Add python 3.13 support (microsoft#22380)

1. Add python 3.13 to our python packaging pipelines
2. Because numpy 2.0.0 doesn't support thread free python, this PR also
upgrades numpy to the latest
3. Delete some unused files.

Author: Changming Sun

tools/ci_build/github/android/build_aar_and_copy_artifacts.sh

@@ -6,7 +6,7 @@
 
 set -e
 set -x
-export PATH=/opt/python/cp38-cp38/bin:$PATH
+export PATH=/opt/python/cp312-cp312/bin:$PATH
 
 ls /build
 ls /build/deps

tools/ci_build/github/azure-pipelines/android-arm64-v8a-QNN-crosscompile-ci-pipeline.yml

@@ -63,9 +63,6 @@ jobs:
 
   - script: |
       set -e -x
-      rm -rf /tmp/scripts
-      cp -r tools/ci_build/github/linux/docker/inference/x86_64/python/cpu/scripts /tmp
-      /tmp/scripts/install_protobuf.sh -p $(Build.BinariesDirectory)/installed -d cmake/deps.txt
       python3 tools/ci_build/build.py \
         --config Release \
         --android \
@@ -78,7 +75,7 @@ jobs:
         --use_qnn \
         --qnn_home $(QnnSDKRootDir) \
         --cmake_generator=Ninja \
-        --skip_tests --path_to_protoc_exe $(Build.BinariesDirectory)/installed/bin/protoc
+        --skip_tests
     displayName: Build QNN EP
 
   - script: |

tools/ci_build/github/azure-pipelines/bigmodels-ci-pipeline.yml

@@ -42,7 +42,7 @@ parameters:
 variables:
   - template: templates/common-variables.yml
   - name: docker_base_image
-    value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241008.1
+    value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241010.2
   - name: linux_trt_version
     value: 10.3.0.26-1.cuda11.8
   - name: Repository

tools/ci_build/github/azure-pipelines/linux-ci-pipeline.yml

@@ -82,22 +82,20 @@ stages:
                     onnxruntimecpubuildcentos8x64 \
                     /bin/bash -c '
                       set -ex; \
-                      python3.9 /onnxruntime_src/tools/ci_build/build.py \
+                      python3.12 /onnxruntime_src/tools/ci_build/build.py \
                         --build_dir /build --cmake_generator 'Ninja' \
                         --config Debug \
                         --skip_submodule_sync \
                         --build_shared_lib \
                         --parallel --use_binskim_compliant_compile_flags \
-                        --build_csharp \
                         --enable_onnx_tests --enable_address_sanitizer \
                         --update --build;
-                      LD_PRELOAD=/usr/lib64/libasan.so.6 python3.9 /onnxruntime_src/tools/ci_build/build.py \
+                      python3.12 /onnxruntime_src/tools/ci_build/build.py \
                         --build_dir /build --cmake_generator 'Ninja' \
                         --config Debug \
                         --skip_submodule_sync \
                         --build_shared_lib \
                         --parallel --use_binskim_compliant_compile_flags \
-                        --build_csharp \
                         --enable_onnx_tests --enable_address_sanitizer \
                         --test;
                       '

tools/ci_build/github/azure-pipelines/linux-cpu-aten-pipeline.yml

@@ -49,9 +49,9 @@ parameters:
 variables:
   - name: docker_base_image
     ${{ if eq(parameters.CudaVersion, '11.8') }}:
-      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241008.1
+      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241010.2
     ${{ if eq(parameters.CudaVersion, '12.2') }}:
-      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20241008.1
+      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20241010.2
 
   - name: Repository
     ${{ if eq(parameters.CudaVersion, '11.8') }}:
@@ -175,13 +175,14 @@ stages:
             -e NVIDIA_TF32_OVERRIDE=0 \
             $(Repository) \
             /bin/bash -c '
+              set -e
               nvidia-smi; \
               /sbin/ldconfig -N -v $(sed "s/:/ /" <<< $LD_LIBRARY_PATH) 2>/dev/null | grep -E "libcudart.so|libcudnn.so|libnvinfer.so"; \
               cat /usr/local/cuda/include/cuda.h | grep -m1 CUDA_VERSION; \
               cat /usr/include/cudnn_version.h | grep CUDNN_MAJOR -m1 -A 2; \
-              ln -s /opt/python/cp38-cp38/bin/python3 /tmp/python3; \
-              /tmp/python3 -m pip install /build/Release/dist/*.whl; \
-              /tmp/python3 -u -c "from onnxruntime.capi._pybind_state import (OrtDevice as C_OrtDevice) ; \
+              export PATH=/opt/python/cp312-cp312/bin:$PATH; \
+              python3 -m pip install /build/Release/dist/*.whl; \
+              python3 -u -c "from onnxruntime.capi._pybind_state import (OrtDevice as C_OrtDevice) ; \
                         ort_device = C_OrtDevice(C_OrtDevice.cuda(), C_OrtDevice.default_memory(), 0); \
                         print(ort_device); print(ort_device.device_type(), C_OrtDevice.cuda()); \
                         assert(ort_device.device_type()==1); assert(C_OrtDevice.cuda()==1);" \
@@ -204,13 +205,13 @@ stages:
             /bin/bash -c '
               set -ex; \
               cp /onnxruntime_src/tools/ci_build/github/linux/docker/scripts/manylinux/requirements.txt /tmp/requirements.txt; \
-              ln -s /opt/python/cp38-cp38/bin/python3 /tmp/python3; \
-              /tmp/python3 -m pip install -r /tmp/requirements.txt; \
-              /tmp/python3 -m pip install /build/Release/dist/*.whl; \
+              export PATH=/opt/python/cp312-cp312/bin:$PATH; \
+              python3 -m pip install -r /tmp/requirements.txt; \
+              python3 -m pip install /build/Release/dist/*.whl; \
               cd /build/Release && xargs -a /build/Release/perms.txt chmod a+x; \
               cd /onnxruntime_src/java && /onnxruntime_src/java/gradlew cmakeCheck -DcmakeBuildDir=/build/Release -DUSE_CUDA=1; \
               cd /tmp; \
-              /tmp/python3 /onnxruntime_src/tools/ci_build/build.py \
+              python3 /onnxruntime_src/tools/ci_build/build.py \
                 --build_dir /build --config Release --test --skip_submodule_sync --build_shared_lib --parallel --use_binskim_compliant_compile_flags --build_wheel --enable_onnx_tests \
                 --enable_transformers_tool_test --use_cuda --cuda_version=${{parameters.CudaVersion}} --cuda_home=/usr/local/cuda --cudnn_home=/usr/local/cuda \
                 --enable_pybind --build_java --ctest_path "" ; \

tools/ci_build/github/azure-pipelines/linux-cpu-eager-pipeline.yml

@@ -39,9 +39,9 @@ parameters:
 variables:
   - name: docker_base_image
     ${{ if eq(parameters.CudaVersion, '11.8') }}:
-      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241008.1
+      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20241010.2
     ${{ if eq(parameters.CudaVersion, '12.2') }}:
-      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20241008.1
+      value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20241010.2
   - name: linux_trt_version
     ${{ if eq(parameters.CudaVersion, '11.8') }}:
       value: 10.4.0.26-1.cuda11.8

tools/ci_build/github/azure-pipelines/linux-gpu-ci-pipeline.yml

@@ -85,7 +85,7 @@ stages:
 
       - task: UsePythonVersion@0
         inputs:
-          versionSpec: '3.8'
+          versionSpec: '3.12'
           addToPath: true
           architecture: ${{ parameters.BuildArch }}