
+        )
+
     def test_mps_backend_floor(self):
         class Floor(torch.nn.Module):
             def forward(self, x):

backends/apple/mps/test/test_mps_utils.py

@@ -26,10 +26,7 @@
 
 # Config for Capturing the weights, will be moved in the future
 
-# TODO(T182928844): Delegate dim order op to backend.
-_EDGE_COMPILE_CONFIG = exir.EdgeCompileConfig(
-    _check_ir_validity=False, _skip_dim_order=True
-)
+_EDGE_COMPILE_CONFIG = exir.EdgeCompileConfig(_check_ir_validity=False)
 
 
 class ansi_colors:
@@ -219,7 +216,6 @@ def lower_module_and_test_output(
             dynamic_shapes=dynamic_shapes,
             edge_compile_config=EdgeCompileConfig(
                 _check_ir_validity=False,
-                _skip_dim_order=True,  # TODO(T182928844): Delegate dim order op to backend.
             ),
         )
 
@@ -250,7 +246,6 @@ def lower_module_and_test_output(
                 export(delegated_program, sample_inputs, strict=True),
                 compile_config=exir.EdgeCompileConfig(
                     _check_ir_validity=False,
-                    _skip_dim_order=True,  # TODO(T182928844): Delegate dim order op to backend.
                 ),
             ).to_executorch(
                 config=ExecutorchBackendConfig(extract_delegate_segments=False)

backends/arm/README.md

@@ -119,7 +119,7 @@ backends/arm/test/setup_testing.sh
 The you can run the tests with
 
 ```
-pytest -c /dev/null -v -n auto backends/arm/test --arm_quantize_io --arm_run_corstoneFVP
+pytest -c /dev/null -v -n auto backends/arm/test --arm_run_corstoneFVP
 ```
 
 ### Code coverage