Skip to content

[TRTLLM-13784][chore] Remove legacy TensorRT-engine Triton backend #15907

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Wanli-Jiang merged 5 commits into from
Jul 9, 2026
Merged

[TRTLLM-13784][chore] Remove legacy TensorRT-engine Triton backend #15907

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
74f9eae
[TRTLLM-13784][chore] Remove legacy TensorRT-engine Triton backend
Wanli-Jiang Jul 2, 2026
09e5337
[https://nvbugs/6414760][fix] Rewrite test_qwen_e2e_cpprunner_large_n…
Wanli-Jiang Jul 7, 2026
f410003
[TRTLLM-13784][chore] Remove TRT-flow Triton dev tests
Wanli-Jiang Jul 7, 2026
920b4a5
[TRTLLM-13784][chore] Remove orphaned triton_server test helpers
Wanli-Jiang Jul 8, 2026
5ae5501
[None][infra] Update Docker image tags to PR 15907 build
tburt-nv Jul 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 2 additions & 68 deletions .pre-commit-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1325,55 +1325,22 @@ common-files: &common_files |
tests/unittest/utils/test_util.py |
tests/unittest/utils/torch_ref.py |
tests/unittest/utils/util.py |
triton_backend/all_models/disaggregated_serving/disaggregated_serving_bls/1/model.py |
triton_backend/all_models/gpt/postprocessing/1/model.py |
triton_backend/all_models/gpt/preprocessing/1/model.py |
triton_backend/all_models/gpt/tensorrt_llm/1/model.py |
triton_backend/all_models/inflight_batcher_llm/postprocessing/1/model.py |
triton_backend/all_models/inflight_batcher_llm/preprocessing/1/model.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm/1/model.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/decode.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/triton_decoder.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/model.py |
triton_backend/all_models/llmapi/tensorrt_llm/1/helpers.py |
triton_backend/all_models/llmapi/tensorrt_llm/1/model.py |
triton_backend/all_models/multimodal/multimodal_encoders/1/model.py |
triton_backend/all_models/multimodal/multimodal_encoders/1/multimodal_utils.py |
triton_backend/all_models/tests/test_decode.py |
triton_backend/all_models/tests/test_llmapi_python_backend.py |
triton_backend/all_models/tests/test_multi_image_preprocess.py |
triton_backend/all_models/tests/test_multimodal_encoders.py |
triton_backend/all_models/tests/test_python_backend.py |
triton_backend/all_models/tests/test_triton_decoder.py |
triton_backend/all_models/whisper/whisper_bls/1/fbank.py |
triton_backend/all_models/whisper/whisper_bls/1/model.py |
triton_backend/all_models/whisper/whisper_bls/1/tokenizer.py |
triton_backend/ci/L0_backend_trtllm/base_metrics_verification_tests.py |
triton_backend/ci/L0_backend_trtllm/custom_metrics_verification_tests.py |
triton_backend/inflight_batcher_llm/client/__init__.py |
triton_backend/inflight_batcher_llm/client/e2e_grpc_speculative_decoding_client.py |
triton_backend/inflight_batcher_llm/client/end_to_end_grpc_client.py |
triton_backend/inflight_batcher_llm/client/inflight_batcher_llm_client.py |
triton_backend/scripts/launch_triton_server.py |
triton_backend/tools/__init__.py |
triton_backend/tools/fill_template.py |
triton_backend/tools/gpt/benchmark_core_model.py |
triton_backend/tools/gpt/client.py |
triton_backend/tools/gpt/client_async.py |
triton_backend/tools/gpt/end_to_end_test.py |
triton_backend/tools/gpt/gen_input_data.py |
triton_backend/tools/inflight_batcher_llm/benchmark_core_model.py |
triton_backend/tools/inflight_batcher_llm/end_to_end_test.py |
triton_backend/tools/inflight_batcher_llm/speculative_decoding_test.py |
triton_backend/tools/inflight_batcher_llm/test_max_queue_size.py |
triton_backend/tools/llmapi_client.py |
triton_backend/tools/multimodal/client.py |
triton_backend/tools/tests/__init__.py |
triton_backend/tools/tests/test_fill_template.py |
triton_backend/tools/tests/test_llmapi_cancel.py |
triton_backend/tools/utils/__init__.py |
triton_backend/tools/utils/utils.py |
triton_backend/tools/whisper/client.py
triton_backend/tools/utils/utils.py
)$

# Used by ruff hooks: main ruff (exclude: *legacy_files) and
Expand Down Expand Up @@ -2699,55 +2666,22 @@ legacy-files: &legacy_files |
tests/unittest/utils/test_util.py |
tests/unittest/utils/torch_ref.py |
tests/unittest/utils/util.py |
triton_backend/all_models/disaggregated_serving/disaggregated_serving_bls/1/model.py |
triton_backend/all_models/gpt/postprocessing/1/model.py |
triton_backend/all_models/gpt/preprocessing/1/model.py |
triton_backend/all_models/gpt/tensorrt_llm/1/model.py |
triton_backend/all_models/inflight_batcher_llm/postprocessing/1/model.py |
triton_backend/all_models/inflight_batcher_llm/preprocessing/1/model.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm/1/model.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/decode.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/triton_decoder.py |
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/model.py |
triton_backend/all_models/llmapi/tensorrt_llm/1/helpers.py |
triton_backend/all_models/llmapi/tensorrt_llm/1/model.py |
triton_backend/all_models/multimodal/multimodal_encoders/1/model.py |
triton_backend/all_models/multimodal/multimodal_encoders/1/multimodal_utils.py |
triton_backend/all_models/tests/test_decode.py |
triton_backend/all_models/tests/test_llmapi_python_backend.py |
triton_backend/all_models/tests/test_multi_image_preprocess.py |
triton_backend/all_models/tests/test_multimodal_encoders.py |
triton_backend/all_models/tests/test_python_backend.py |
triton_backend/all_models/tests/test_triton_decoder.py |
triton_backend/all_models/whisper/whisper_bls/1/fbank.py |
triton_backend/all_models/whisper/whisper_bls/1/model.py |
triton_backend/all_models/whisper/whisper_bls/1/tokenizer.py |
triton_backend/ci/L0_backend_trtllm/base_metrics_verification_tests.py |
triton_backend/ci/L0_backend_trtllm/custom_metrics_verification_tests.py |
triton_backend/inflight_batcher_llm/client/__init__.py |
triton_backend/inflight_batcher_llm/client/e2e_grpc_speculative_decoding_client.py |
triton_backend/inflight_batcher_llm/client/end_to_end_grpc_client.py |
triton_backend/inflight_batcher_llm/client/inflight_batcher_llm_client.py |
triton_backend/scripts/launch_triton_server.py |
triton_backend/tools/__init__.py |
triton_backend/tools/fill_template.py |
triton_backend/tools/gpt/benchmark_core_model.py |
triton_backend/tools/gpt/client.py |
triton_backend/tools/gpt/client_async.py |
triton_backend/tools/gpt/end_to_end_test.py |
triton_backend/tools/gpt/gen_input_data.py |
triton_backend/tools/inflight_batcher_llm/benchmark_core_model.py |
triton_backend/tools/inflight_batcher_llm/end_to_end_test.py |
triton_backend/tools/inflight_batcher_llm/speculative_decoding_test.py |
triton_backend/tools/inflight_batcher_llm/test_max_queue_size.py |
triton_backend/tools/llmapi_client.py |
triton_backend/tools/multimodal/client.py |
triton_backend/tools/tests/__init__.py |
triton_backend/tools/tests/test_fill_template.py |
triton_backend/tools/tests/test_llmapi_cancel.py |
triton_backend/tools/utils/__init__.py |
triton_backend/tools/utils/utils.py |
triton_backend/tools/whisper/client.py
triton_backend/tools/utils/utils.py
)$
# <<<< END AUTO-GENERATED >>>>

Expand Down
12 changes: 0 additions & 12 deletions docker/Dockerfile.multi
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -181,21 +181,9 @@ ENV TRT_LLM_GIT_COMMIT=${GIT_COMMIT} \
RUN --mount=type=bind,source=scripts/generate_container_oss_attribution.sh,target=/mnt/gen_attribution.sh \
bash /mnt/gen_attribution.sh "release" "${TRT_LLM_VER}" "${TARGETARCH}"

FROM wheel AS tritonbuild

WORKDIR /src/tensorrt_llm
RUN pip install /src/tensorrt_llm/build/tensorrt_llm*.whl
COPY ./triton_backend/ ./triton_backend/
ARG TRITON_BASE_TAG
RUN bash ./triton_backend/inflight_batcher_llm/scripts/build.sh -s "r${TRITON_BASE_TAG%-py3}"


FROM release AS tritonrelease

WORKDIR /app/tensorrt_llm
COPY ./triton_backend/all_models ./triton_backend/all_models
COPY ./triton_backend/scripts ./triton_backend/scripts
COPY ./triton_backend/tools ./triton_backend/tools
COPY ./triton_backend/inflight_batcher_llm/scripts ./triton_backend/inflight_batcher_llm/scripts
COPY ./triton_backend/inflight_batcher_llm/client ./triton_backend/inflight_batcher_llm/client
COPY --from=tritonbuild /opt/tritonserver/backends/tensorrtllm /opt/tritonserver/backends/tensorrtllm
12 changes: 6 additions & 6 deletions docs/source/developer-guide/ci-overview.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,14 +24,14 @@ Integration tests are listed under `tests/integration/test_lists/test-db/`. Most
- `stage`: either `pre_merge` or `post_merge`.
- `backend`: `pytorch`, `tensorrt` or `triton`.

Example from `l0_a100.yml`:
Example from `l0_a30.yml`:

```yaml
terms:
stage: post_merge
backend: triton
stage: pre_merge
backend: pytorch
tests:
- triton_server/test_triton.py::test_gpt_ib_ptuning[gpt-ib-ptuning]
- unittest/_torch/sampler/test_beam_search.py
```

## Unit tests
Expand Down Expand Up @@ -61,8 +61,8 @@ Manually searching YAML and Groovy files can be tedious. The helper script
`scripts/test_to_stage_mapping.py` automates the lookup:

```bash
python scripts/test_to_stage_mapping.py --tests "triton_server/test_triton.py::test_gpt_ib_ptuning[gpt-ib-ptuning]"
python scripts/test_to_stage_mapping.py --tests gpt_ib_ptuning
python scripts/test_to_stage_mapping.py --tests "unittest/_torch/sampler/test_beam_search.py"
python scripts/test_to_stage_mapping.py --tests test_beam_search
python scripts/test_to_stage_mapping.py --stages A100X-Triton-Post-Merge-1
python scripts/test_to_stage_mapping.py --test-list my_tests.txt
python scripts/test_to_stage_mapping.py --test-list my_tests.yml
Expand Down
16 changes: 2 additions & 14 deletions jenkins/Build.groovy
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -427,30 +427,18 @@ def runLLMBuild(pipeline, buildFlags, tarName, is_linux_x86_64)
sh "cd ${LLM_ROOT} && python3 scripts/build_cpp_examples.py"
}

// Build tritonserver artifacts
def llmPath = sh (script: "realpath ${LLM_ROOT}",returnStdout: true).trim()
// TODO: Remove after the cmake version is upgraded to 3.31.8
// Get triton tag from docker/dockerfile.multi
def tritonShortTag = "r26.04"
sh "cd ${LLM_ROOT}/triton_backend/inflight_batcher_llm && mkdir build && cd build && cmake .. -DTRTLLM_DIR=${llmPath} -DTRITON_COMMON_REPO_TAG=${tritonShortTag} -DTRITON_CORE_REPO_TAG=${tritonShortTag} -DTRITON_THIRD_PARTY_REPO_TAG=${tritonShortTag} -DTRITON_BACKEND_REPO_TAG=${tritonShortTag} -DUSE_CXX11_ABI=ON && make -j${buildJobs} install"

// Step 3: packaging wheels into tarfile
sh "cp ${LLM_ROOT}/build/tensorrt_llm-*.whl TensorRT-LLM/"

// Step 4: packaging tritonserver artifacts into tarfile
sh "mkdir -p TensorRT-LLM/triton_backend/inflight_batcher_llm/"
sh "cp ${LLM_ROOT}/triton_backend/inflight_batcher_llm/build/libtriton_tensorrtllm.so TensorRT-LLM/triton_backend/inflight_batcher_llm/"
sh "cp ${LLM_ROOT}/triton_backend/inflight_batcher_llm/build/trtllmExecutorWorker TensorRT-LLM/triton_backend/inflight_batcher_llm/"

// Step 5: packaging benchmark and required cpp dependencies into tarfile
// Step 4: packaging benchmark and required cpp dependencies into tarfile
sh "mkdir -p TensorRT-LLM/benchmarks/cpp"
sh "cp ${LLM_ROOT}/cpp/build/benchmarks/bertBenchmark TensorRT-LLM/benchmarks/cpp"
sh "cp ${LLM_ROOT}/cpp/build/benchmarks/gptManagerBenchmark TensorRT-LLM/benchmarks/cpp"
sh "cp ${LLM_ROOT}/cpp/build/benchmarks/disaggServerBenchmark TensorRT-LLM/benchmarks/cpp"
sh "cp ${LLM_ROOT}/cpp/build/tensorrt_llm/libtensorrt_llm.so TensorRT-LLM/benchmarks/cpp"
sh "cp ${LLM_ROOT}/cpp/build/tensorrt_llm/plugins/libnvinfer_plugin_tensorrt_llm.so TensorRT-LLM/benchmarks/cpp"

// Step 6: packaging attribution files into tarfile when they exist
// Step 5: packaging attribution files into tarfile when they exist
sh "mkdir -p TensorRT-LLM/attribution"
sh "cp ${LLM_ROOT}/cpp/build/attribution/missing_files.json TensorRT-LLM/attribution/ || true"
sh "cp ${LLM_ROOT}/cpp/build/attribution/import_payload.json TensorRT-LLM/attribution/ || true"
Expand Down
10 changes: 2 additions & 8 deletions jenkins/L0_Test.groovy
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3570,12 +3570,6 @@ def runLLMTestlistOnPlatformImpl(pipeline, platform, testList, config=VANILLA_CO
}
}

trtllm_utils.llmExecStepWithRetry(pipeline, script: "mkdir -p /opt/tritonserver/backends/tensorrtllm")
def isAarch64 = config.contains("aarch64")
if (!isAarch64) {
trtllm_utils.llmExecStepWithRetry(pipeline, script: "cd ${llmPath} && cp TensorRT-LLM/triton_backend/inflight_batcher_llm/libtriton_tensorrtllm.so /opt/tritonserver/backends/tensorrtllm/")
trtllm_utils.llmExecStepWithRetry(pipeline, script: "cd ${llmPath} && cp TensorRT-LLM/triton_backend/inflight_batcher_llm/trtllmExecutorWorker /opt/tritonserver/backends/tensorrtllm/")
}
trtllm_utils.llmExecStepWithRetry(pipeline, script: "git config --global --add safe.directory \"*\"")
} // timeout 45 min
}
Expand Down Expand Up @@ -4484,8 +4478,8 @@ def launchTestJobs(pipeline, testFilter)
// "A30-TensorRT-Post-Merge-6": ["a30", "l0_a30", 6, 6],
"A30-CPP-Post-Merge-1": ["a30", "l0_a30", 1, 2],
"A30-CPP-Post-Merge-2": ["a30", "l0_a30", 2, 2],
"A30-Triton-Post-Merge-1": ["a30", "l0_a30", 1, 2],
"A30-Triton-Post-Merge-2": ["a30", "l0_a30", 2, 2],
// "A30-Triton-Post-Merge-1": ["a30", "l0_a30", 1, 2],
// "A30-Triton-Post-Merge-2": ["a30", "l0_a30", 2, 2],
// "A100X-TensorRT-Post-Merge-1": ["a100x", "l0_a100", 1, 6],
// "A100X-TensorRT-Post-Merge-2": ["a100x", "l0_a100", 2, 6],
// "A100X-TensorRT-Post-Merge-3": ["a100x", "l0_a100", 3, 6],
Expand Down
10 changes: 5 additions & 5 deletions jenkins/current_image_tags.properties
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,8 @@
# images are adopted from PostMerge pipelines, the abbreviated commit hash is used instead.
IMAGE_NAME=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm

LLM_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-26.04-py3-x86_64-ubuntu24.04-trt10.16.1.11-skip-tritondevel-202606292100-15694
LLM_SBSA_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-26.04-py3-aarch64-ubuntu24.04-trt10.16.1.11-skip-tritondevel-202606292100-15694
LLM_ROCKYLINUX8_PY310_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-rocky8-x86_64-rocky8-py310-trt10.16.1.11-skip-tritondevel-202606292100-15694
LLM_ROCKYLINUX8_PY312_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-rocky8-x86_64-rocky8-py312-trt10.16.1.11-skip-tritondevel-202606292100-15694
LLM_SBSA_WHEEL_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-ubuntu24.04-sbsa-ubuntu24.04-py312-trt10.16.1.11-skip-tritondevel-202606292100-15694
LLM_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-26.04-py3-x86_64-ubuntu24.04-trt10.16.1.11-skip-tritondevel-202607072201-15907
LLM_SBSA_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-26.04-py3-aarch64-ubuntu24.04-trt10.16.1.11-skip-tritondevel-202607072201-15907
LLM_ROCKYLINUX8_PY310_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-rocky8-x86_64-rocky8-py310-trt10.16.1.11-skip-tritondevel-202607072201-15907
LLM_ROCKYLINUX8_PY312_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-rocky8-x86_64-rocky8-py312-trt10.16.1.11-skip-tritondevel-202607072201-15907
LLM_SBSA_WHEEL_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-13.2.1-devel-ubuntu24.04-sbsa-ubuntu24.04-py312-trt10.16.1.11-skip-tritondevel-202607072201-15907
33 changes: 0 additions & 33 deletions legacy-files.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1317,52 +1317,19 @@ tests/unittest/utils/test_prebuilt_whl_cpp_extensions.py
tests/unittest/utils/test_util.py
tests/unittest/utils/torch_ref.py
tests/unittest/utils/util.py
triton_backend/all_models/disaggregated_serving/disaggregated_serving_bls/1/model.py
triton_backend/all_models/gpt/postprocessing/1/model.py
triton_backend/all_models/gpt/preprocessing/1/model.py
triton_backend/all_models/gpt/tensorrt_llm/1/model.py
triton_backend/all_models/inflight_batcher_llm/postprocessing/1/model.py
triton_backend/all_models/inflight_batcher_llm/preprocessing/1/model.py
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm/1/model.py
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/decode.py
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/lib/triton_decoder.py
triton_backend/all_models/inflight_batcher_llm/tensorrt_llm_bls/1/model.py
triton_backend/all_models/llmapi/tensorrt_llm/1/helpers.py
triton_backend/all_models/llmapi/tensorrt_llm/1/model.py
triton_backend/all_models/multimodal/multimodal_encoders/1/model.py
triton_backend/all_models/multimodal/multimodal_encoders/1/multimodal_utils.py
triton_backend/all_models/tests/test_decode.py
triton_backend/all_models/tests/test_llmapi_python_backend.py
triton_backend/all_models/tests/test_multi_image_preprocess.py
triton_backend/all_models/tests/test_multimodal_encoders.py
triton_backend/all_models/tests/test_python_backend.py
triton_backend/all_models/tests/test_triton_decoder.py
triton_backend/all_models/whisper/whisper_bls/1/fbank.py
triton_backend/all_models/whisper/whisper_bls/1/model.py
triton_backend/all_models/whisper/whisper_bls/1/tokenizer.py
triton_backend/ci/L0_backend_trtllm/base_metrics_verification_tests.py
triton_backend/ci/L0_backend_trtllm/custom_metrics_verification_tests.py
triton_backend/inflight_batcher_llm/client/__init__.py
triton_backend/inflight_batcher_llm/client/e2e_grpc_speculative_decoding_client.py
triton_backend/inflight_batcher_llm/client/end_to_end_grpc_client.py
triton_backend/inflight_batcher_llm/client/inflight_batcher_llm_client.py
triton_backend/scripts/launch_triton_server.py
triton_backend/tools/__init__.py
triton_backend/tools/fill_template.py
triton_backend/tools/gpt/benchmark_core_model.py
triton_backend/tools/gpt/client.py
triton_backend/tools/gpt/client_async.py
triton_backend/tools/gpt/end_to_end_test.py
triton_backend/tools/gpt/gen_input_data.py
triton_backend/tools/inflight_batcher_llm/benchmark_core_model.py
triton_backend/tools/inflight_batcher_llm/end_to_end_test.py
triton_backend/tools/inflight_batcher_llm/speculative_decoding_test.py
triton_backend/tools/inflight_batcher_llm/test_max_queue_size.py
triton_backend/tools/llmapi_client.py
triton_backend/tools/multimodal/client.py
triton_backend/tools/tests/__init__.py
triton_backend/tools/tests/test_fill_template.py
triton_backend/tools/tests/test_llmapi_cancel.py
triton_backend/tools/utils/__init__.py
triton_backend/tools/utils/utils.py
triton_backend/tools/whisper/client.py
Loading
Loading