diff --git a/.pipelines/OneBranch.Nuget-WindowsAI-Pipeline.Official.yml b/.pipelines/OneBranch.Nuget-WindowsAI-Pipeline.Official.yml
index fa2b475fe9..b9de1b79e1 100644
--- a/.pipelines/OneBranch.Nuget-WindowsAI-Pipeline.Official.yml
+++ b/.pipelines/OneBranch.Nuget-WindowsAI-Pipeline.Official.yml
@@ -351,6 +351,31 @@ extends:
- script: |
dir $(Build.SourcesDirectory)\unzipped\runtimes\win-x64\_native
+ - task: EsrpCodeSigning@2
+ displayName: "Sign Nuget package"
+ inputs:
+ ConnectedServiceName: 'OnnxRuntime CodeSign 20190817'
+ FolderPath: $(Build.ArtifactStagingDirectory)
+ Pattern: '*.nupkg'
+ signConfigType: inlineSignParams
+ inlineOperation: |
+ [
+ {
+ "keyCode": "CP-401405",
+ "operationSetCode": "NuGetSign",
+ "parameters": [ ],
+ "toolName": "sign",
+ "toolVersion": "1.0"
+ },
+ {
+ "keyCode": "CP-401405",
+ "operationSetCode": "NuGetVerify",
+ "parameters": [ ],
+ "toolName": "sign",
+ "toolVersion": "1.0"
+ }
+ ]
+
- job: NuGet_Publishing
pool:
type: windows
diff --git a/.pipelines/nuget_config/x64/packages.config b/.pipelines/nuget_config/x64/packages.config
index 2ec9b577c1..5dfeb0836f 100644
--- a/.pipelines/nuget_config/x64/packages.config
+++ b/.pipelines/nuget_config/x64/packages.config
@@ -1,6 +1,6 @@
-
+
diff --git a/.pipelines/nuget_config/x86/packages.config b/.pipelines/nuget_config/x86/packages.config
index 6f9e3ef496..f8e63133f4 100644
--- a/.pipelines/nuget_config/x86/packages.config
+++ b/.pipelines/nuget_config/x86/packages.config
@@ -1,6 +1,6 @@
-
+
diff --git a/.pipelines/windowsai-steps.yml b/.pipelines/windowsai-steps.yml
index 0b736da427..a29e2fe6c8 100644
--- a/.pipelines/windowsai-steps.yml
+++ b/.pipelines/windowsai-steps.yml
@@ -80,11 +80,11 @@ jobs:
# must call vsdevcmd first to add cmake to PATH
- script: |
- curl -O -L https://github.com/Kitware/CMake/releases/download/v3.24.3/cmake-3.24.3-windows-x86_64.zip
- 7z x cmake-3.24.3-windows-x86_64.zip
+ curl -O -L https://github.com/Kitware/CMake/releases/download/v3.26.3/cmake-3.26.3-windows-x86_64.zip
+ 7z x cmake-3.26.3-windows-x86_64.zip
set PYTHONHOME=$(Build.BinariesDirectory)\${{ parameters.PythonPackageName }}.3.9.7\tools
set PYTHONPATH=$(Build.BinariesDirectory)\${{ parameters.PythonPackageName }}.3.9.7\tools
- $(Build.BinariesDirectory)\${{ parameters.PythonPackageName }}.3.9.7\tools\python.exe "$(Build.SourcesDirectory)\tools\ci_build\build.py" --build_dir $(Build.BinariesDirectory) --build_shared_lib --enable_onnx_tests --ms_experimental --use_dml --use_winml --cmake_generator "Visual Studio 16 2019" --update --config RelWithDebInfo --enable_lto --use_telemetry --disable_rtti --enable_wcos $(BuildFlags) --cmake_extra_defines CMAKE_SYSTEM_VERSION=10.0.19041.0 --cmake_path $(Build.BinariesDirectory)\cmake-3.24.3-windows-x86_64\bin\cmake.exe --ctest_path $(Build.BinariesDirectory)\cmake-3.24.3-windows-x86_64\bin\ctest.exe
+ $(Build.BinariesDirectory)\${{ parameters.PythonPackageName }}.3.9.7\tools\python.exe "$(Build.SourcesDirectory)\tools\ci_build\build.py" --build_dir $(Build.BinariesDirectory) --build_shared_lib --enable_onnx_tests --ms_experimental --use_dml --use_winml --cmake_generator "Visual Studio 16 2019" --update --config RelWithDebInfo --enable_lto --use_telemetry --disable_rtti --enable_wcos $(BuildFlags) --cmake_extra_defines CMAKE_SYSTEM_VERSION=10.0.19041.0 --cmake_path $(Build.BinariesDirectory)\cmake-3.26.3-windows-x86_64\bin\cmake.exe --ctest_path $(Build.BinariesDirectory)\cmake-3.26.3-windows-x86_64\bin\ctest.exe
workingDirectory: '$(Build.BinariesDirectory)'
displayName: 'Generate cmake config'
diff --git a/cmake/CMakeLists.txt b/cmake/CMakeLists.txt
index 0b2f980ab2..17fdc9a86e 100644
--- a/cmake/CMakeLists.txt
+++ b/cmake/CMakeLists.txt
@@ -272,6 +272,9 @@ if (onnxruntime_USE_ROCM)
file(GLOB rocm_cmake_components ${onnxruntime_ROCM_HOME}/lib/cmake/*)
list(APPEND CMAKE_PREFIX_PATH ${rocm_cmake_components})
+ # Force cmake to accept the configured HIP compiler. Because the configured CMAKE_PREFIX_PATH does not work during
+ # enable_language(HIP), we might need to move configuring of CMAKE_PREFIX_PATH to build.py (in the future).
+ set(CMAKE_HIP_COMPILER_FORCED ON)
enable_language(HIP)
# NOTE: Flags -mllvm -amdgpu-early-inline-all=true are critical for gpu kernel code performance. -mllvm passes the
diff --git a/cmake/external/dml.cmake b/cmake/external/dml.cmake
index 161db093a1..4dec75deca 100644
--- a/cmake/external/dml.cmake
+++ b/cmake/external/dml.cmake
@@ -41,7 +41,7 @@ if (NOT onnxruntime_USE_CUSTOM_DIRECTML)
set(NUGET_CONFIG ${PROJECT_SOURCE_DIR}/../NuGet.config)
set(PACKAGES_CONFIG ${PROJECT_SOURCE_DIR}/../packages.config)
get_filename_component(PACKAGES_DIR ${CMAKE_CURRENT_BINARY_DIR}/../packages ABSOLUTE)
- set(DML_PACKAGE_DIR ${PACKAGES_DIR}/Microsoft.AI.DirectML.1.10.1)
+ set(DML_PACKAGE_DIR ${PACKAGES_DIR}/Microsoft.AI.DirectML.1.11.0)
# Restore nuget packages, which will pull down the DirectML redist package.
add_custom_command(
diff --git a/docs/OperatorKernels.md b/docs/OperatorKernels.md
index 5d9b3d582e..97051e99ef 100644
--- a/docs/OperatorKernels.md
+++ b/docs/OperatorKernels.md
@@ -1193,7 +1193,6 @@ Do not modify directly.*
|DequantizeLinear|*in* x:**T1**
*in* x_scale:**T2**
*in* x_zero_point:**T1**
*out* y:**T2**|1+|**T1** = tensor(float)
**T2** = tensor(uint8)|
|EmbedLayerNormalization|*in* input_ids:**T1**
*in* segment_ids:**T1**
*in* word_embedding:**T**
*in* position_embedding:**T**
*in* segment_embedding:**T**
*in* gamma:**T**
*in* beta:**T**
*in* mask:**T1**
*in* position_ids:**T1**
*out* output:**T**
*out* mask_index:**T1**
*out* embedding_sum:**T**|1+|**T** = tensor(float), tensor(float16)|
|FusedMatMul|*in* A:**T**
*in* B:**T**
*out* Y:**T**|1+|**T** = tensor(float), tensor(float16)|
-|FusedMatMulActivation|*in* A:**T**
*in* B:**T**
*out* Y:**T**|1+|**T** = tensor(float), tensor(float16)|
|Gelu|*in* X:**T**
*out* Y:**T**|1+|**T** = tensor(float), tensor(float16)|
|GroupNorm|*in* X:**T**
*in* gamma:**M**
*in* beta:**M**
*out* Y:**T**|1+|**M** = tensor(float), tensor(float16)
**T** = tensor(float), tensor(float16)|
|NhwcConv|*in* X:**T**
*in* W:**T**
*in* B:**T**
*out* Y:**T**|1+|**T** = tensor(float), tensor(float16)|
diff --git a/onnxruntime/core/optimizer/graph_transformer_utils.cc b/onnxruntime/core/optimizer/graph_transformer_utils.cc
index 7196465e5d..73403f2c62 100644
--- a/onnxruntime/core/optimizer/graph_transformer_utils.cc
+++ b/onnxruntime/core/optimizer/graph_transformer_utils.cc
@@ -182,7 +182,6 @@ InlinedVector> GenerateTransformers(
#ifndef DISABLE_CONTRIB_OPS
const InlinedHashSet cpu_ep = {onnxruntime::kCpuExecutionProvider};
#endif
- const InlinedHashSet dml_ep = {onnxruntime::kDmlExecutionProvider};
switch (level) {
case TransformerLevel::Level1: {
// RewriteRule optimizations are the simplest (they generally remove unnecessary nodes and are cheap to run)
@@ -309,7 +308,6 @@ InlinedVector> GenerateTransformers(
transformers.emplace_back(std::make_unique(cpu_cuda_dml_rocm_eps));
transformers.emplace_back(std::make_unique(cpu_cuda_dml_rocm_eps));
- transformers.emplace_back(std::make_unique(dml_ep));
// GeluApproximation has side effects which may change results. It needs to be manually enabled,
// or alternatively the model can be updated offline using a model conversion script
diff --git a/onnxruntime/core/providers/dml/DmlExecutionProvider/src/Operators/OperatorRegistration.cpp b/onnxruntime/core/providers/dml/DmlExecutionProvider/src/Operators/OperatorRegistration.cpp
index 34a8f496d7..fd0ad8385f 100644
--- a/onnxruntime/core/providers/dml/DmlExecutionProvider/src/Operators/OperatorRegistration.cpp
+++ b/onnxruntime/core/providers/dml/DmlExecutionProvider/src/Operators/OperatorRegistration.cpp
@@ -871,7 +871,6 @@ constexpr static OperatorRegistrationInformation operatorRegistrationInformation
{REG_INFO_MS( 1, Gelu, typeNameListDefault, supportedTypeListFloat16to32, DmlGraphSupport::Supported)},
{REG_INFO_MS( 1, BiasGelu, typeNameListDefault, supportedTypeListFloat16to32, DmlGraphSupport::Supported)},
{REG_INFO_MS( 1, FusedMatMul, typeNameListDefault, supportedTypeListFloat16to32, DmlGraphSupport::Supported)},
- {REG_INFO_MS( 1, FusedMatMulActivation, typeNameListDefault, supportedTypeListFloat16to32, DmlGraphSupport::Supported)},
{REG_INFO_MS( 1, QLinearSigmoid, typeNameListDefault, supportedTypeListQLinearSigmoid, DmlGraphSupport::Supported, requiredConstantCpuInputs(), std::nullopt, QueryQLinearSigmoid)},
{REG_INFO_MS( 1, Attention, typeNameListAttention, supportedTypeListAttention, DmlGraphSupport::Supported, requiredConstantCpuInputs(), std::nullopt, QueryAttention)},
diff --git a/onnxruntime/test/shared_lib/test_inference.cc b/onnxruntime/test/shared_lib/test_inference.cc
index af1df0ed63..328652f03a 100644
--- a/onnxruntime/test/shared_lib/test_inference.cc
+++ b/onnxruntime/test/shared_lib/test_inference.cc
@@ -2827,7 +2827,7 @@ TEST(CApiTest, TestMultiStreamInferenceSimpleSSD) {
}
#endif
-#if !defined(ORT_MINIMAL_BUILD)
+#if !defined(ORT_MINIMAL_BUILD) && !defined(REDUCED_OPS_BUILD)
TEST(MultiKernelSingleSchemaTest, valid) {
Ort::SessionOptions session_options;
session_options.SetIntraOpNumThreads(1);
diff --git a/packages.config b/packages.config
index d679d2ceab..97d03b55b6 100644
--- a/packages.config
+++ b/packages.config
@@ -1,6 +1,6 @@
-
+
diff --git a/tools/android_custom_build/Dockerfile b/tools/android_custom_build/Dockerfile
index 1613e63351..7e67cc9214 100644
--- a/tools/android_custom_build/Dockerfile
+++ b/tools/android_custom_build/Dockerfile
@@ -24,9 +24,9 @@ RUN apt-get update && apt-get install --yes --no-install-recommends \
unzip
# cmake
-RUN CMAKE_VERSION=3.25.2 && \
+RUN CMAKE_VERSION=3.26.3 && \
aria2c -q -d /tmp -o cmake-${CMAKE_VERSION}-linux-x86_64.tar.gz \
- --checksum=sha-256=783da74f132fd1fea91b8236d267efa4df5b91c5eec1dea0a87f0cf233748d99 \
+ --checksum=sha-256=28d4d1d0db94b47d8dfd4f7dec969a3c747304f4a28ddd6fd340f553f2384dc2 \
https://github.com/Kitware/CMake/releases/download/v${CMAKE_VERSION}/cmake-${CMAKE_VERSION}-linux-x86_64.tar.gz && \
tar -zxf /tmp/cmake-${CMAKE_VERSION}-linux-x86_64.tar.gz --strip=1 -C /usr
diff --git a/tools/ci_build/github/pai/rocm-ci-pipeline-env.Dockerfile b/tools/ci_build/github/pai/rocm-ci-pipeline-env.Dockerfile
index 605427ab7d..f6b55af38b 100644
--- a/tools/ci_build/github/pai/rocm-ci-pipeline-env.Dockerfile
+++ b/tools/ci_build/github/pai/rocm-ci-pipeline-env.Dockerfile
@@ -10,7 +10,7 @@ RUN cd /opt/mpi_install/ucx/build &&\
make install
# CMake
-ENV CMAKE_VERSION=3.24.2
+ENV CMAKE_VERSION=3.26.3
RUN cd /usr/local && \
wget -q -O - https://github.com/Kitware/CMake/releases/download/v${CMAKE_VERSION}/cmake-${CMAKE_VERSION}-Linux-x86_64.tar.gz | tar zxf -
ENV PATH=/usr/local/cmake-${CMAKE_VERSION}-linux-x86_64/bin:${PATH}
diff --git a/tools/nuget/generate_nuspec_for_native_nuget.py b/tools/nuget/generate_nuspec_for_native_nuget.py
index 71e998e2c7..f6cfcc3ce2 100644
--- a/tools/nuget/generate_nuspec_for_native_nuget.py
+++ b/tools/nuget/generate_nuspec_for_native_nuget.py
@@ -190,7 +190,7 @@ def generate_repo_url(line_list, repo_url, commit_id):
def generate_dependencies(xml_text, package_name, version):
- dml_dependency = ''
+ dml_dependency = ''
if package_name == "Microsoft.AI.MachineLearning":
xml_text.append("")