diff --git a/.github/scripts/install-tensorrt-rtx.sh b/.github/scripts/install-tensorrt-rtx.sh
index ef5c2945f3..61a0ce2ae6 100644
--- a/.github/scripts/install-tensorrt-rtx.sh
+++ b/.github/scripts/install-tensorrt-rtx.sh
@@ -17,8 +17,8 @@ install_tensorrt_rtx() {
             curl -L https://developer.nvidia.com/downloads/trt/rtx_sdk/secure/1.0/TensorRT-RTX-1.0.0.21.Windows.win10.cuda-12.9.zip -o TensorRT-RTX-1.0.0.21.Windows.win10.cuda-12.9.zip
             unzip TensorRT-RTX-1.0.0.21.Windows.win10.cuda-12.9.zip
             rtx_lib_dir=${PWD}/TensorRT-RTX-1.0.0.21/lib
-            export LD_LIBRARY_PATH=${rtx_lib_dir}:$LD_LIBRARY_PATH
-            echo "LD_LIBRARY_PATH: $LD_LIBRARY_PATH"
+            export PATH=${rtx_lib_dir}:$PATH
+            echo "PATH: $PATH"
             if [[ ${install_wheel_or_not} == true ]]; then
                 pip install TensorRT-RTX-1.0.0.21/python/tensorrt_rtx-1.0.0.21-${CPYTHON_TAG}-none-win_amd64.whl
             fi
diff --git a/.github/workflows/build-test-linux-x86_64.yml b/.github/workflows/build-test-linux-x86_64.yml
index 4b18ef559d..f2b2c8acd7 100644
--- a/.github/workflows/build-test-linux-x86_64.yml
+++ b/.github/workflows/build-test-linux-x86_64.yml
@@ -79,7 +79,7 @@ jobs:
       use-rtx: false
 
   tests-py-torchscript-fe:
-    name: Test torchscript frontend [Python]
+    name: Test resnet50 and bert [Python]
     needs: [filter-matrix, build]
     strategy:
       fail-fast: false
@@ -100,23 +100,18 @@ jobs:
       build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
       pre-script: ${{ matrix.pre-script }}
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
         export LD_LIBRARY_PATH=/usr/lib64:$LD_LIBRARY_PATH
-        pushd .
-        cd tests/modules
-        python hub.py
-        popd
-        pushd .
-        cd tests/py/ts
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_api_test_results.xml api/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_models_test_results.xml models/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_integrations_test_results.xml integrations/
-        popd
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        python tools/perf/resnet50.py
+        python tools/perf/bert.py
 
   tests-py-dynamo-converters:
-    name: Test dynamo converters [Python]
+    name: Test Qwen/Qwen3-0.6B [Python]
     needs: [filter-matrix, build]
     strategy:
       fail-fast: false
@@ -137,208 +132,205 @@ jobs:
       build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
       pre-script: ${{ matrix.pre-script }}
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/py/dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml -n 4 conversion/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_automatic_plugin.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_automatic_plugin_with_attrs.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_flashinfer_rmsnorm.py
-        popd
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        python tools/llm/run_llm.py --model Qwen/Qwen3-0.6B --prompt "Gemma is Google’s family of open-weights LLMs, launched in February 2024 (Gemma 1.0, 2B/7B sizes). Designed to be lightweight, efficient, can you please tell me the difference between gemma3 and deepseek? which one is better in performance?" --precision FP16 --num_tokens 128 --benchmark
 
-  tests-py-dynamo-fe:
-    name: Test dynamo frontend [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-fe
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
-        popd
+  # tests-py-dynamo-fe:
+  #   name: Test dynamo frontend [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-fe
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
+  #       popd
 
-  tests-py-dynamo-serde:
-    name: Test dynamo export serde [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-serde
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
-        popd
+  # tests-py-dynamo-serde:
+  #   name: Test dynamo export serde [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-serde
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
+  #       popd
 
-  tests-py-torch-compile-be:
-    name: Test torch compile backend [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-torch-compile-be
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
-        popd
+  # tests-py-torch-compile-be:
+  #   name: Test torch compile backend [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-torch-compile-be
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
+  #       popd
 
-  tests-py-dynamo-core:
-    name: Test dynamo core [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-core
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
-        popd
+  # tests-py-dynamo-core:
+  #   name: Test dynamo core [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-core
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
+  #       popd
 
-  tests-py-dynamo-cudagraphs:
-    name: Test dynamo cudagraphs [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-cudagraphs
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        nvidia-smi
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py || true
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py || true
-        popd
+  # tests-py-dynamo-cudagraphs:
+  #   name: Test dynamo cudagraphs [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-cudagraphs
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       nvidia-smi
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py || true
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py || true
+  #       popd
 
-  tests-py-core:
-    name: Test core [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-core
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py/core
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
-        popd
+  # tests-py-core:
+  #   name: Test core [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-core
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py/core
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
+  #       popd
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-tensorrt-${{ inputs.repository }}-${{ github.event_name == 'workflow_dispatch' }}-${{ inputs.job-name }}
diff --git a/.github/workflows/build-test-linux-x86_64_rtx.yml b/.github/workflows/build-test-linux-x86_64_rtx.yml
index 24d0169e2c..f4900cb2bd 100644
--- a/.github/workflows/build-test-linux-x86_64_rtx.yml
+++ b/.github/workflows/build-test-linux-x86_64_rtx.yml
@@ -78,7 +78,7 @@ jobs:
       use-rtx: true
 
   tests-py-torchscript-fe:
-    name: Test torchscript frontend [Python]
+    name: Test resnet50 and bert [Python]
     needs: [filter-matrix, build]
     strategy:
       fail-fast: false
@@ -100,23 +100,18 @@ jobs:
       pre-script: ${{ matrix.pre-script }}
       use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
         export LD_LIBRARY_PATH=/usr/lib64:$LD_LIBRARY_PATH
-        pushd .
-        cd tests/modules
-        python hub.py
-        popd
-        pushd .
-        cd tests/py/ts
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_api_test_results.xml api/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_models_test_results.xml models/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_integrations_test_results.xml integrations/
-        popd
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        python tools/perf/resnet50.py
+        python tools/perf/bert.py
 
   tests-py-dynamo-converters:
-    name: Test dynamo converters [Python]
+    name: Test Qwen/Qwen3-0.6B [Python]
     needs: [filter-matrix, build]
     strategy:
       fail-fast: false
@@ -138,214 +133,211 @@ jobs:
       pre-script: ${{ matrix.pre-script }}
       use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/py/dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml -n 4 conversion/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_automatic_plugin.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_automatic_plugin_with_attrs.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml automatic_plugin/test_flashinfer_rmsnorm.py
-        popd
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        python tools/llm/run_llm.py --model Qwen/Qwen3-0.6B --prompt "Gemma is Google’s family of open-weights LLMs, launched in February 2024 (Gemma 1.0, 2B/7B sizes). Designed to be lightweight, efficient, can you please tell me the difference between gemma3 and deepseek? which one is better in performance?" --precision FP16 --num_tokens 128 --benchmark
 
-  tests-py-dynamo-fe:
-    name: Test dynamo frontend [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-fe
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
-        popd
+  # tests-py-dynamo-fe:
+  #   name: Test dynamo frontend [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-fe
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
+  #       popd
 
-  tests-py-dynamo-serde:
-    name: Test dynamo export serde [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-serde
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
-        popd
+  # tests-py-dynamo-serde:
+  #   name: Test dynamo export serde [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-serde
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
+  #       popd
 
-  tests-py-torch-compile-be:
-    name: Test torch compile backend [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-torch-compile-be
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
-        popd
+  # tests-py-torch-compile-be:
+  #   name: Test torch compile backend [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-torch-compile-be
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
+  #       popd
 
-  tests-py-dynamo-core:
-    name: Test dynamo core [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-core
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
-        popd
+  # tests-py-dynamo-core:
+  #   name: Test dynamo core [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-core
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
+  #       popd
 
-  tests-py-dynamo-cudagraphs:
-    name: Test dynamo cudagraphs [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-dynamo-cudagraphs
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        nvidia-smi
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py || true
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py || true
-        popd
+  # tests-py-dynamo-cudagraphs:
+  #   name: Test dynamo cudagraphs [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-cudagraphs
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       nvidia-smi
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py || true
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py || true
+  #       popd
 
-  tests-py-core:
-    name: Test core [Python]
-    needs: [filter-matrix, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-            pre-script: packaging/pre_build_script.sh
-            post-script: packaging/post_build_script.sh
-            smoke-test-script: packaging/smoke_test_script.sh
-    uses: ./.github/workflows/linux-test.yml
-    with:
-      job-name: tests-py-core
-      repository: "pytorch/tensorrt"
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
-      pre-script: ${{ matrix.pre-script }}
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py/core
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
-        popd
+  # tests-py-core:
+  #   name: Test core [Python]
+  #   needs: [filter-matrix, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #           pre-script: packaging/pre_build_script.sh
+  #           post-script: packaging/post_build_script.sh
+  #           smoke-test-script: packaging/smoke_test_script.sh
+  #   uses: ./.github/workflows/linux-test.yml
+  #   with:
+  #     job-name: tests-py-core
+  #     repository: "pytorch/tensorrt"
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.filter-matrix.outputs.matrix }}
+  #     pre-script: ${{ matrix.pre-script }}
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py/core
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
+  #       popd
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-tensorrt-rtx-${{ inputs.repository }}-${{ github.event_name == 'workflow_dispatch' }}-${{ inputs.job-name }}
diff --git a/.github/workflows/build-test-windows.yml b/.github/workflows/build-test-windows.yml
index 2d402a8799..a88adbe682 100644
--- a/.github/workflows/build-test-windows.yml
+++ b/.github/workflows/build-test-windows.yml
@@ -85,8 +85,9 @@ jobs:
       trigger-event: ${{ github.event_name }}
       timeout: 120
 
+
   tests-py-torchscript-fe:
-    name: Test torchscript frontend [Python]
+    name: Test Restnet/Bert [Python]
     needs: [substitute-runner, build]
     strategy:
       fail-fast: false
@@ -103,23 +104,21 @@ jobs:
       test-infra-ref: main
       build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
       pre-script: packaging/driver_upgrade.bat
+      use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/modules
-        python hub.py
-        popd
-        pushd .
-        cd tests/py/ts
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_api_test_results.xml api/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_models_test_results.xml models/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_integrations_test_results.xml integrations/
-        popd
+        export USE_TRT_RTX=false
+        echo $LD_LIBRARY_PATH
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        packaging/vc_env_helper.bat python tools/perf/resnet50.py
+        packaging/vc_env_helper.bat python tools/perf/bert.py
 
   tests-py-dynamo-converters:
-    name: Test dynamo converters [Python]
+    name: Test Qwen/Qwen3-0.6B [Python]
     needs: [substitute-runner, build]
     strategy:
       fail-fast: false
@@ -136,188 +135,191 @@ jobs:
       test-infra-ref: main
       build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
       pre-script: packaging/driver_upgrade.bat
+      use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml -n 4 conversion/
-        popd
+        export USE_TRT_RTX=false
+        echo $LD_LIBRARY_PATH
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        packaging/vc_env_helper.bat  python tools/llm/run_llm.py --model Qwen/Qwen3-0.6B --prompt "Gemma is Google’s family of open-weights LLMs, launched in February 2024 (Gemma 1.0, 2B/7B sizes). Designed to be lightweight, efficient, can you please tell me the difference between gemma3 and deepseek? which one is better in performance?" --precision FP16 --num_tokens 128 --benchmark
 
-  tests-py-dynamo-fe:
-    name: Test dynamo frontend [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-fe
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
-        popd
 
-  tests-py-dynamo-serde:
-    name: Test dynamo export serde [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-serde
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
-        popd
+  # tests-py-dynamo-fe:
+  #   name: Test dynamo frontend [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-fe
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
+  #       popd
 
-  tests-py-torch-compile-be:
-    name: Test torch compile backend [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-torch-compile-be
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
-        popd
+  # tests-py-dynamo-serde:
+  #   name: Test dynamo export serde [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-serde
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
+  #       popd
 
-  tests-py-dynamo-core:
-    name: Test dynamo core [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-core
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
-        popd
+  # tests-py-torch-compile-be:
+  #   name: Test torch compile backend [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-torch-compile-be
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
+  #       popd
 
-  tests-py-dynamo-cudagraphs:
-    name: Test dynamo cudagraphs [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-cudagraphs
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py
-        popd
+  # tests-py-dynamo-core:
+  #   name: Test dynamo core [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-core
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
+  #       popd
 
-  tests-py-core:
-    name: Test core [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-core
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py/core
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
-        popd
+  # tests-py-dynamo-cudagraphs:
+  #   name: Test dynamo cudagraphs [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-cudagraphs
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py
+  #       popd
+
+  # tests-py-core:
+  #   name: Test core [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-core
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py/core
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
+  #       popd
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-${{ inputs.repository }}-${{ github.event_name == 'workflow_dispatch' }}-${{ inputs.job-name }}
diff --git a/.github/workflows/build-test-windows_rtx.yml b/.github/workflows/build-test-windows_rtx.yml
index 6321cd8a52..044320b6a8 100644
--- a/.github/workflows/build-test-windows_rtx.yml
+++ b/.github/workflows/build-test-windows_rtx.yml
@@ -88,7 +88,7 @@ jobs:
       timeout: 120
 
   tests-py-torchscript-fe:
-    name: Test torchscript frontend [Python]
+    name: Test Restnet/Bert [Python]
     needs: [substitute-runner, build]
     strategy:
       fail-fast: false
@@ -107,22 +107,19 @@ jobs:
       pre-script: packaging/driver_upgrade.bat
       use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/modules
-        python hub.py
-        popd
-        pushd .
-        cd tests/py/ts
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_api_test_results.xml api/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_models_test_results.xml models/
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/ts_integrations_test_results.xml integrations/
-        popd
+        export USE_TRT_RTX=true
+        echo $LD_LIBRARY_PATH
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        packaging/vc_env_helper.bat python tools/perf/resnet50.py
+        packaging/vc_env_helper.bat python tools/perf/bert.py
 
   tests-py-dynamo-converters:
-    name: Test dynamo converters [Python]
+    name: Test Qwen/Qwen3-0.6B [Python]
     needs: [substitute-runner, build]
     strategy:
       fail-fast: false
@@ -141,193 +138,194 @@ jobs:
       pre-script: packaging/driver_upgrade.bat
       use-rtx: true
       script: |
-        set -euo pipefail
+        set -x
         export USE_HOST_DEPS=1
         export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dynamo_converters_test_results.xml -n 4 conversion/
-        popd
+        export USE_TRT_RTX=true
+        echo $LD_LIBRARY_PATH
+        nvidia-smi
+        python -c "import torch; print(torch.cuda.get_device_capability())"
+        python -c "import torch; print(torch.cuda.get_device_properties(0))"
+        packaging/vc_env_helper.bat  python tools/llm/run_llm.py --model Qwen/Qwen3-0.6B --prompt "Gemma is Google’s family of open-weights LLMs, launched in February 2024 (Gemma 1.0, 2B/7B sizes). Designed to be lightweight, efficient, can you please tell me the difference between gemma3 and deepseek? which one is better in performance?" --precision FP16 --num_tokens 128 --benchmark
 
-  tests-py-dynamo-fe:
-    name: Test dynamo frontend [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-fe
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
-        popd
+  # tests-py-dynamo-fe:
+  #   name: Test dynamo frontend [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-fe
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/dyn_models_export.xml --ir dynamo models/
+  #       popd
 
-  tests-py-dynamo-serde:
-    name: Test dynamo export serde [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-serde
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
-        popd
+  # tests-py-dynamo-serde:
+  #   name: Test dynamo export serde [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-serde
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/export_serde_test_results.xml --ir dynamo models/test_export_serde.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/reexport_test_results.xml --ir dynamo models/test_reexport.py
+  #       popd
 
-  tests-py-torch-compile-be:
-    name: Test torch compile backend [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-torch-compile-be
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
-        popd
+  # tests-py-torch-compile-be:
+  #   name: Test torch compile backend [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-torch-compile-be
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra -n 10 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_be_test_results.xml backend/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_complete_be_e2e_test_results.xml --ir torch_compile models/test_models.py
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/torch_compile_dyn_models_export.xml --ir torch_compile models/test_dyn_models.py
+  #       popd
 
-  tests-py-dynamo-core:
-    name: Test dynamo core [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-core
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
-        ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
-        popd
+  # tests-py-dynamo-core:
+  #   name: Test dynamo core [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-core
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_test_results.xml --ignore runtime/test_002_cudagraphs_py.py --ignore runtime/test_002_cudagraphs_cpp.py runtime/
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_partitioning_test_results.xml partitioning/
+  #       ../../../packaging/vc_env_helper.bat python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_lowering_test_results.xml lowering/
+  #       popd
 
-  tests-py-dynamo-cudagraphs:
-    name: Test dynamo cudagraphs [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-dynamo-cudagraphs
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py
-        cd dynamo
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py
-        python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py
-        popd
+  # tests-py-dynamo-cudagraphs:
+  #   name: Test dynamo cudagraphs [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-dynamo-cudagraphs
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py
+  #       cd dynamo
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_cpp_test_results.xml runtime/test_002_cudagraphs_cpp.py
+  #       python -m pytest -ra --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_dynamo_core_runtime_cudagraphs_py_test_results.xml runtime/test_002_cudagraphs_py.py
+  #       popd
 
-  tests-py-core:
-    name: Test core [Python]
-    needs: [substitute-runner, build]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - repository: pytorch/tensorrt
-            package-name: torch_tensorrt
-    uses: ./.github/workflows/windows-test.yml
-    with:
-      job-name: tests-py-core
-      repository: ${{ matrix.repository }}
-      ref: ""
-      test-infra-repository: pytorch/test-infra
-      test-infra-ref: main
-      build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
-      pre-script: packaging/driver_upgrade.bat
-      use-rtx: true
-      script: |
-        set -euo pipefail
-        export USE_HOST_DEPS=1
-        export CI_BUILD=1
-        pushd .
-        cd tests/py/core
-        python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
-        popd
+  # tests-py-core:
+  #   name: Test core [Python]
+  #   needs: [substitute-runner, build]
+  #   strategy:
+  #     fail-fast: false
+  #     matrix:
+  #       include:
+  #         - repository: pytorch/tensorrt
+  #           package-name: torch_tensorrt
+  #   uses: ./.github/workflows/windows-test.yml
+  #   with:
+  #     job-name: tests-py-core
+  #     repository: ${{ matrix.repository }}
+  #     ref: ""
+  #     test-infra-repository: pytorch/test-infra
+  #     test-infra-ref: main
+  #     build-matrix: ${{ needs.substitute-runner.outputs.matrix }}
+  #     pre-script: packaging/driver_upgrade.bat
+  #     use-rtx: true
+  #     script: |
+  #       set -euo pipefail
+  #       export USE_HOST_DEPS=1
+  #       export CI_BUILD=1
+  #       pushd .
+  #       cd tests/py/core
+  #       python -m pytest -ra -n 4 --junitxml=${RUNNER_TEST_RESULTS_DIR}/tests_py_core_test_results.xml .
+  #       popd
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-tensorrt-rtx-${{ inputs.repository }}-${{ github.event_name == 'workflow_dispatch' }}-${{ inputs.job-name }}
diff --git a/tools/perf/bert.py b/tools/perf/bert.py
new file mode 100644
index 0000000000..9f8817026f
--- /dev/null
+++ b/tools/perf/bert.py
@@ -0,0 +1,95 @@
+import tokenize
+from pickle import FALSE, TRUE
+from time import perf_counter
+
+import torch
+import torch_tensorrt
+from torch.export import Dim
+from transformers import BertModel, BertTokenizer
+
+dtypes = [torch.float32, torch.float16, torch.bfloat16]
+dynamic_shape_supports = [False]
+iterations = 20
+
+results = {}
+
+
+def run_bert(dtype: torch.dtype):
+    model = BertModel.from_pretrained("bert-base-uncased").cuda().eval().to(dtype)
+    tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+    inputs = tokenizer("Hello, how are you?", return_tensors="pt").to("cuda")
+    input_ids = inputs["input_ids"]
+    token_type_ids = inputs["token_type_ids"]
+    inputs = (input_ids, token_type_ids)
+    exp_program = torch.export.export(model, args=inputs)
+    compile_start = perf_counter()
+    compiled_model = torch_tensorrt.dynamo.compile(
+        exp_program,
+        inputs=inputs,
+        use_explicit_type=True,
+        min_block_size=1,
+        immutable_weights=True,
+        cache_built_engines=False,
+        reuse_cached_engines=False,
+        use_python_runtime=False,
+    )
+    compile_end = perf_counter()
+    compile_time = compile_end - compile_start
+
+    with torch.no_grad():
+        start = perf_counter()
+        compiled_model(*inputs)
+        end = perf_counter()
+        first_inference_time = end - start
+        start = perf_counter()
+        for i in range(iterations):
+            compiled_model(*inputs)
+        end = perf_counter()
+        second_inference_time = (end - start) / iterations
+    return compile_time, first_inference_time, second_inference_time
+
+
+for dtype in dtypes:
+    results[dtype] = {}
+    for dynamic_shape_support in dynamic_shape_supports:
+        results[dtype][dynamic_shape_support] = {}
+        total_compile_time = 0
+        total_first_inference_time = 0
+        total_second_inference_time = 0
+        for i in range(iterations):
+            compile_time, first_inference_time, second_inference_time = run_bert(dtype)
+            total_compile_time += compile_time
+            total_first_inference_time += first_inference_time
+            total_second_inference_time += second_inference_time
+
+        results[dtype][dynamic_shape_support]["compile_time"] = (
+            total_compile_time / iterations
+        )
+        results[dtype][dynamic_shape_support]["first_inference_time"] = (
+            total_first_inference_time / iterations
+        )
+        results[dtype][dynamic_shape_support]["second_inference_time"] = (
+            total_second_inference_time / iterations
+        )
+
+for dype in dtypes:
+    for dynamic_shape_support in dynamic_shape_supports:
+        print(f"")
+        print(
+            f"================================================================================"
+        )
+        print(f"========Bert model: {dtype=} {dynamic_shape_support=}=========")
+        print(f"compile_time: {results[dtype][dynamic_shape_support]['compile_time']}")
+        print(
+            f"first_inference_time: {results[dtype][dynamic_shape_support]['first_inference_time']}"
+        )
+        print(
+            f"second_inference_time: {results[dtype][dynamic_shape_support]['second_inference_time']}"
+        )
+        print(
+            f"================================================================================"
+        )
+        print(
+            f"================================================================================"
+        )
+        print(f"")
diff --git a/tools/perf/resnet50.py b/tools/perf/resnet50.py
new file mode 100644
index 0000000000..5e290e42c0
--- /dev/null
+++ b/tools/perf/resnet50.py
@@ -0,0 +1,99 @@
+from pickle import FALSE, TRUE
+from time import perf_counter
+
+import torch
+import torch_tensorrt
+import torchvision.models as models
+from torch.export import Dim
+
+dtypes = [torch.float32, torch.float16, torch.bfloat16]
+dynamic_shape_supports = [True, False]
+
+iterations = 20
+
+results = {}
+
+
+def run_resnet(dtype: torch.dtype, dynamic_shape_support: bool):
+    model = models.resnet50(pretrained=True).to(dtype).eval().to("cuda")
+    inputs = (torch.randn((100, 3, 224, 224)).to(dtype).to("cuda"),)
+    if dynamic_shape_support:
+        batch_dim = Dim("batch_dim", min=2, max=100)
+        exp_program = torch.export.export(
+            model, args=inputs, dynamic_shapes=({0: batch_dim},)
+        )
+    else:
+        exp_program = torch.export.export(model, args=inputs)
+    compile_start = perf_counter()
+    compiled_model = torch_tensorrt.dynamo.compile(
+        exp_program,
+        inputs=inputs,
+        use_explicit_type=True,
+        min_block_size=1,
+        immutable_weights=True,
+        cache_built_engines=False,
+        reuse_cached_engines=False,
+        use_python_runtime=False,
+    )
+    compile_end = perf_counter()
+    compile_time = compile_end - compile_start
+
+    with torch.no_grad():
+        start = perf_counter()
+        compiled_model(*inputs)
+        end = perf_counter()
+        first_inference_time = end - start
+        start = perf_counter()
+        for i in range(iterations):
+            compiled_model(*inputs)
+        end = perf_counter()
+        second_inference_time = (end - start) / iterations
+    return compile_time, first_inference_time, second_inference_time
+
+
+for dtype in dtypes:
+    results[dtype] = {}
+    for dynamic_shape_support in dynamic_shape_supports:
+        results[dtype][dynamic_shape_support] = {}
+        total_compile_time = 0
+        total_first_inference_time = 0
+        total_second_inference_time = 0
+        for i in range(iterations):
+            compile_time, first_inference_time, second_inference_time = run_resnet(
+                dtype, dynamic_shape_support
+            )
+            total_compile_time += compile_time
+            total_first_inference_time += first_inference_time
+            total_second_inference_time += second_inference_time
+
+        results[dtype][dynamic_shape_support]["compile_time"] = (
+            total_compile_time / iterations
+        )
+        results[dtype][dynamic_shape_support]["first_inference_time"] = (
+            total_first_inference_time / iterations
+        )
+        results[dtype][dynamic_shape_support]["second_inference_time"] = (
+            total_second_inference_time / iterations
+        )
+
+for dype in dtypes:
+    for dynamic_shape_support in dynamic_shape_supports:
+        print(f"")
+        print(
+            f"================================================================================"
+        )
+        print(f"========Resnet50 model: {dtype=} {dynamic_shape_support=}=========")
+        print(f"compile_time: {results[dtype][dynamic_shape_support]['compile_time']}")
+        print(
+            f"first_inference_time: {results[dtype][dynamic_shape_support]['first_inference_time']}"
+        )
+        print(
+            f"second_inference_time: {results[dtype][dynamic_shape_support]['second_inference_time']}"
+        )
+        print(
+            f"================================================================================"
+        )
+        print(
+            f"================================================================================"
+        )
+        print(f"")