Fix Nuget Cuda pipeline package pipeline (#20741)

### Description
<!-- Describe your changes. -->

This PR adding protoc.exe to make the Nuget Cuda Pipleine, which also
allowing it to get build Java for various CUDA version

### 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. -->

tools/ci_build/github/azure-pipelines/cuda-packaging-pipeline.yml

@@ -102,7 +102,7 @@ stages:
       PreReleaseVersionSuffixNumber: ${{ parameters.PreReleaseVersionSuffixNumber }}
 
   # this is needed for certain artifacts to be published
-  - stage: Linux_C_API_Packaging_CPU_x64
+  - stage: Linux_C_API_Packaging_CPU
     dependsOn: [ ]
     jobs:
     - template: templates/c-api-linux-cpu.yml

tools/ci_build/github/azure-pipelines/stages/java-cuda-packaging-stage.yml

@@ -7,11 +7,13 @@ parameters:
   type: string
 
 stages:
-- stage: Java_GPU_Packaging
+- stage: Jar_Packaging_GPU
   dependsOn:
-  - Linux_C_API_Packaging_Combined_CUDA
-  - Windows_Packaging_CUDA
-  - Windows_Packaging_TensorRT
+  - Linux_C_API_Packaging_GPU
+# Because Java Jar is published only after Windows Packaging GPU Testing stage we need to depend on the Testing stages
+# TODO: change Windows_Packaging_*_Testing to Windows_Packaging_* once we finish PRODUCT BACKLOG ITEM 34666
+  - Windows_Packaging_CUDA_Testing
+  - Windows_Packaging_TensorRT_Testing
   - Download_Java_Tools
   jobs:
   - job: Jar_Packaging_GPU

tools/ci_build/github/azure-pipelines/stages/nuget-combine-cuda-stage.yml

@@ -57,6 +57,7 @@ stages:
     CudaVersion: ${{ parameters.CudaVersion }}
     win_trt_home: ${{ parameters.win_trt_home }}
     win_cuda_home: ${{ parameters.win_cuda_home }}
+    buildJava: ${{ parameters.buildJava }}
 
 - template: nuget-cuda-packaging-stage.yml
   parameters:

tools/ci_build/github/azure-pipelines/stages/nuget-cuda-packaging-stage.yml

@@ -16,10 +16,12 @@ stages:
     - Setup
     - Windows_Packaging_CUDA
     - Windows_Packaging_TensorRT
-    - Linux_C_API_Packaging_Combined_CUDA
+    - Linux_C_API_Packaging_GPU
+#    This is need for Download Linux CustomOp TestData
+    - Linux_C_API_Packaging_CPU
   condition: succeeded()
   jobs:
-    - job:
+    - job: NuGet_Packaging_GPU
       workspace:
         clean: all
       pool: 'Onnxruntime-Win-CPU-2022'
@@ -56,12 +58,6 @@ stages:
             ArtifactName: 'onnxruntime-linux-x64-tensorrt'
             TargetPath: '$(Build.BinariesDirectory)/nuget-artifact'
 
-        - template: ../templates/flex-downloadPipelineArtifact.yml
-          parameters:
-            StepName: 'Download Pipeline Artifact - NuGet'
-            ArtifactName: 'drop-extra'
-            TargetPath: '$(Build.BinariesDirectory)/extra-artifact'
-
         # Reconstruct the build dir
         - task: PowerShell@2
           displayName: 'PS: Extract nuget files gpu'

tools/ci_build/github/azure-pipelines/stages/nuget-linux-cuda-packaging-stage.yml

@@ -8,13 +8,11 @@ parameters:
   type: string
 - name: buildJava
   type: boolean
-  default: false
 - name: buildNodejs
   type: boolean
-  default: false
 
 stages:
-- stage: Linux_C_API_Packaging_Combined_CUDA
+- stage: Linux_C_API_Packaging_GPU
   dependsOn: []
   jobs:
   - job: Linux_C_API_Packaging_CUDA
@@ -51,14 +49,16 @@ stages:
       workingDirectory: $(Build.SourcesDirectory)
       displayName: 'Build CUDA C API Package'
 
-    - template: ../templates/java-api-artifacts-package-and-publish-steps-posix.yml
-      parameters:
-        arch: 'linux-x64'
-        buildConfig: 'Release'
-        artifactName: 'onnxruntime-java-linux-x64-cuda'
-        version: '$(OnnxRuntimeVersion)'
-        libraryName: 'libonnxruntime.so'
-        nativeLibraryName: 'libonnxruntime4j_jni.so'
+    - ${{ if eq(parameters.buildJava, true) }}:
+      - template: ../templates/java-api-artifacts-package-and-publish-steps-posix.yml
+        parameters:
+          arch: 'linux-x64'
+          buildConfig: 'Release'
+          artifactName: 'onnxruntime-java-linux-x64-cuda'
+          version: '$(OnnxRuntimeVersion)'
+          libraryName: 'libonnxruntime.so'
+          nativeLibraryName: 'libonnxruntime4j_jni.so'
+
     - template: ../templates/c-api-artifacts-package-and-publish-steps-posix.yml
       parameters:
         buildConfig: 'Release'
@@ -147,10 +147,11 @@ stages:
         condition: 'succeeded'
     - template: ../templates/clean-agent-build-directory-step.yml
   # Linux CUDA Combined Testing and Publishing
+- stage: Linux_Packaging_combined_CUDA
+  dependsOn:
+  - Linux_C_API_Packaging_GPU
+  jobs:
   - job: Linux_Packaging_combined_CUDA
-    dependsOn:
-    - Linux_C_API_Packaging_CUDA
-    - Linux_C_API_Packaging_TensorRT
     condition: succeeded()
     workspace:
       clean: all

tools/ci_build/github/azure-pipelines/stages/nuget-win-cuda-packaging-stage.yml

@@ -34,6 +34,9 @@ parameters:
   displayName: Specific Artifact's BuildId
   type: string
   default: '0'
+
+- name: buildJava
+  type: boolean
 
 stages:
 # Windows CUDA without TensorRT Packaging
@@ -49,7 +52,7 @@ stages:
     CudaVersion: ${{ parameters.CudaVersion }}
     buildparameter: --use_cuda --cuda_home=${{ parameters.win_cuda_home }} --enable_onnx_tests --enable_wcos --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=60;61;70;75;80"
     runTests: ${{ parameters.RunOnnxRuntimeTests }}
-    buildJava: true
+    buildJava: ${{ parameters.buildJava }}
     java_artifact_id: onnxruntime_gpu
     UseIncreasedTimeoutForTests: ${{ parameters.UseIncreasedTimeoutForTests }}
     SpecificArtifact: ${{ parameters.SpecificArtifact }}
@@ -67,7 +70,7 @@ stages:
     packageName: x64-tensorrt
     buildparameter: --use_tensorrt --tensorrt_home=${{ parameters.win_trt_home }} --cuda_home=${{ parameters.win_cuda_home }}  --enable_onnx_tests --enable_wcos --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=60;61;70;75;80"
     runTests: ${{ parameters.RunOnnxRuntimeTests }}
-    buildJava: true
+    buildJava: ${{ parameters.buildJava }}
     java_artifact_id: onnxruntime_gpu
     UseIncreasedTimeoutForTests: ${{ parameters.UseIncreasedTimeoutForTests }}
     SpecificArtifact: ${{ parameters.SpecificArtifact }}

tools/ci_build/github/azure-pipelines/templates/c-api-cpu.yml

@@ -511,7 +511,7 @@ stages:
   - Windows_CI_GPU_DML_Dev
   - Windows_CI_GPU_DML_Dev_arm64
   - Linux_C_API_Packaging_CPU
-  - Linux_C_API_Packaging_Combined_CUDA
+  - Linux_C_API_Packaging_GPU
   - MacOS_C_API_Package_Publish
   condition: succeeded()
   jobs:

tools/ci_build/github/azure-pipelines/templates/win-ci.yml

@@ -307,9 +307,11 @@ stages:
       - template: component-governance-component-detection-steps.yml
         parameters :
           condition : 'succeeded'
-  - ${{ if contains(parameters.ort_build_pool_name, 'GPU') }}:
+- ${{ if contains(parameters.ort_build_pool_name, 'GPU') }}:
+  - stage: Windows_Packaging_${{ parameters.stage_name_suffix }}_Testing
+    dependsOn: Windows_Packaging_${{ parameters.stage_name_suffix }}
+    jobs:
     - job: Windows_Packaging_${{ parameters.stage_name_suffix }}_Testing
-      dependsOn: Windows_Packaging_${{ parameters.stage_name_suffix }}
       workspace:
         clean: all
       pool: ${{ parameters.ort_build_pool_name }}

tools/ci_build/github/windows/extract_nuget_files_gpu.ps1

@@ -29,7 +29,13 @@ Foreach-Object {
 New-Item -Path $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\_deps\protobuf-build\RelWithDebInfo -ItemType directory
 
 Copy-Item -Path $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\RelWithDebInfo\nuget-artifacts\onnxruntime-win-x64-cuda-*\lib\* -Destination $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\RelWithDebInfo
-Copy-Item -Path $Env:BUILD_BINARIESDIRECTORY\extra-artifact\protoc.exe $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\_deps\protobuf-build\RelWithDebInfo
+
+$protocInstallDir = "$Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\_deps\protobuf-build"
+dotnet new console
+dotnet add package Google.Protobuf.Tools --version 3.21.12 --package-directory $protocInstallDir
+$protocDir = Get-ChildItem -Path $protocInstallDir -Recurse -Filter "protoc.exe" | Select-Object -ExpandProperty DirectoryName -First 1
+Write-Output $protocDir
+Copy-Item -Path $protocDir -Destination $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\_deps\protobuf-build\RelWithDebInfo
 
 $ort_dirs = Get-ChildItem -Path $Env:BUILD_BINARIESDIRECTORY\RelWithDebInfo\RelWithDebInfo\nuget-artifacts\onnxruntime-* -Directory
 foreach ($ort_dir in $ort_dirs)