Integrate Torch-MLIR to llvm/torch-mlir@1828c5053 (#24691)
Integrates Torch-MLIR via `iree-org/torch-mlir/sm-iree-integrates/torch-20260707` branch.
Includes a CMake fix for https://github.com/llvm/torch-mlir/commit/35010fc69.
Promotes ONNX ops test case `test_convtranspose_output_shape` to an expected pass.
On top of the integrated upstream commit, the 3 latest fixes from
`sm-iree-integrates/llvm-20260617` are carried over:
"Adaptions to llvm/llvm-project@75b7809248cc for integrating LLVM into
IREE" by @schuermans-roofline as commit ec8720b
"Adaptions to llvm/llvm-project@917117ceeebf for integrating LLVM into
IREE" by @schuermans-roofline as commit 67c8c41
"Bump LLVM to llvm/llvm-project@0f3ca6bb9ca5" by @lialan as commit
ba9dad2
---------
Signed-off-by: Artem Gindinson <gindinson@roofline.ai>
diff --git a/compiler/bindings/python/CMakeLists.txt b/compiler/bindings/python/CMakeLists.txt
index 35c70da..5129153 100644
--- a/compiler/bindings/python/CMakeLists.txt
+++ b/compiler/bindings/python/CMakeLists.txt
@@ -181,6 +181,7 @@
ADD_TO_PARENT IREEPythonSources
ROOT_DIR "${TORCH_MLIR_ROOT_DIR}/python/torch_mlir"
SOURCES
+ extras/fx_as_strided.py
extras/fx_importer.py
extras/onnx_importer.py
)
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json
index ef4bdd2..09edc6a 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O0.json
@@ -310,7 +310,6 @@
"onnx/node/generated/test_bernoulli_double",
"onnx/node/generated/test_bernoulli_double_expanded",
"onnx/node/generated/test_bernoulli_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_det_2d",
"onnx/node/generated/test_gridsample_nearest",
"onnx/node/generated/test_gridsample_nearest_align_corners_0_additional_1",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json
index 15bfd0b..9f9dae1 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_cpu_llvm_sync_O2.json
@@ -311,7 +311,6 @@
"onnx/node/generated/test_bernoulli_double",
"onnx/node/generated/test_bernoulli_double_expanded",
"onnx/node/generated/test_bernoulli_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_det_2d",
"onnx/node/generated/test_gridsample_nearest",
"onnx/node/generated/test_gridsample_nearest_align_corners_0_additional_1",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_cuda_O0.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_cuda_O0.json
index 758590f..da4c191 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_cuda_O0.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_cuda_O0.json
@@ -112,7 +112,6 @@
"onnx/node/generated/test_compress_negative_axis",
"onnx/node/generated/test_convtranspose_autopad_same",
"onnx/node/generated/test_convtranspose_kernel_shape",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_cumsum_1d",
"onnx/node/generated/test_cumsum_1d_exclusive",
"onnx/node/generated/test_cumsum_1d_reverse",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O0.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O0.json
index 8611125..c2a18bc 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O0.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O0.json
@@ -314,7 +314,6 @@
"onnx/node/generated/test_cast_FLOAT_to_BFLOAT16",
"onnx/node/generated/test_castlike_FLOAT_to_BFLOAT16",
"onnx/node/generated/test_castlike_FLOAT_to_BFLOAT16_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_gridsample_nearest",
"onnx/node/generated/test_gridsample_nearest_align_corners_0_additional_1",
"onnx/node/generated/test_gridsample_nearest_align_corners_1_additional_1",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json
index 7d3132e..e03c849 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna3_O3.json
@@ -308,7 +308,7 @@
"onnx/node/generated/test_bernoulli_double",
"onnx/node/generated/test_bernoulli_double_expanded",
"onnx/node/generated/test_bernoulli_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
+ "onnx/node/generated/",
"onnx/node/generated/test_det_2d",
"onnx/node/generated/test_gridsample_nearest",
"onnx/node/generated/test_gridsample_nearest_align_corners_0_additional_1",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json
index a8a701b..8744f2a 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_hip_rdna4_O3.json
@@ -308,7 +308,6 @@
"onnx/node/generated/test_bernoulli_double",
"onnx/node/generated/test_bernoulli_double_expanded",
"onnx/node/generated/test_bernoulli_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_det_2d",
"onnx/node/generated/test_gridsample_nearest",
"onnx/node/generated/test_gridsample_nearest_align_corners_0_additional_1",
diff --git a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json
index 78de724..5ccfb66 100644
--- a/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json
+++ b/tests/external/iree-test-suites/onnx_ops/onnx_ops_gpu_vulkan_rdna3_O0.json
@@ -330,7 +330,6 @@
"onnx/node/generated/test_castlike_FLOAT_to_DOUBLE_expanded",
"onnx/node/generated/test_castlike_FLOAT_to_FLOAT16",
"onnx/node/generated/test_castlike_FLOAT_to_FLOAT16_expanded",
- "onnx/node/generated/test_convtranspose_output_shape",
"onnx/node/generated/test_det_2d",
"onnx/node/generated/test_dynamicquantizelinear",
"onnx/node/generated/test_dynamicquantizelinear_expanded",
diff --git a/third_party/torch-mlir b/third_party/torch-mlir
index 091bb8d..ec8720b 160000
--- a/third_party/torch-mlir
+++ b/third_party/torch-mlir
@@ -1 +1 @@
-Subproject commit 091bb8d3221457d25c079bed9750c5b6927f39bc
+Subproject commit ec8720bf26b0c7b35f98b93730b75fb2f4b0fccb
