Remove pybind from CI/CD

Signed-off-by: Linda-Stadter <[email protected]>

Author: Linda-Stadter

.github/CODEOWNERS

@@ -185,8 +185,6 @@ docs/source/performance/perf-benchmarking.md @NVIDIA/trtllm-bench-reviewers
 /tensorrt_llm/_torch/pyexecutor/resource_manager.py @NVIDIA/trt-llm-kv-cache-manager-devs
 /cpp/tensorrt_llm/nanobind/batch_manager/kvCacheManager.h @NVIDIA/trt-llm-kv-cache-manager-devs
 /cpp/tensorrt_llm/nanobind/batch_manager/kvCacheManager.cpp @NVIDIA/trt-llm-kv-cache-manager-devs
-/cpp/tensorrt_llm/pybind/batch_manager/kvCacheManager.h @NVIDIA/trt-llm-kv-cache-manager-devs
-/cpp/tensorrt_llm/pybind/batch_manager/kvCacheManager.cpp @NVIDIA/trt-llm-kv-cache-manager-devs
 
 # The rule below requires that any PR modifying public APIs must be approved by at least one member
 # of the NVIDIA/trt-llm-committed-api-review-committee or NVIDIA/trt-llm-noncommitted-api-review-committee team.

docs/source/legacy/advanced/executor.md

@@ -141,7 +141,7 @@ Two C++ examples are provided that shows how to use the Executor API and can be
 
 ## Python Bindings for the Executor API
 
-Python bindings for the Executor API are also available to use the Executor API from Python. The Python bindings are defined in [bindings.cpp](source:cpp/tensorrt_llm/pybind/executor/bindings.cpp) and once built, are available in package `tensorrt_llm.bindings.executor`. Running `'help('tensorrt_llm.bindings.executor')` in a Python interpreter will provide an overview of the classes available.
+Python bindings for the Executor API are also available to use the Executor API from Python. The Python bindings are defined in [bindings.cpp](source:cpp/tensorrt_llm/nanobind/executor/bindings.cpp) and once built, are available in package `tensorrt_llm.bindings.executor`. Running `'help('tensorrt_llm.bindings.executor')` in a Python interpreter will provide an overview of the classes available.
 
 In addition, three Python examples are provided to demonstrate how to use the Python bindings to the Executor API for single and multi-GPU models. They can be found in [`examples/bindings`](source:examples/bindings).
 

jenkins/Build.groovy

@@ -49,11 +49,6 @@ def CONFIG_LINUX_AARCH64 = "linux_aarch64"
 @Field
 def CONFIG_LINUX_AARCH64_LLVM = "linux_aarch64_LLVM"
 
-@Field
-def CONFIG_LINUX_X86_64_PYBIND = "linux_x86_64_Pybind"
-
-@Field
-def CONFIG_LINUX_AARCH64_PYBIND = "linux_aarch64_Pybind"
 
 @Field
 def BUILD_CONFIGS = [
@@ -64,11 +59,6 @@ def BUILD_CONFIGS = [
     (TARNAME) : "TensorRT-LLM.tar.gz",
     (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
   ],
-  (CONFIG_LINUX_X86_64_PYBIND) : [
-    (WHEEL_EXTRA_ARGS) : "--binding_type pybind --extra-cmake-vars ENABLE_MULTI_DEVICE=1 --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars NIXL_ROOT=/opt/nvidia/nvda_nixl --extra-cmake-vars MOONCAKE_ROOT=/usr/local/Mooncake --micro_benchmarks",
-    (TARNAME) : "pybind-TensorRT-LLM.tar.gz",
-    (WHEEL_ARCHS): "80-real;86-real;89-real;90-real;100-real;103-real;120-real",
-  ],
   (CONFIG_LINUX_X86_64_SINGLE_DEVICE) : [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars ENABLE_MULTI_DEVICE=0 --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars ENABLE_UCX=0 --micro_benchmarks",
     (TARNAME) : "single-device-TensorRT-LLM.tar.gz",
@@ -85,12 +75,6 @@ def BUILD_CONFIGS = [
     (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
     (BUILD_JOBS_FOR_CONFIG): "4", // TODO: Remove after fix the build OOM issue on SBSA
   ],
-  (CONFIG_LINUX_AARCH64_PYBIND): [
-    (WHEEL_EXTRA_ARGS) : "--binding_type pybind --extra-cmake-vars WARNING_IS_ERROR=ON --extra-cmake-vars NIXL_ROOT=/opt/nvidia/nvda_nixl --extra-cmake-vars MOONCAKE_ROOT=/usr/local/Mooncake",
-    (TARNAME) : "pybind-TensorRT-LLM-GH200.tar.gz",
-    (WHEEL_ARCHS): "90-real;100-real;103-real;120-real",
-    (BUILD_JOBS_FOR_CONFIG): "4", // TODO: Remove after fix the build OOM issue on SBSA
-  ],
   (CONFIG_LINUX_AARCH64_LLVM) : [
     (WHEEL_EXTRA_ARGS) : "--extra-cmake-vars WARNING_IS_ERROR=ON -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_CUDA_HOST_COMPILER=clang -DCMAKE_LINKER_TYPE=LLD",
     (TARNAME) : "llvm-TensorRT-LLM-GH200.tar.gz",
@@ -549,8 +533,6 @@ def launchStages(pipeline, cpu_arch, enableFailFast, globalVars)
             pipeline, cpu_arch == AARCH64_TRIPLE ? CONFIG_LINUX_AARCH64 : CONFIG_LINUX_X86_64_VANILLA),
         "Build TRT-LLM LLVM": [LLM_DOCKER_IMAGE] + prepareLLMBuild(
             pipeline, cpu_arch == AARCH64_TRIPLE ? CONFIG_LINUX_AARCH64_LLVM : CONFIG_LINUX_X86_64_LLVM),
-        "Build TRT-LLM Pybind": [LLM_DOCKER_IMAGE] + prepareLLMBuild(
-            pipeline, cpu_arch == AARCH64_TRIPLE ? CONFIG_LINUX_AARCH64_PYBIND : CONFIG_LINUX_X86_64_PYBIND),
     ]
 
     if (cpu_arch == X86_64_TRIPLE) {

jenkins/L0_MergeRequest.groovy

@@ -694,7 +694,7 @@ def getMultiGpuFileChanged(pipeline, testFilter, globalVars)
         "cpp/tensorrt_llm/plugins/gptAttentionPlugin/gptAttentionPlugin.cpp",
         "cpp/tensorrt_llm/plugins/gptAttentionPlugin/gptAttentionPlugin.h",
         "cpp/tensorrt_llm/plugins/ncclPlugin/",
-        "cpp/tensorrt_llm/pybind/",
+        "cpp/tensorrt_llm/nanobind/",
         "cpp/tensorrt_llm/runtime/ipcUtils.cpp",
         "cpp/tensorrt_llm/runtime/ncclCommunicator.cpp",
         "cpp/tensorrt_llm/runtime/utils/mpiUtils.cpp",

jenkins/L0_Test.groovy

@@ -65,17 +65,13 @@ def LLVM_CONFIG = "LLVM"
 @Field
 def LINUX_AARCH64_CONFIG = "linux_aarch64"
 
-@Field
-def PYBIND_CONFIG = "Pybind"
-
 @Field
 def BUILD_CONFIGS = [
   // Vanilla TARNAME is used for packaging in runLLMPackage
   (VANILLA_CONFIG) : [(TARNAME) : "TensorRT-LLM.tar.gz"],
   (SINGLE_DEVICE_CONFIG) : [(TARNAME) : "single-device-TensorRT-LLM.tar.gz"],
   (LLVM_CONFIG) : [(TARNAME) : "llvm-TensorRT-LLM.tar.gz"],
   (LINUX_AARCH64_CONFIG) : [(TARNAME) : "TensorRT-LLM-GH200.tar.gz"],
-  (PYBIND_CONFIG) : [(TARNAME) : "pybind-TensorRT-LLM.tar.gz"],
 ]
 
 // TODO: Move common variables to an unified location
@@ -3147,7 +3143,6 @@ def launchTestJobs(pipeline, testFilter)
         "A10-TensorRT-3": ["a10", "l0_a10", 3, 5],
         "A10-TensorRT-4": ["a10", "l0_a10", 4, 5],
         "A10-TensorRT-5": ["a10", "l0_a10", 5, 5],
-        "A10-Pybind": ["a10", "l0_a10_pybind", 1, 1],
         "A30-Triton-1": ["a30", "l0_a30", 1, 1],
         "A30-PyTorch-1": ["a30", "l0_a30", 1, 2],
         "A30-PyTorch-2": ["a30", "l0_a30", 2, 2],
@@ -3241,9 +3236,6 @@ def launchTestJobs(pipeline, testFilter)
         if (key.contains("llvm")) {
             config = LLVM_CONFIG
         }
-        if (key.contains("Pybind")) {
-            config = PYBIND_CONFIG
-        }
         runLLMTestlistOnPlatform(pipeline, values[0], values[1], config, key.contains("-Perf-"), key, values[2], values[3])
     }]]}
     fullSet = parallelJobs.keySet()

tests/integration/test_lists/test-db/l0_a10.yml

@@ -275,19 +275,3 @@ l0_a10:
       backend: fmha
   tests:
   - test_fmha.py::test_fmha TIMEOUT (90)
-l0_a10_pybind:
-- condition:
-    ranges:
-      system_gpu_count:
-        gte: 1
-        lte: 1
-    wildcards:
-      gpu:
-      - '*a10*'
-      linux_distribution_name: ubuntu*
-    terms:
-      stage: pre_merge
-  tests:
-  - unittest/bindings
-  - test_e2e.py::test_openai_chat_example[trt]
-  - test_e2e.py::test_openai_chat_example[pytorch] TIMEOUT (90)