diff --git a/cgmanifests/generate_cgmanifest.py b/cgmanifests/generate_cgmanifest.py
index a85516011a402..bd2c85762115e 100644
--- a/cgmanifests/generate_cgmanifest.py
+++ b/cgmanifests/generate_cgmanifest.py
@@ -59,6 +59,7 @@ def add_github_dep(name, parsed_url):
         if dep not in git_deps:
             git_deps[dep] = name
     else:
+        # TODO: support urls like: https://github.com/onnx/onnx-tensorrt/archive/refs/tags/release/7.1.zip
         if len(segments) == 5:
             tag = PurePosixPath(segments[4]).stem
             if tag.endswith(".tar"):
@@ -72,6 +73,7 @@ def add_github_dep(name, parsed_url):
             return
         # Make a REST call to convert to tag to a git commit
         url = "https://api.github.com/repos/%s/%s/git/refs/tags/%s" % (org_name, repo_name, tag)
+        print("requesting %s ..." % url)
         res = requests.get(url, auth=(args.username, args.token))
         response_json = res.json()
         tag_object = response_json["object"]
@@ -148,7 +150,6 @@ def normalize_path_separators(path):
         "submodule",
         "foreach",
         "--quiet",
-        "--recursive",
         "'{}' '{}' $toplevel/$sm_path".format(
             normalize_path_separators(sys.executable),
             normalize_path_separators(os.path.join(SCRIPT_DIR, "print_submodule_info.py")),
diff --git a/cmake/deps.txt b/cmake/deps.txt
index 1cc6ed26dcd95..ee487872f9246 100644
--- a/cmake/deps.txt
+++ b/cmake/deps.txt
@@ -24,8 +24,8 @@ microsoft_wil;https://github.com/microsoft/wil/archive/5f4caba4e7a9017816e47becd
 mimalloc;https://github.com/microsoft/mimalloc/archive/refs/tags/v2.0.3.zip;e4f37b93b2da78a5816c2495603a4188d316214b
 mp11;https://github.com/boostorg/mp11/archive/refs/tags/boost-1.79.0.zip;c8f04e378535ededbe5af52c8f969d2dedbe73d5
 onnx;https://github.com/onnx/onnx/archive/5a5f8a5935762397aa68429b5493084ff970f774.zip;edc8e1338c02f3ab222f3d803a24e17608c13895
-#Branch name: 8.4-GA
-onnx_tensorrt;https://github.com/onnx/onnx-tensorrt/archive/87c7a70688fd98fb355b8976f41425b40e4fe52f.zip;b97d112d9d6efa180c9b94e05268f2ff3294a534
+#use the commit where it's several commits after 8.5-GA branch (https://github.com/onnx/onnx-tensorrt/commit/369d6676423c2a6dbf4a5665c4b5010240d99d3c)
+onnx_tensorrt;https://github.com/onnx/onnx-tensorrt/archive/369d6676423c2a6dbf4a5665c4b5010240d99d3c.zip;62119892edfb78689061790140c439b111491275
 protobuf;https://github.com/protocolbuffers/protobuf/archive/refs/tags/v3.18.3.zip;b95bf7e9de9c2249b6c1f2ca556ace49999e90bd
 psimd;https://github.com/Maratyszcza/psimd/archive/072586a71b55b7f8c584153d223e95687148a900.zip;1f5454b01f06f9656b77e4a5e2e31d7422487013
 pthreadpool;https://github.com/Maratyszcza/pthreadpool/archive/1787867f6183f056420e532eec640cba25efafea.zip;e43e80781560c5ab404a4da20f34d846f5f5d101
@@ -33,4 +33,4 @@ pybind11;https://github.com/pybind/pybind11/archive/refs/tags/v2.10.1.zip;769b6a
 pytorch_cpuinfo;https://github.com/pytorch/cpuinfo/archive/5916273f79a21551890fd3d56fc5375a78d1598d.zip;2be4d2ae321fada97cb39eaf4eeba5f8c85597cf
 re2;https://github.com/google/re2/archive/refs/tags/2022-06-01.zip;aa77313b76e91b531ee7f3e45f004c6a502a5374
 safeint;https://github.com/dcleblanc/SafeInt/archive/ff15c6ada150a5018c5ef2172401cb4529eac9c0.zip;913a4046e5274d329af2806cb53194f617d8c0ab
-tensorboard;https://github.com/tensorflow/tensorboard/archive/373eb09e4c5d2b3cc2493f0949dc4be6b6a45e81.zip;67b833913605a4f3f499894ab11528a702c2b381
\ No newline at end of file
+tensorboard;https://github.com/tensorflow/tensorboard/archive/373eb09e4c5d2b3cc2493f0949dc4be6b6a45e81.zip;67b833913605a4f3f499894ab11528a702c2b381
diff --git a/cmake/onnxruntime_providers.cmake b/cmake/onnxruntime_providers.cmake
index 99c985f9ee98e..b80e75e8c5342 100644
--- a/cmake/onnxruntime_providers.cmake
+++ b/cmake/onnxruntime_providers.cmake
@@ -635,10 +635,12 @@ if (onnxruntime_USE_TENSORRT)
     FetchContent_Declare(
       onnx_tensorrt
       URL ${DEP_URL_onnx_tensorrt}
-	    URL_HASH SHA1=${DEP_SHA1_onnx_tensorrt}
+      URL_HASH SHA1=${DEP_SHA1_onnx_tensorrt}
     )
-    FetchContent_MakeAvailable(onnx_tensorrt)
-	include_directories(${onnx_tensorrt_SOURCE_DIR})
+    # The onnx_tensorrt repo contains a test program, getSupportedAPITest, which doesn't support Windows. It uses
+    # unistd.h. So we must exclude it from our build. onnxruntime_fetchcontent_makeavailable is for the purpose.
+    onnxruntime_fetchcontent_makeavailable(onnx_tensorrt)
+    include_directories(${onnx_tensorrt_SOURCE_DIR})
     set(CMAKE_CXX_FLAGS ${OLD_CMAKE_CXX_FLAGS})
     if ( CMAKE_COMPILER_IS_GNUCC )
       set(CMAKE_CXX_FLAGS  "${CMAKE_CXX_FLAGS} -Wno-unused-parameter")
@@ -1479,7 +1481,7 @@ if (onnxruntime_USE_TVM)
     # wd4100: identifier' : unreferenced formal parameter
     # wd4127: conditional expression is constant
     # wd4244: conversion from 'int' to 'char', possible loss of data
-	# TODO: 4244 should not be disabled
+    # TODO: 4244 should not be disabled
     target_compile_options(onnxruntime_providers_tvm PRIVATE "/wd4100" "/wd4127" "/wd4244")
   else()
     target_compile_options(onnxruntime_providers_tvm PRIVATE "-Wno-error=type-limits")
diff --git a/cmake/onnxruntime_unittests.cmake b/cmake/onnxruntime_unittests.cmake
index 43e79a2dd6e88..aa224d753a997 100644
--- a/cmake/onnxruntime_unittests.cmake
+++ b/cmake/onnxruntime_unittests.cmake
@@ -708,13 +708,21 @@ endif()
 
 set(test_all_args)
 if (onnxruntime_USE_TENSORRT)
-    # TRT EP CI takes much longer time when updating to TRT 8.2
-    # So, we only run trt ep and exclude other eps to reduce CI test time.
-    #
-    # The test names of model tests were using sequential number in the past.
-    # This PR https://github.com/microsoft/onnxruntime/pull/10220 (Please see ExpandModelName function in model_tests.cc for more details)
-    # made test name contain the "ep" and "model path" information, so we can easily filter the tests using cuda ep or other ep with *cpu__* or *xxx__*.
-    list(APPEND test_all_args "--gtest_filter=-*cpu__*:*cuda__*" )
+    if (onnxruntime_SKIP_AND_PERFORM_FILTERED_TENSORRT_TESTS)
+       # TRT EP package pipelines takes much longer time to run tests with TRT 8.5. We can't use placeholder to reduce testing time due to application test deadlock. 
+       # Therefore we only run filtered TRT EP tests.
+      list(APPEND test_all_args "--gtest_filter=*tensorrt_*:*TensorrtExecutionProviderTest*" )
+      #list(APPEND test_all_args "--gtest_filter=-*cpu_*:*cuda_*:*ContribOpTest*:*QuantGemmTest*:*QLinearConvTest*:*MurmurHash3OpTest*:*PadOpTest*:*QLinearConvTest*" )
+    else()
+      # TRT EP CI takes much longer time when updating to TRT 8.2
+      # So, we only run trt ep and exclude other eps to reduce CI test time.
+      #
+      # The test names of model tests were using sequential number in the past.
+      # This PR https://github.com/microsoft/onnxruntime/pull/10220 (Please see ExpandModelName function in model_tests.cc for more details)
+      # made test name contain the "ep" and "model path" information, so we can easily filter the tests using cuda ep or other ep with *cpu_* or *xxx_*.
+      list(APPEND test_all_args "--gtest_filter=-*cpu_*:*cuda_*" )
+    endif()
+
 endif ()
 
 AddTest(
diff --git a/onnxruntime/test/contrib_ops/quantize_ops_test.cc b/onnxruntime/test/contrib_ops/quantize_ops_test.cc
index af0b1a2dbbaf7..5b01544d4bb2b 100644
--- a/onnxruntime/test/contrib_ops/quantize_ops_test.cc
+++ b/onnxruntime/test/contrib_ops/quantize_ops_test.cc
@@ -203,7 +203,7 @@ void TestQuantizeLinearPerTensorFloatUint8(bool use_initializer_except_x) {
                            255, 0,
                            255, 0,
                            255, 0});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 TEST(QuantizeLinearContribOpTest, QuantizeLinear_per_tensor_float_uint8) {
@@ -270,7 +270,7 @@ TEST(QuantizeLinearContribOpTest, QuantizeLinear_per_tensor_half_uint8) {
                            255, 0,
                            255, 0,
                            255, 0});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 TEST(QuantizeLinearContribOpTest, QuantizeLinear_per_tensor_half_int8) {
@@ -317,7 +317,7 @@ TEST(QuantizeLinearContribOpTest, QuantizeLinear_per_channel) {
                           {0, 2, 3, 255,
                            0, 1, 2, 255,
                            0, 0, 1, 250});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 // quantize with broadcasting and negative axis (-2 resolves to axis 0)
@@ -335,7 +335,7 @@ TEST(QuantizeLinearContribOpTest, QuantizeLinear_per_channel_negative_axis) {
                           {0, 2, 3, 255,
                            0, 1, 2, 255,
                            0, 0, 1, 250});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 }  // namespace test
 }  // namespace onnxruntime
diff --git a/onnxruntime/test/contrib_ops/tensor_op_test.cc b/onnxruntime/test/contrib_ops/tensor_op_test.cc
index 6ed563909560c..96a2a39748c56 100644
--- a/onnxruntime/test/contrib_ops/tensor_op_test.cc
+++ b/onnxruntime/test/contrib_ops/tensor_op_test.cc
@@ -120,7 +120,7 @@ void MeanVarianceNormalizationAcrossChannels(bool across_channels, bool normaliz
   test.AddAttribute("normalize_variance", normalize_variance ? one : zero);
   test.AddInput<float>("input", {N, C, H, W}, X);
   test.AddOutput<float>("output", {N, C, H, W}, result);
-  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kOpenVINOExecutionProvider}); //OpenVINO doesn't support MVN operator below opset 9
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kOpenVINOExecutionProvider, kTensorrtExecutionProvider}); //OpenVINO doesn't support MVN operator below opset 9. TensorRT doesn't support opset 8 of MVN operator.
 }
 
 void MeanVarianceNormalizationPerChannel(bool across_channels, bool normalize_variance) {
@@ -187,7 +187,7 @@ void MeanVarianceNormalizationPerChannel(bool across_channels, bool normalize_va
   test.AddAttribute("normalize_variance", normalize_variance ? one : zero);
   test.AddInput<float>("input", {N, C, H, W}, X);
   test.AddOutput<float>("output", {N, C, H, W}, result);
-  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kOpenVINOExecutionProvider}); //OpenVINO doesn't support MVN operator below opset 9
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kOpenVINOExecutionProvider, kTensorrtExecutionProvider}); //OpenVINO doesn't support MVN operator below opset 9. TensorRT doesn't support opset 8 of MVN operator.
 }
 
 TEST(MVNContribOpTest, MeanVarianceNormalizationCPUTest_Version1_TO_8) {
diff --git a/onnxruntime/test/providers/cpu/generator/random_test.cc b/onnxruntime/test/providers/cpu/generator/random_test.cc
index e3101297b25f3..85ad297cab814 100644
--- a/onnxruntime/test/providers/cpu/generator/random_test.cc
+++ b/onnxruntime/test/providers/cpu/generator/random_test.cc
@@ -71,7 +71,8 @@ void RunRandomNormalLike3DFloat(bool infer_dtype = false) {
 
   test.AddOutput<float>("Y", dims, expected_output);
 
-  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kCudaExecutionProvider, kRocmExecutionProvider});
+  // TensorRT does not support manual seed overrides and there will be result mismatch
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kCudaExecutionProvider, kRocmExecutionProvider, kTensorrtExecutionProvider});
 }
 
 TEST(Random, RandomNormalLike3DDouble) {
diff --git a/onnxruntime/test/providers/cpu/math/element_wise_ops_test.cc b/onnxruntime/test/providers/cpu/math/element_wise_ops_test.cc
index a005d3e823ea6..2000c5cededbc 100644
--- a/onnxruntime/test/providers/cpu/math/element_wise_ops_test.cc
+++ b/onnxruntime/test/providers/cpu/math/element_wise_ops_test.cc
@@ -2883,7 +2883,7 @@ TEST(ModOpTest, Int8_mixed_sign) {
   test.AddInput<int8_t>("Y", {6}, {2, -3, 8, -2, 3, 5});
   test.AddOutput<int8_t>("Z", {6}, {0, -2, 5, 0, 2, 3});
 
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); // For TensorRT running in these in INT8 quantization scales are needed, so skip it now
 }
 
 TEST(ModOpTest, Int8_mixed_sign_fmod) {
@@ -2894,7 +2894,7 @@ TEST(ModOpTest, Int8_mixed_sign_fmod) {
   test.AddInput<int8_t>("Y", {6}, {2, -3, 8, -2, 3, 5});
   test.AddOutput<int8_t>("Z", {6}, {0, 1, 5, 0, -1, 3});
 
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); // For TensorRT running in these in INT8 quantization scales are needed, so skip it now
 }
 
 TEST(ModOpTest, UInt8_mod) {
diff --git a/onnxruntime/test/providers/cpu/nn/shrink_test.cc b/onnxruntime/test/providers/cpu/nn/shrink_test.cc
index 709f6a16d2277..4a1e18db21d47 100644
--- a/onnxruntime/test/providers/cpu/nn/shrink_test.cc
+++ b/onnxruntime/test/providers/cpu/nn/shrink_test.cc
@@ -97,7 +97,7 @@ const std::vector<MLFloat16> ConvertFloatToMLFloat16(const std::vector<float>& f
 
 TEST(MathOpTest, ShrinkInt8Type) {
   const auto& test_cases = GenerateSignedTestCases<int8_t>();
-  RunShrinkTest<int8_t>(test_cases);
+  RunShrinkTest<int8_t>(test_cases, {kTensorrtExecutionProvider}); // For TensorRT running in these in INT8 quantization scales are needed, so skip it now
 }
 
 TEST(MathOpTest, ShrinkUint8Type) {
diff --git a/onnxruntime/test/providers/cpu/tensor/quantize_linear_test.cc b/onnxruntime/test/providers/cpu/tensor/quantize_linear_test.cc
index 63b911805c36c..0b286f628582f 100644
--- a/onnxruntime/test/providers/cpu/tensor/quantize_linear_test.cc
+++ b/onnxruntime/test/providers/cpu/tensor/quantize_linear_test.cc
@@ -228,7 +228,7 @@ TEST(QuantizeLinearOpTest, Uint8) {
   test.AddInput<float>("y_scale", {}, {2.0f});
   test.AddInput<uint8_t>("y_zero_point", {}, {128});
   test.AddOutput<uint8_t>("y", dims, {128, 129, 130, 255, 1, 0});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 // quantize with scalar zero point and scale
@@ -296,7 +296,7 @@ TEST(QuantizeLinearOpTest, 2D) {
                           {0, 0, 1, 250,
                            0, 0, 1, 250,
                            0, 0, 1, 250});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 // quantize with scalar data
@@ -306,7 +306,7 @@ TEST(QuantizeLinearOpTest, Scalar) {
   test.AddInput<float>("y_scale", {}, {2.0f});
   test.AddInput<uint8_t>("y_zero_point", {}, {128});
   test.AddOutput<uint8_t>("y", {}, {130});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 // quantize with scalar data
@@ -315,7 +315,7 @@ TEST(QuantizeLinearOpTest, DISABLED_QuantizeLinear_Without_Zero_Point) {
   test.AddInput<float>("x", {}, {3});
   test.AddInput<float>("y_scale", {}, {2.0f});
   test.AddOutput<uint8_t>("y", {}, {2});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 TEST(QuantizeLinearOpTest, Per_Channel_Axis_Default) {
@@ -331,7 +331,7 @@ TEST(QuantizeLinearOpTest, Per_Channel_Axis_Default) {
                           {64, 101, 127, 177,
                            65, 100, 128, 182,
                            66, 102, 128, 187});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 TEST(QuantizeLinearOpTest, Per_Channel_Axis_0) {
@@ -348,7 +348,7 @@ TEST(QuantizeLinearOpTest, Per_Channel_Axis_0) {
                           {0, 2, 3, 255,
                            0, 1, 2, 255,
                            0, 0, 1, 250});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 // quantize with per-channel and negative axis (-2 resolves to axis 0)
@@ -366,7 +366,7 @@ TEST(QuantizeLinearOpTest, Per_Channel_Axis_neg) {
                           {0, 2, 3, 255,
                            0, 1, 2, 255,
                            0, 0, 1, 250});
-  test.Run();
+  test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider}); //TensorRT doesn't support support UINT8 for quantization
 }
 
 }  // namespace test
diff --git a/tools/ci_build/build.py b/tools/ci_build/build.py
index a6e6b4cdc715b..62901a876f2f1 100644
--- a/tools/ci_build/build.py
+++ b/tools/ci_build/build.py
@@ -491,6 +491,12 @@ def convert_arg_line_to_args(self, arg_line):
         "--tensorrt_placeholder_builder", action="store_true", help="Instantiate Placeholder TensorRT Builder"
     )
     parser.add_argument("--tensorrt_home", help="Path to TensorRT installation dir")
+    parser.add_argument("--test_all_timeout", default="10800", help="Set timeout for onnxruntime_test_all")
+    parser.add_argument(
+        "--skip_and_perform_filtered_tests",
+        action="store_true",
+        help="Skip time-consuming and only perform filtered tests for TensorRT EP",
+    )
     parser.add_argument("--use_migraphx", action="store_true", help="Build with MIGraphX")
     parser.add_argument("--migraphx_home", help="Path to MIGraphX installation dir")
     parser.add_argument("--use_full_protobuf", action="store_true", help="Use the full protobuf library")
@@ -876,6 +882,8 @@ def generate_build_tree(
         "-Donnxruntime_ENABLE_MICROSOFT_INTERNAL=" + ("ON" if args.enable_msinternal else "OFF"),
         "-Donnxruntime_USE_VITISAI=" + ("ON" if args.use_vitisai else "OFF"),
         "-Donnxruntime_USE_TENSORRT=" + ("ON" if args.use_tensorrt else "OFF"),
+        "-Donnxruntime_SKIP_AND_PERFORM_FILTERED_TENSORRT_TESTS="
+        + ("ON" if args.test_all_timeout == "10800" else "OFF"),
         "-Donnxruntime_USE_TENSORRT_BUILTIN_PARSER=" + ("ON" if args.use_tensorrt_builtin_parser else "OFF"),
         "-Donnxruntime_TENSORRT_PLACEHOLDER_BUILDER=" + ("ON" if args.tensorrt_placeholder_builder else "OFF"),
         # set vars for TVM
@@ -1881,7 +1889,7 @@ def run_onnxruntime_tests(args, source_dir, ctest_path, build_dir, configs):
                     run_subprocess([os.path.join(cwd, exe)], cwd=cwd, dll_path=dll_path)
 
         else:
-            ctest_cmd = [ctest_path, "--build-config", config, "--verbose", "--timeout", "10800"]
+            ctest_cmd = [ctest_path, "--build-config", config, "--verbose", "--timeout", args.test_all_timeout]
             run_subprocess(ctest_cmd, cwd=cwd, dll_path=dll_path)
 
         if args.enable_pybind:
diff --git a/tools/ci_build/github/azure-pipelines/c-api-noopenmp-packaging-pipelines.yml b/tools/ci_build/github/azure-pipelines/c-api-noopenmp-packaging-pipelines.yml
index e3f614e59ced7..b13bd25e61d47 100644
--- a/tools/ci_build/github/azure-pipelines/c-api-noopenmp-packaging-pipelines.yml
+++ b/tools/ci_build/github/azure-pipelines/c-api-noopenmp-packaging-pipelines.yml
@@ -4,6 +4,11 @@ parameters:
   type: boolean
   default: true
 
+- name: UseIncreasedTimeoutForTests 
+  displayName: Increase timeout for tests? Set it to false if you are doing an Onnx Runtime release.
+  type: boolean
+  default: false
+
 - name: DoCompliance
   displayName: Run Compliance Tasks?
   type: boolean
@@ -116,10 +121,11 @@ jobs:
     buildArch: x64
     msbuildPlatform: x64
     packageName: x64-tensorrt
-    buildparameter: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --enable_onnx_tests --enable_wcos --build_java --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;52;60;61;70;75;80"
+    buildparameter: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --enable_onnx_tests --enable_wcos --build_java --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;52;60;61;70;75;80"
     runTests: ${{ parameters.RunOnnxRuntimeTests }}
     buildJava: true
     java_artifact_id: onnxruntime_gpu
+    UseIncreasedTimeoutForTests: ${{ parameters.UseIncreasedTimeoutForTests }}
 
 - job: Jar_Packaging_GPU
   workspace:
@@ -294,11 +300,11 @@ jobs:
       Steps:
       - script: |
           tools/ci_build/get_docker_image.py \
-            --dockerfile tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_4 \
+            --dockerfile tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5 \
             --context tools/ci_build/github/linux/docker \
             --docker-build-args "--network=host --build-arg POLICY=manylinux2014 --build-arg PLATFORM=x86_64 --build-arg DEVTOOLSET_ROOTPATH=/opt/rh/devtoolset-11/root --build-arg PREPEND_PATH=/opt/rh/devtoolset-11/root/usr/bin: --build-arg LD_LIBRARY_PATH_ARG=/opt/rh/devtoolset-11/root/usr/lib64:/opt/rh/devtoolset-11/root/usr/lib:/opt/rh/devtoolset-11/root/usr/lib64/dyninst:/opt/rh/devtoolset-11/root/usr/lib/dyninst:/usr/local/lib64 --build-arg BUILD_UID=$( id -u )" \
             --container-registry onnxruntimebuildcache \
-            --repository onnxruntimecuda116xtrt84build
+            --repository onnxruntimecuda116xtrt85build
         displayName: "Getonnxruntimecuda116xtrt84build image for tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_4"
         workingDirectory: $(Build.SourcesDirectory)/onnxruntime
       ContainerRegistry: onnxruntimebuildcache
@@ -351,7 +357,7 @@ jobs:
     inputs:
       script: |
         docker run --gpus all -e CC=/opt/rh/devtoolset-11/root/usr/bin/cc -e CXX=/opt/rh/devtoolset-11/root/usr/bin/c++ -e CFLAGS="-Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fstack-protector-strong -fstack-clash-protection -fcf-protection -O3 -Wl,--strip-all" -e CXXFLAGS="-Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fstack-protector-strong -fstack-clash-protection -fcf-protection -O3 -Wl,--strip-all" -e NVIDIA_VISIBLE_DEVICES=all --rm --volume $(Build.SourcesDirectory):/src_dir \
-        --volume $(Build.ArtifactStagingDirectory):/artifact_src -e NIGHTLY_BUILD onnxruntimecuda116xtrt84build \
+        --volume $(Build.ArtifactStagingDirectory):/artifact_src -e NIGHTLY_BUILD onnxruntimecuda116xtrt85build \
         /src_dir/onnxruntime-inference-examples/c_cxx/squeezenet/run_capi_application.sh -o /src_dir/onnxruntime -p /artifact_src/onnxruntime-linux-x64-gpu-$(OnnxRuntimeVersion).tgz -w /src_dir/onnxruntime-inference-examples/c_cxx/squeezenet
       workingDirectory: '$(Build.ArtifactStagingDirectory)'
 
diff --git a/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-ci-pipeline.yml b/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-ci-pipeline.yml
index 1c51a82d164b9..5f13510515d6e 100644
--- a/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-ci-pipeline.yml
+++ b/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-ci-pipeline.yml
@@ -22,10 +22,10 @@ jobs:
 
   - template: templates/get-docker-image-steps.yml
     parameters:
-      Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_4
+      Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5
       Context: tools/ci_build/github/linux/docker
       DockerBuildArgs: "--network=host --build-arg POLICY=manylinux2014 --build-arg PLATFORM=x86_64  --build-arg DEVTOOLSET_ROOTPATH=/opt/rh/devtoolset-11/root --build-arg PREPEND_PATH=/opt/rh/devtoolset-11/root/usr/bin: --build-arg LD_LIBRARY_PATH_ARG=/opt/rh/devtoolset-11/root/usr/lib64:/opt/rh/devtoolset-11/root/usr/lib:/opt/rh/devtoolset-11/root/usr/lib64/dyninst:/opt/rh/devtoolset-11/root/usr/lib/dyninst:/usr/local/lib64 --build-arg BUILD_UID=$( id -u )"
-      Repository: onnxruntimetensorrt84gpubuild
+      Repository: onnxruntimetensorrt85gpubuild
 
   - task: CmdLine@2
     inputs:
@@ -39,7 +39,7 @@ jobs:
             -e ALLOW_RELEASED_ONNX_OPSET_ONLY=0 \
             -e NIGHTLY_BUILD \
             -e BUILD_BUILDNUMBER \
-            onnxruntimetensorrt84gpubuild \
+            onnxruntimetensorrt85gpubuild \
             /opt/python/cp38-cp38/bin/python3 /onnxruntime_src/tools/ci_build/build.py \
               --build_dir /build --cmake_generator Ninja \
               --config Release \
diff --git a/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-packaging-pipeline.yml b/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-packaging-pipeline.yml
index f78e741584f6f..b6844710a4806 100644
--- a/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-packaging-pipeline.yml
+++ b/tools/ci_build/github/azure-pipelines/linux-gpu-tensorrt-packaging-pipeline.yml
@@ -30,7 +30,7 @@ jobs:
       submodules: recursive
     - template: templates/get-docker-image-steps.yml
       parameters:
-        Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_4
+        Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5
         Context: tools/ci_build/github/linux/docker
         DockerBuildArgs: "--network=host --build-arg POLICY=manylinux2014 --build-arg PLATFORM=x86_64 --build-arg DEVTOOLSET_ROOTPATH=/opt/rh/devtoolset-11/root --build-arg PREPEND_PATH=/opt/rh/devtoolset-11/root/usr/bin: --build-arg LD_LIBRARY_PATH_ARG=/opt/rh/devtoolset-11/root/usr/lib64:/opt/rh/devtoolset-11/root/usr/lib:/opt/rh/devtoolset-11/root/usr/lib64/dyninst:/opt/rh/devtoolset-11/root/usr/lib/dyninst:/usr/local/lib64 --build-arg BUILD_UID=$( id -u )"
         Repository: onnxruntimecuda114xtrt82build
diff --git a/tools/ci_build/github/azure-pipelines/templates/download-deps.yml b/tools/ci_build/github/azure-pipelines/templates/download-deps.yml
index 03693edf7cf7e..5cbe400e9f455 100644
--- a/tools/ci_build/github/azure-pipelines/templates/download-deps.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/download-deps.yml
@@ -11,7 +11,7 @@ steps:
       packageType: upack
       feed: '/7424c8e4-5c62-490e-95c4-79446f31017c'
       definition: '517c4f6f-5437-4392-a70d-4f15ec5be2f0'
-      version: 1.0.11
+      version: 1.0.15
       downloadPath: $(Build.BinariesDirectory)/deps
 
 # The private ADO project
@@ -22,7 +22,7 @@ steps:
       packageType: upack
       feed: '/4c7631f5-24c0-4307-8822-1aa8f180c325'
       definition: 'fd9dd5ad-b73e-4678-890e-edcf680dbc1a'
-      version: 1.0.11
+      version: 1.0.15
       downloadPath: $(Build.BinariesDirectory)/deps
 
-# You can add more ADO accounts at here.
\ No newline at end of file
+# You can add more ADO accounts at here.
diff --git a/tools/ci_build/github/azure-pipelines/templates/py-packaging-selectable-stage.yml b/tools/ci_build/github/azure-pipelines/templates/py-packaging-selectable-stage.yml
index 4ff80f1fa7130..c631a58463658 100644
--- a/tools/ci_build/github/azure-pipelines/templates/py-packaging-selectable-stage.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/py-packaging-selectable-stage.yml
@@ -301,10 +301,10 @@ stages:
 
       - template: get-docker-image-steps.yml
         parameters:
-          Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_4
+          Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5
           Context: tools/ci_build/github/linux/docker
           DockerBuildArgs: "--network=host --build-arg POLICY=manylinux2014 --build-arg PLATFORM=x86_64 --build-arg DEVTOOLSET_ROOTPATH=/opt/rh/devtoolset-11/root --build-arg PREPEND_PATH=/opt/rh/devtoolset-11/root/usr/bin: --build-arg LD_LIBRARY_PATH_ARG=/opt/rh/devtoolset-11/root/usr/lib64:/opt/rh/devtoolset-11/root/usr/lib:/opt/rh/devtoolset-11/root/usr/lib64/dyninst:/opt/rh/devtoolset-11/root/usr/lib/dyninst:/usr/local/lib64 --build-arg BUILD_UID=$( id -u )"
-          Repository: onnxruntimecuda116xtrt84build
+          Repository: onnxruntimecuda116xtrt85build
 
       - task: CmdLine@2
         displayName: 'Build Python Wheel'
@@ -318,7 +318,7 @@ stages:
               --volume /data/models:/build/models:ro \
               --volume $HOME/.onnx:/home/onnxruntimedev/.onnx \
               -e BUILD_BUILDNUMBER \
-              onnxruntimecuda116xtrt84build \
+              onnxruntimecuda116xtrt85build \
                 $(PythonManylinuxDir)/bin/python3 /onnxruntime_src/tools/ci_build/build.py \
                   --build_dir /build --cmake_generator Ninja \
                   --config Release --update --build \
@@ -386,7 +386,7 @@ stages:
       variables:
         CUDA_VERSION: '11.6'
         buildArch: x64
-        EpBuildFlags: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=$(CUDA_VERSION) --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v$(CUDA_VERSION)" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
+        EpBuildFlags: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=$(CUDA_VERSION) --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v$(CUDA_VERSION)" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
         EnvSetupScript: setup_env_gpu.bat
         EP_NAME: gpu
       strategy:
diff --git a/tools/ci_build/github/azure-pipelines/templates/py-packaging-stage.yml b/tools/ci_build/github/azure-pipelines/templates/py-packaging-stage.yml
index 9f5004995ab5b..6fe1cc1abd8d4 100644
--- a/tools/ci_build/github/azure-pipelines/templates/py-packaging-stage.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/py-packaging-stage.yml
@@ -271,7 +271,7 @@ stages:
         parameters:
           MACHINE_POOL: 'onnxruntime-gpu-winbuild-t4'
           PYTHON_VERSION: '3.7'
-          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
+          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
           ENV_SETUP_SCRIPT: setup_env_gpu.bat
           EP_NAME: gpu
 
@@ -279,7 +279,7 @@ stages:
         parameters:
           MACHINE_POOL: 'onnxruntime-gpu-winbuild-t4'
           PYTHON_VERSION: '3.8'
-          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
+          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
           ENV_SETUP_SCRIPT: setup_env_gpu.bat
           EP_NAME: gpu
 
@@ -287,7 +287,7 @@ stages:
         parameters:
           MACHINE_POOL: 'onnxruntime-gpu-winbuild-t4'
           PYTHON_VERSION: '3.9'
-          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
+          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
           ENV_SETUP_SCRIPT: setup_env_gpu.bat
           EP_NAME: gpu
 
@@ -295,7 +295,7 @@ stages:
         parameters:
           MACHINE_POOL: 'onnxruntime-gpu-winbuild-t4'
           PYTHON_VERSION: '3.10'
-          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
+          EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6"  --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
           ENV_SETUP_SCRIPT: setup_env_gpu.bat
           EP_NAME: gpu
 
diff --git a/tools/ci_build/github/azure-pipelines/templates/win-ci.yml b/tools/ci_build/github/azure-pipelines/templates/win-ci.yml
index 59c0c53fd8e39..dad7d23f57b24 100644
--- a/tools/ci_build/github/azure-pipelines/templates/win-ci.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/win-ci.yml
@@ -52,12 +52,20 @@ parameters:
   type: string
   default: 'onnxruntime'
 
+- name: UseIncreasedTimeoutForTests 
+  displayName: Increase timeout for tests? Set it to false if you are doing an Onnx Runtime release.
+  type: boolean
+  default: false
+
 jobs:
 - job: Windows_Packaging_${{ parameters.job_name_suffix }}
   workspace:
     clean: all
   pool: ${{ parameters.ort_build_pool_name }}
-  timeoutInMinutes:  300  
+  ${{ if eq(parameters['UseIncreasedTimeoutForTests'], 'true') }}:
+    timeoutInMinutes: 1200 
+  ${{ else }}:
+    timeoutInMinutes: 300 
 
   steps:
     - checkout: self
@@ -106,7 +114,10 @@ jobs:
       displayName: 'Generate cmake config'
       inputs:
         scriptPath: '$(Build.SourcesDirectory)\tools\ci_build\build.py'
-        arguments: '--config RelWithDebInfo --enable_lto --disable_rtti --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --update --cmake_generator "Visual Studio 16 2019" --enable_onnx_tests $(TelemetryOption) ${{ parameters.buildparameter }}'
+        ${{ if eq(parameters['UseIncreasedTimeoutForTests'], 'true') }}:
+          arguments: '--config RelWithDebInfo --enable_lto --disable_rtti --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --update --cmake_generator "Visual Studio 16 2019" --enable_onnx_tests $(TelemetryOption) ${{ parameters.buildparameter }} --test_all_timeout 72000'
+        ${{ else }}:
+          arguments: '--config RelWithDebInfo --enable_lto --disable_rtti --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --update --cmake_generator "Visual Studio 16 2019" --enable_onnx_tests $(TelemetryOption) ${{ parameters.buildparameter }} '
         workingDirectory: '$(Build.BinariesDirectory)'
  
     - task: VSBuild@1
diff --git a/tools/ci_build/github/azure-pipelines/win-gpu-tensorrt-ci-pipeline.yml b/tools/ci_build/github/azure-pipelines/win-gpu-tensorrt-ci-pipeline.yml
index 5ff784bf1bc2b..35f3e1df7fc40 100644
--- a/tools/ci_build/github/azure-pipelines/win-gpu-tensorrt-ci-pipeline.yml
+++ b/tools/ci_build/github/azure-pipelines/win-gpu-tensorrt-ci-pipeline.yml
@@ -54,7 +54,7 @@ jobs:
     displayName: 'Generate cmake config'
     inputs:
       scriptPath: '$(Build.SourcesDirectory)\tools\ci_build\build.py'
-      arguments: '--config $(BuildConfig) --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --update --cmake_generator "Visual Studio 16 2019" --build_wheel --enable_onnx_tests --use_tensorrt --tensorrt_placeholder_builder --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=75'
+      arguments: '--config $(BuildConfig) --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --update --cmake_generator "Visual Studio 16 2019" --build_wheel --enable_onnx_tests --use_tensorrt --tensorrt_placeholder_builder --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=75'
       workingDirectory: '$(Build.BinariesDirectory)'
 
   - task: VSBuild@1
@@ -84,7 +84,7 @@ jobs:
      del wheel_filename_file
      python.exe -m pip install -q --upgrade %WHEEL_FILENAME%
      set PATH=$(Build.BinariesDirectory)\$(BuildConfig)\$(BuildConfig);%PATH%
-     python $(Build.SourcesDirectory)\tools\ci_build\build.py --config $(BuildConfig) --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --test --cmake_generator "Visual Studio 16 2019" --build_wheel --enable_onnx_tests --use_tensorrt --tensorrt_placeholder_builder --tensorrt_home="C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=75
+     python $(Build.SourcesDirectory)\tools\ci_build\build.py --config $(BuildConfig) --build_dir $(Build.BinariesDirectory) --skip_submodule_sync --build_shared_lib --test --cmake_generator "Visual Studio 16 2019" --build_wheel --enable_onnx_tests --use_tensorrt --tensorrt_placeholder_builder --tensorrt_home="C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6" --cuda_version=11.6 --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=75
 
     workingDirectory: '$(Build.BinariesDirectory)\$(BuildConfig)\$(BuildConfig)'
     displayName: 'Run tests'
diff --git a/tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5 b/tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5
new file mode 100644
index 0000000000000..5deedc71d0a4d
--- /dev/null
+++ b/tools/ci_build/github/linux/docker/Dockerfile.manylinux2014_cuda11_6_tensorrt8_5
@@ -0,0 +1,184 @@
+ARG BASEIMAGE=nvidia/cuda:11.6.1-cudnn8-devel-centos7
+ARG POLICY=manylinux2014
+ARG PLATFORM=x86_64
+ARG DEVTOOLSET_ROOTPATH=
+ARG LD_LIBRARY_PATH_ARG=
+ARG PREPEND_PATH=
+
+#We need CUDA, TensorRT and manylinux. But the CUDA Toolkit End User License Agreement says NVIDIA CUDA Driver Libraries(libcuda.so, libnvidia-ptxjitcompiler.so) are only distributable in applications that meet this criteria:
+#1. The application was developed starting from a NVIDIA CUDA container obtained from Docker Hub or the NVIDIA GPU Cloud, and
+#2. The resulting application is packaged as a Docker container and distributed to users on Docker Hub or the NVIDIA GPU Cloud only.
+#So we use CUDA as the base image then add manylinux and TensorRT on top of it.
+
+#Build manylinux2014 docker image begin
+FROM $BASEIMAGE AS runtime_base
+ARG POLICY
+ARG PLATFORM
+ARG DEVTOOLSET_ROOTPATH
+ARG LD_LIBRARY_PATH_ARG
+ARG PREPEND_PATH
+LABEL maintainer="The ManyLinux project"
+
+ENV AUDITWHEEL_POLICY=${POLICY} AUDITWHEEL_ARCH=${PLATFORM} AUDITWHEEL_PLAT=${POLICY}_${PLATFORM}
+ENV LC_ALL=en_US.UTF-8 LANG=en_US.UTF-8 LANGUAGE=en_US.UTF-8
+ENV DEVTOOLSET_ROOTPATH=${DEVTOOLSET_ROOTPATH}
+ENV LD_LIBRARY_PATH=${LD_LIBRARY_PATH_ARG}
+ENV PATH=${PREPEND_PATH}${PATH}
+ENV PKG_CONFIG_PATH=/usr/local/lib/pkgconfig
+
+# first copy the fixup mirrors script, keep the script around
+COPY build_scripts/fixup-mirrors.sh /usr/local/sbin/fixup-mirrors
+
+# setup entrypoint, this will wrap commands with `linux32` with i686 images
+COPY build_scripts/install-entrypoint.sh \
+     build_scripts/update-system-packages.sh \
+     build_scripts/build_utils.sh \
+     /build_scripts/
+
+RUN /build_scripts/install-entrypoint.sh && rm -rf /build_scripts
+COPY manylinux-entrypoint /usr/local/bin/manylinux-entrypoint
+ENTRYPOINT ["manylinux-entrypoint"]
+
+COPY build_scripts/install-runtime-packages.sh \
+     build_scripts/update-system-packages.sh \
+     build_scripts/build_utils.sh \
+     /build_scripts/
+RUN manylinux-entrypoint /build_scripts/install-runtime-packages.sh && rm -rf /build_scripts/
+
+COPY build_scripts/build_utils.sh /build_scripts/
+
+COPY build_scripts/install-autoconf.sh /build_scripts/
+RUN export AUTOCONF_ROOT=autoconf-2.71 && \
+    export AUTOCONF_HASH=431075ad0bf529ef13cb41e9042c542381103e80015686222b8a9d4abef42a1c && \
+    export AUTOCONF_DOWNLOAD_URL=http://ftp.gnu.org/gnu/autoconf && \
+    manylinux-entrypoint /build_scripts/install-autoconf.sh
+
+COPY build_scripts/install-automake.sh /build_scripts/
+RUN export AUTOMAKE_ROOT=automake-1.16.5 && \
+    export AUTOMAKE_HASH=07bd24ad08a64bc17250ce09ec56e921d6343903943e99ccf63bbf0705e34605 && \
+    export AUTOMAKE_DOWNLOAD_URL=http://ftp.gnu.org/gnu/automake && \
+    manylinux-entrypoint /build_scripts/install-automake.sh
+
+COPY build_scripts/install-libtool.sh /build_scripts/
+RUN export LIBTOOL_ROOT=libtool-2.4.7 && \
+    export LIBTOOL_HASH=04e96c2404ea70c590c546eba4202a4e12722c640016c12b9b2f1ce3d481e9a8 && \
+    export LIBTOOL_DOWNLOAD_URL=http://ftp.gnu.org/gnu/libtool && \
+    manylinux-entrypoint /build_scripts/install-libtool.sh
+
+COPY build_scripts/install-libxcrypt.sh /build_scripts/
+RUN export LIBXCRYPT_VERSION=4.4.28 && \
+    export LIBXCRYPT_HASH=db7e37901969cb1d1e8020cb73a991ef81e48e31ea5b76a101862c806426b457 && \
+    export LIBXCRYPT_DOWNLOAD_URL=https://github.com/besser82/libxcrypt/archive && \
+    export PERL_ROOT=perl-5.34.0 && \
+    export PERL_HASH=551efc818b968b05216024fb0b727ef2ad4c100f8cb6b43fab615fa78ae5be9a && \
+    export PERL_DOWNLOAD_URL=https://www.cpan.org/src/5.0 && \
+    manylinux-entrypoint /build_scripts/install-libxcrypt.sh
+
+FROM runtime_base AS build_base
+COPY build_scripts/install-build-packages.sh /build_scripts/
+RUN manylinux-entrypoint /build_scripts/install-build-packages.sh
+
+
+FROM build_base AS build_git
+COPY build_scripts/build-git.sh /build_scripts/
+RUN export GIT_ROOT=git-2.36.2 && \
+    export GIT_HASH=6dc2cdea5fb23d823ba4871cc23222c1db31dfbb6d6c6ff74c4128700df57c68 && \
+    export GIT_DOWNLOAD_URL=https://www.kernel.org/pub/software/scm/git && \
+    manylinux-entrypoint /build_scripts/build-git.sh
+
+
+FROM build_base AS build_cpython
+COPY build_scripts/build-sqlite3.sh /build_scripts/
+RUN export SQLITE_AUTOCONF_ROOT=sqlite-autoconf-3390200 && \
+    export SQLITE_AUTOCONF_HASH=852be8a6183a17ba47cee0bbff7400b7aa5affd283bf3beefc34fcd088a239de && \
+    export SQLITE_AUTOCONF_DOWNLOAD_URL=https://www.sqlite.org/2022 && \
+    manylinux-entrypoint /build_scripts/build-sqlite3.sh
+
+COPY build_scripts/build-openssl.sh /build_scripts/
+RUN export OPENSSL_ROOT=openssl-1.1.1q && \
+    export OPENSSL_HASH=d7939ce614029cdff0b6c20f0e2e5703158a489a72b2507b8bd51bf8c8fd10ca && \
+    export OPENSSL_DOWNLOAD_URL=https://www.openssl.org/source && \
+    manylinux-entrypoint /build_scripts/build-openssl.sh
+
+COPY build_scripts/build-cpython.sh /build_scripts/
+
+
+FROM build_cpython AS build_cpython37
+COPY build_scripts/cpython-pubkeys.txt /build_scripts/cpython-pubkeys.txt
+RUN manylinux-entrypoint /build_scripts/build-cpython.sh 3.7.13
+
+
+FROM build_cpython AS build_cpython38
+COPY build_scripts/ambv-pubkey.txt /build_scripts/cpython-pubkeys.txt
+RUN manylinux-entrypoint /build_scripts/build-cpython.sh 3.8.13
+
+
+FROM build_cpython AS build_cpython39
+COPY build_scripts/ambv-pubkey.txt /build_scripts/cpython-pubkeys.txt
+RUN manylinux-entrypoint /build_scripts/build-cpython.sh 3.9.13
+
+
+FROM build_cpython AS build_cpython310
+COPY build_scripts/cpython-pubkey-310-311.txt /build_scripts/cpython-pubkeys.txt
+RUN manylinux-entrypoint /build_scripts/build-cpython.sh 3.10.5
+
+FROM build_cpython AS build_cpython311
+COPY build_scripts/cpython-pubkey-310-311.txt /build_scripts/cpython-pubkeys.txt
+RUN manylinux-entrypoint /build_scripts/build-cpython.sh 3.11.0b5
+
+FROM build_cpython AS all_python
+COPY build_scripts/install-pypy.sh \
+     build_scripts/pypy.sha256 \
+     build_scripts/finalize-python.sh \
+     /build_scripts/
+RUN manylinux-entrypoint /build_scripts/install-pypy.sh 3.7 7.3.9
+RUN manylinux-entrypoint /build_scripts/install-pypy.sh 3.8 7.3.9
+RUN manylinux-entrypoint /build_scripts/install-pypy.sh 3.9 7.3.9
+COPY --from=build_cpython37 /opt/_internal /opt/_internal/
+COPY --from=build_cpython38 /opt/_internal /opt/_internal/
+COPY --from=build_cpython39 /opt/_internal /opt/_internal/
+COPY --from=build_cpython310 /opt/_internal /opt/_internal/
+COPY --from=build_cpython311 /opt/_internal /opt/_internal/
+RUN manylinux-entrypoint /build_scripts/finalize-python.sh
+
+
+FROM runtime_base
+COPY --from=build_git /manylinux-rootfs /
+COPY --from=build_cpython /manylinux-rootfs /
+COPY --from=all_python /opt/_internal /opt/_internal/
+COPY build_scripts/finalize.sh \
+     build_scripts/update-system-packages.sh \
+     build_scripts/python-tag-abi-tag.py \
+     build_scripts/requirements3.7.txt \
+     build_scripts/requirements3.8.txt \
+     build_scripts/requirements3.9.txt \
+     build_scripts/requirements3.10.txt \
+     build_scripts/requirements3.11.txt \
+     build_scripts/requirements-base-tools.txt \
+     /build_scripts/
+COPY build_scripts/requirements-tools/* /build_scripts/requirements-tools/
+RUN manylinux-entrypoint /build_scripts/finalize.sh && rm -rf /build_scripts
+
+ENV SSL_CERT_FILE=/opt/_internal/certs.pem
+
+CMD ["/bin/bash"]
+
+#Build manylinux2014 docker image end
+
+#Install TensorRT 8.5.1.7
+#RUN yum install -y wget
+RUN v="8.5.1-1.cuda11.8" &&\
+    yum-config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel7/x86_64/cuda-rhel7.repo &&\
+    yum -y install libnvinfer8-${v} libnvparsers8-${v} libnvonnxparsers8-${v} libnvinfer-plugin8-${v} \
+        libnvinfer-devel-${v} libnvparsers-devel-${v} libnvonnxparsers-devel-${v} libnvinfer-plugin-devel-${v}
+
+#Add our own dependencies
+ADD scripts /tmp/scripts
+RUN cd /tmp/scripts && /tmp/scripts/manylinux/install_centos.sh && /tmp/scripts/manylinux/install_deps.sh && rm -rf /tmp/scripts
+
+ARG BUILD_UID=1001
+ARG BUILD_USER=onnxruntimedev
+RUN adduser --uid $BUILD_UID $BUILD_USER
+WORKDIR /home/$BUILD_USER
+USER $BUILD_USER
+ENV PATH /usr/local/gradle/bin:/usr/local/dotnet:$PATH
diff --git a/tools/ci_build/github/windows/setup_env_gpu.bat b/tools/ci_build/github/windows/setup_env_gpu.bat
index faaff0387d56b..57adb11d9056b 100644
--- a/tools/ci_build/github/windows/setup_env_gpu.bat
+++ b/tools/ci_build/github/windows/setup_env_gpu.bat
@@ -1,5 +1,6 @@
 REM Copyright (c) Microsoft Corporation. All rights reserved.
 REM Licensed under the MIT License.
 
-set PATH=C:\local\TensorRT-8.4.1.5.Windows10.x86_64.cuda-11.6.cudnn8.4\lib;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\extras\CUPTI\lib64;C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\MSBuild\Current\Bin;%PATH%
+set PATH=C:\local\TensorRT-8.5.1.7.Windows10.x86_64.cuda-11.8.cudnn8.6\lib;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\bin;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\extras\CUPTI\lib64;C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\MSBuild\Current\Bin;%PATH%
 set GRADLE_OPTS=-Dorg.gradle.daemon=false
+set CUDA_MODULE_LOADING=LAZY