rel-1.22.2 cherry-pick 1 (#25633)

### Description
- #24265
- #24616
- #24640
- #24707
- #24646
- #24750
- #24809
- #24895
- #24820
- #25002
- #25171
- #25283
- #24818
- #25351
- #25361
- #25388
- #25520
- #25158




### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->

---------

Co-authored-by: quic-zhaoxul <[email protected]>
Co-authored-by: Yuduo Wu <[email protected]>
Co-authored-by: Hector Li <[email protected]>
Co-authored-by: chenweng-quic <[email protected]>
Co-authored-by: qti-yuduo <[email protected]>
Co-authored-by: Akupadhye <[email protected]>
Co-authored-by: Jeff Kilpatrick <[email protected]>
Co-authored-by: Jeff Kilpatrick <[email protected]>
Co-authored-by: George Wu <[email protected]>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: quic-calvnguy <[email protected]>
Co-authored-by: Changming Sun <[email protected]>
Co-authored-by: Yulong Wang <[email protected]>

Author: 14 people

.github/workflows/linux_cuda_ci.yml

@@ -28,7 +28,7 @@ jobs:
       dockerfile_path: tools/ci_build/github/linux/docker/Dockerfile.manylinux2_28_cuda
       docker_build_args: '--build-arg BASEIMAGE=onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20250124.1'
       docker_image_repo: onnxruntimecuda12manylinuxbuild
-      extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --enable_cuda_profiling --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=80 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
+      extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --enable_cuda_profiling --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=90 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
       python_path_prefix: 'PATH=/opt/python/cp310-cp310/bin:$PATH'
       run_tests: false            # <<< Do not run tests in this job
       upload_build_output: true   # <<< Upload the build/Release directory
@@ -41,7 +41,7 @@ jobs:
     needs: build-linux-cuda-x64-release
     runs-on:
       - self-hosted
-      - "1ES.Pool=Onnxruntime-github-Linux-GPU-A100-WUS3"
+      - "1ES.Pool=Onnxruntime-github-Linux-GPU-H100"
     permissions:
       contents: read
       packages: read
@@ -98,5 +98,5 @@ jobs:
           build_config: Release
           mode: 'test' # Set mode to test
           execution_providers: 'cuda'
-          extra_build_flags: '--use_binskim_compliant_compile_flags --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --enable_cuda_profiling --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=80 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
+          extra_build_flags: '--use_binskim_compliant_compile_flags --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --enable_cuda_profiling --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=90 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
           python_path_prefix: 'PATH=/opt/python/cp310-cp310/bin:$PATH'

.github/workflows/linux_tensorrt_ci.yml

@@ -28,7 +28,7 @@ jobs:
       dockerfile_path: tools/ci_build/github/linux/docker/Dockerfile.manylinux2_28_cuda
       docker_build_args: '--build-arg BASEIMAGE=onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20250124.1 --build-arg TRT_VERSION=10.9.0.34-1.cuda12.8 --network=host'
       docker_image_repo: onnxruntimetensorrt86gpubuild
-      extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --use_tensorrt --tensorrt_home /usr  --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=80 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
+      extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --use_tensorrt --tensorrt_home /usr  --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=90 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
       python_path_prefix: 'PATH=/opt/python/cp310-cp310/bin:$PATH'
       run_tests: false            # <<< Do not run tests in this job
       upload_build_output: true   # <<< Upload the build/Release directory
@@ -41,7 +41,7 @@ jobs:
     needs: build-linux-TensorRT-x64-release
     runs-on:
       - self-hosted
-      - "1ES.Pool=Onnxruntime-github-Linux-GPU-A100-WUS3"
+      - "1ES.Pool=Onnxruntime-github-Linux-GPU-H100"
     permissions:
       contents: read
       packages: read
@@ -100,5 +100,5 @@ jobs:
           build_config: Release
           mode: 'test' # Set mode to test
           execution_providers: 'cuda tensorrt'
-          extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --use_tensorrt --tensorrt_home /usr  --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=80 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
+          extra_build_flags: '--use_binskim_compliant_compile_flags --build_wheel --cuda_version=12.2 --cuda_home=/usr/local/cuda-12.2 --cudnn_home=/usr/local/cuda-12.2 --use_tensorrt --tensorrt_home /usr  --build_java --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=90 onnxruntime_BUILD_UNIT_TESTS=ON onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS=ON'
           python_path_prefix: 'PATH=/opt/python/cp310-cp310/bin:$PATH'

VERSION_NUMBER

@@ -1 +1 @@
-1.22.1
+1.22.2

cgmanifests/cgmanifest.json

@@ -36,7 +36,7 @@
       "component": {
         "type": "git",
         "git": {
-          "commitHash": "bee4d1dd8dc1ee4a1fd8fa6a96476c2f8b7492a3",
+          "commitHash": "5c210da409e7f1e51ddf445134a4376fdbd70d7d",
           "repositoryUrl": "https://github.com/dmlc/dlpack.git"
         }
       }
@@ -316,16 +316,6 @@
         "comments": "gtest-ios-framework"
       }
     },
-    {
-      "component": {
-        "type": "git",
-        "git": {
-          "commitHash": "277508879878e0a5b5b43599b1bea11f66eb3c6c",
-          "repositoryUrl": "https://github.com/dmlc/dlpack.git"
-        },
-        "comments": "dlpack"
-      }
-    },
     {
       "component": {
         "Type": "other",

cmake/CMakeLists.txt

@@ -150,6 +150,7 @@ option(onnxruntime_DISABLE_SPARSE_TENSORS "Disable sparse tensors data types" OF
 option(onnxruntime_DISABLE_OPTIONAL_TYPE "Disable optional type" OFF)
 option(onnxruntime_DISABLE_FLOAT8_TYPES "Disable float 8 types" OFF)
 option(onnxruntime_MINIMAL_BUILD "Exclude as much as possible from the build. Support ORT format models. No support for ONNX format models." OFF)
+option(onnxruntime_CLIENT_PACKAGE_BUILD "Enables default settings that are more appropriate for client/on-device workloads." OFF)
 cmake_dependent_option(onnxruntime_DISABLE_RTTI "Disable RTTI" ON "NOT onnxruntime_ENABLE_PYTHON;NOT onnxruntime_USE_CUDA" OFF)
 # For now onnxruntime_DISABLE_EXCEPTIONS will only work with onnxruntime_MINIMAL_BUILD, more changes (ONNX, non-CPU EP, ...) are required to run this standalone
 cmake_dependent_option(onnxruntime_DISABLE_EXCEPTIONS "Disable exception handling. Requires onnxruntime_MINIMAL_BUILD currently." ON "onnxruntime_MINIMAL_BUILD;NOT onnxruntime_ENABLE_PYTHON" OFF)

cmake/adjust_global_compile_flags.cmake

@@ -98,6 +98,11 @@ if (onnxruntime_MINIMAL_BUILD)
   endif()
 endif()
 
+# ORT build with default settings more appropriate for client/on-device workloads.
+if (onnxruntime_CLIENT_PACKAGE_BUILD)
+  add_compile_definitions(ORT_CLIENT_PACKAGE_BUILD)
+endif()
+
 if (onnxruntime_ENABLE_LTO)
     include(CheckIPOSupported)
     check_ipo_supported(RESULT ipo_enabled OUTPUT ipo_output)

cmake/deps.txt

@@ -16,7 +16,7 @@ abseil_cpp;https://github.com/abseil/abseil-cpp/archive/refs/tags/20240722.0.zip
 coremltools;https://github.com/apple/coremltools/archive/refs/tags/7.1.zip;f1bab0f30966f2e217d8e01207d518f230a1641a
 cxxopts;https://github.com/jarro2783/cxxopts/archive/3c73d91c0b04e2b59462f0a741be8c07024c1bc0.zip;6c6ca7f8480b26c8d00476e0e24b7184717fe4f0
 date;https://github.com/HowardHinnant/date/archive/refs/tags/v3.0.1.zip;2dac0c81dc54ebdd8f8d073a75c053b04b56e159
-dlpack;https://github.com/dmlc/dlpack/archive/refs/tags/v0.6.zip;4d565dd2e5b31321e5549591d78aa7f377173445
+dlpack;https://github.com/dmlc/dlpack/archive/5c210da409e7f1e51ddf445134a4376fdbd70d7d.zip;e499c86e4e5c5268a87661d7ea39c27fae10907c
 # This Eigen commit id matches the eigen archive being consumed from https://gitlab.com/libeigen/eigen/-/archive/3.4/eigen-3.4.zip
 # prior to the 3.4.1 RC changing the bits and invalidating the hash.
 # it contains changes on top of 3.4.0 which are required to fix build issues.

cmake/onnxruntime_mlas.cmake

@@ -31,6 +31,7 @@ onnxruntime_add_static_library(onnxruntime_mlas
   ${MLAS_SRC_DIR}/eltwise.cpp
   ${MLAS_SRC_DIR}/erf.cpp
   ${MLAS_SRC_DIR}/compute.cpp
+  ${MLAS_SRC_DIR}/dequantize.cpp
   ${MLAS_SRC_DIR}/quantize.cpp
   ${MLAS_SRC_DIR}/qgemm_kernel_default.cpp
   ${MLAS_SRC_DIR}/qladd.cpp

cmake/onnxruntime_providers_qnn.cmake

@@ -66,10 +66,10 @@
         COMMAND ${CMAKE_COMMAND} -E copy ${QNN_LIB_FILES} $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>
         )
     endif()
-    if (EXISTS "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf")
+    if (EXISTS "${onnxruntime_QNN_HOME}/LICENSE.pdf")
       add_custom_command(
         TARGET ${onnxruntime_providers_qnn_target} POST_BUILD
-        COMMAND ${CMAKE_COMMAND} -E copy "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf" $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>
+        COMMAND ${CMAKE_COMMAND} -E copy "${onnxruntime_QNN_HOME}/LICENSE.pdf" $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>/Qualcomm_LICENSE.pdf
         )
     endif()
   else()
@@ -154,10 +154,10 @@
         COMMAND ${CMAKE_COMMAND} -E copy ${QNN_LIB_FILES} $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>
         )
     endif()
-    if (EXISTS "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf")
+    if (EXISTS "${onnxruntime_QNN_HOME}/LICENSE.pdf")
       add_custom_command(
         TARGET ${onnxruntime_providers_qnn_target} POST_BUILD
-        COMMAND ${CMAKE_COMMAND} -E copy "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf" $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>
+        COMMAND ${CMAKE_COMMAND} -E copy "${onnxruntime_QNN_HOME}/LICENSE.pdf" $<TARGET_FILE_DIR:${onnxruntime_providers_qnn_target}>/Qualcomm_LICENSE.pdf
         )
     endif()
   endif()

cmake/onnxruntime_python.cmake

@@ -1050,18 +1050,10 @@ if (onnxruntime_USE_QNN)
         ${QNN_LIB_FILES}
         $<TARGET_FILE_DIR:${build_output_target}>/onnxruntime/capi/
   )
-  add_custom_command(
-    TARGET onnxruntime_pybind11_state POST_BUILD
-    COMMAND ${CMAKE_COMMAND} -E copy
-        $<TARGET_FILE:ep_weight_sharing_ctx_gen>
-        $<TARGET_FILE_DIR:${build_output_target}>/onnxruntime/capi/
-  )
-  if (EXISTS "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf")
+  if (EXISTS "${onnxruntime_QNN_HOME}/LICENSE.pdf")
     add_custom_command(
       TARGET onnxruntime_pybind11_state POST_BUILD
-      COMMAND ${CMAKE_COMMAND} -E copy
-          "${onnxruntime_QNN_HOME}/Qualcomm AI Hub Proprietary License.pdf"
-          $<TARGET_FILE_DIR:${build_output_target}>/onnxruntime/
+        COMMAND ${CMAKE_COMMAND} -E copy "${onnxruntime_QNN_HOME}/LICENSE.pdf" $<TARGET_FILE_DIR:${build_output_target}>/onnxruntime/Qualcomm_LICENSE.pdf
     )
   endif()
 endif()