Cherry-picks into rel-1.22.0 (microsoft#24580) (#680)

### Description

Cherry pick the following into
[rel-1.22.0](https://github.com/microsoft/onnxruntime/tree/rel-1.22.0)


- (microsoft#24487)
- (microsoft#24466)
- (microsoft#24493)
- (microsoft#24484)
- (microsoft#24494)
- (microsoft#24489)
- (microsoft#24504)
- (microsoft#24510)
- (microsoft#24456)
- (microsoft#24537)
- (microsoft#24501)
- (microsoft#24519)
- (microsoft#24513)
- (microsoft#24539)
- (microsoft#24514)
- (microsoft#24542)
- (microsoft#24585)

Not added:

Planning to cherry pick Cuda Matmulnbits PRs once the fix for failing
cuda pipeline is ready
- (microsoft#24491)
- (microsoft#24509)
- (microsoft#24564)

---------

Co-authored-by: vraspar <vrajang@outlook.com>
Co-authored-by: Adrian Lizarraga <adlizarraga@microsoft.com>
Co-authored-by: minfhong-quic <quic_minfhong@quicinc.com>
Co-authored-by: minfhong-quic <minfhong-quic@quicinc.com>
Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com>
Co-authored-by: Prathik Rao <prathik.rao@gmail.com>
Co-authored-by: Edward Chen <18449977+edgchen1@users.noreply.github.com>
Co-authored-by: Ankan Banerjee <ankan.ban@gmail.com>
Co-authored-by: Maximilian Müller <maximilianm@nvidia.com>
Co-authored-by: Gaurav Garg <gaugarg@nvidia.com>
Co-authored-by: iraut <iraut@nvidia.com>
Co-authored-by: Hrishikesh Manohar <hrishikeshm@nvidia.com>
Co-authored-by: Maximilian Müller <44298237+gedoensmax@users.noreply.github.com>
Co-authored-by: Scott McKay <skottmckay@gmail.com>
Co-authored-by: Jiajia Qin <jiajiaqin@microsoft.com>
Co-authored-by: kunal-vaishnavi <115581922+kunal-vaishnavi@users.noreply.github.com>
Co-authored-by: xhcao <xinghua.cao@intel.com>

Author: 17 people

cmake/CMakeLists.txt

@@ -107,6 +107,7 @@ option(onnxruntime_ENABLE_MICROSOFT_INTERNAL "Use this option to enable/disable
 option(onnxruntime_USE_VITISAI "Build with Vitis-AI" OFF)
 option(onnxruntime_USE_TENSORRT "Build with TensorRT support" OFF)
 option(onnxruntime_USE_TENSORRT_BUILTIN_PARSER "Use TensorRT builtin parser" OFF)
+option(onnxruntime_USE_NV "Build with TensorRT support" OFF)
 option(onnxruntime_ENABLE_LTO "Enable link time optimization" OFF)
 option(onnxruntime_CROSS_COMPILING "Cross compiling onnx runtime" OFF)
 option(onnxruntime_GCOV_COVERAGE "Compile with options necessary to run code coverage" OFF)
@@ -250,6 +251,7 @@ option(onnxruntime_USE_LOCK_FREE_QUEUE "Build with lock-free task queue for thre
 option(onnxruntime_FORCE_GENERIC_ALGORITHMS "Disable optimized arch-specific algorithms. Use only for testing and debugging generic algorithms." OFF)
 
 option(onnxruntime_USE_TENSORRT_INTERFACE "Build ONNXRuntime shared lib which is compatible with TensorRT EP interface" OFF)
+option(onnxruntime_USE_NV_INTERFACE "Build ONNXRuntime shared lib which is compatible with NV EP interface" OFF)
 option(onnxruntime_USE_CUDA_INTERFACE "Build ONNXRuntime shared lib which is compatible with Cuda EP interface" OFF)
 option(onnxruntime_USE_OPENVINO_INTERFACE "Build ONNXRuntime shared lib which is compatible with OpenVINO EP interface" OFF)
 option(onnxruntime_USE_VITISAI_INTERFACE "Build ONNXRuntime shared lib which is compatible with Vitis-AI EP interface" OFF)
@@ -946,6 +948,15 @@ if (onnxruntime_USE_TENSORRT_INTERFACE AND (NOT onnxruntime_USE_TENSORRT))
     list(APPEND ORT_INTERFACE_FLAGS -DUSE_TENSORRT=1)
 endif()
 
+if (onnxruntime_USE_NV)
+    list(APPEND ORT_PROVIDER_FLAGS -DUSE_NV=1)
+    list(APPEND ONNXRUNTIME_PROVIDER_NAMES nv_tensorrt_rtx)
+endif()
+
+if (onnxruntime_USE_NV_INTERFACE AND (NOT onnxruntime_USE_NV))
+    list(APPEND ORT_INTERFACE_FLAGS -DUSE_NV=1)
+endif()
+
 if (onnxruntime_USE_RKNPU)
     list(APPEND ORT_PROVIDER_FLAGS -DUSE_RKNPU=1)
     list(APPEND ONNXRUNTIME_PROVIDER_NAMES rknpu)

cmake/onnxruntime_framework.cmake

@@ -63,7 +63,7 @@ endif()
 if(onnxruntime_ENABLE_INSTRUMENT)
   target_compile_definitions(onnxruntime_framework PRIVATE ONNXRUNTIME_ENABLE_INSTRUMENT)
 endif()
-if(onnxruntime_USE_TENSORRT OR onnxruntime_USE_NCCL)
+if(onnxruntime_USE_TENSORRT OR onnxruntime_USE_NCCL OR onnxruntime_USE_NV)
 # TODO: for now, core framework depends on CUDA. It should be moved to TensorRT EP
 # TODO: provider_bridge_ort.cc should not include nccl.h
 target_include_directories(onnxruntime_framework PRIVATE ${ONNXRUNTIME_ROOT} PUBLIC ${CMAKE_CURRENT_BINARY_DIR} ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})