mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-05-14 20:48:00 +00:00
[Running CI] [TensorRT EP] support TensorRT 10.3-GA (#21742)
### Description - TensorRT 10.2.0.19 -> 10.3.0.26 ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. -->
This commit is contained in:
jingyanwangms
GitHub
parent
c4ade796d6
commit
c018ba43ef
31 changed files with 84 additions and 85 deletions
|
|
@ -216,7 +216,7 @@
|
|||
"component": {
|
||||
"type": "git",
|
||||
"git": {
|
||||
"commitHash": "f161f95883b4ebd8cb789de5efc67b73c0a6e694",
|
||||
"commitHash": "62bdde2a04fcd53c2409cb895ee18db445b7e755",
|
||||
"repositoryUrl": "https://github.com/onnx/onnx-tensorrt.git"
|
||||
},
|
||||
"comments": "onnx_tensorrt"
|
||||
|
|
|
|||
|
|
@ -37,8 +37,8 @@ mimalloc;https://github.com/microsoft/mimalloc/archive/refs/tags/v2.1.1.zip;d5ee
|
|||
mp11;https://github.com/boostorg/mp11/archive/refs/tags/boost-1.82.0.zip;9bc9e01dffb64d9e0773b2e44d2f22c51aace063
|
||||
neural_speed;https://github.com/intel/neural-speed/archive/refs/tags/v0.3.zip;5ec64e3071edc7347ebd8a81679cf06e2bb9b851
|
||||
onnx;https://github.com/onnx/onnx/archive/refs/tags/v1.16.1.zip;2eb9198bb352757d5ff13977cbe0634898e0837c
|
||||
#use the latest commit of 10.2-GA
|
||||
onnx_tensorrt;https://github.com/onnx/onnx-tensorrt/archive/f161f95883b4ebd8cb789de5efc67b73c0a6e694.zip;2148d0c79a171abf2b9451f3bfec164e85caf2ef
|
||||
#use the latest commit of 10.3-GA
|
||||
onnx_tensorrt;https://github.com/onnx/onnx-tensorrt/archive/62bdde2a04fcd53c2409cb895ee18db445b7e755.zip;980a455b07dfa67aa70b9e49d37dd9d4cdf690a0
|
||||
protobuf;https://github.com/protocolbuffers/protobuf/archive/refs/tags/v21.12.zip;7cf2733949036c7d52fda017badcab093fe73bfa
|
||||
protoc_win64;https://github.com/protocolbuffers/protobuf/releases/download/v21.12/protoc-21.12-win64.zip;b4521f7ada5b260380f94c4bd7f1b7684c76969a
|
||||
protoc_win32;https://github.com/protocolbuffers/protobuf/releases/download/v21.12/protoc-21.12-win32.zip;3688010318192c46ce73213cdfb6b3e5656da874
|
||||
|
|
|
|||
|
|
@ -17,8 +17,8 @@ from typing import List, Optional
|
|||
TRT_DOCKER_FILES = {
|
||||
"8.6.cuda_11_8_cudnn_8": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda11_8_tensorrt8_6",
|
||||
"8.6.cuda_12_3_cudnn_9": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda12_3_tensorrt8_6",
|
||||
"10.2.cuda_11_8_cudnn_8": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda11_tensorrt10",
|
||||
"10.2.cuda_12_5_cudnn_9": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda12_tensorrt10",
|
||||
"10.3.cuda_11_8_cudnn_8": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda11_tensorrt10",
|
||||
"10.3.cuda_12_5_cudnn_9": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_cuda12_tensorrt10",
|
||||
"BIN": "tools/ci_build/github/linux/docker/Dockerfile.ubuntu_tensorrt_bin",
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -43,7 +43,7 @@ variables:
|
|||
- name: docker_base_image
|
||||
value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20240531.1
|
||||
- name: linux_trt_version
|
||||
value: 10.2.0.19-1.cuda11.8
|
||||
value: 10.3.0.26-1.cuda11.8
|
||||
- name: Repository
|
||||
value: 'onnxruntimecuda11manylinuxbuild'
|
||||
|
||||
|
|
|
|||
|
|
@ -83,7 +83,7 @@ variables:
|
|||
value: 11.8
|
||||
|
||||
- name: win_trt_home
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8
|
||||
- name: win_cuda_home
|
||||
value: $(Agent.TempDirectory)\v11.8
|
||||
|
||||
|
|
|
|||
|
|
@ -68,9 +68,9 @@ variables:
|
|||
value: nvidia/cuda:12.2.2-cudnn8-devel-ubi8
|
||||
- name: win_trt_home
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5
|
||||
value: $(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5
|
||||
- name: win_cuda_home
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: $(Agent.TempDirectory)\v11.8
|
||||
|
|
|
|||
|
|
@ -43,9 +43,9 @@ variables:
|
|||
value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20240719.1
|
||||
- name: linux_trt_version
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: 10.2.0.19-1.cuda11.8
|
||||
value: 10.3.0.26-1.cuda11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: 10.2.0.19-1.cuda12.5
|
||||
value: 10.3.0.26-1.cuda12.5
|
||||
|
||||
jobs:
|
||||
- job: Linux_Build
|
||||
|
|
|
|||
|
|
@ -12,8 +12,8 @@ parameters:
|
|||
values:
|
||||
- 8.6.cuda_11_8_cudnn_8
|
||||
- 8.6.cuda_12_3_cudnn_9
|
||||
- 10.2.cuda_11_8_cudnn_8
|
||||
- 10.2.cuda_12_5_cudnn_9
|
||||
- 10.3.cuda_11_8_cudnn_8
|
||||
- 10.3.cuda_12_5_cudnn_9
|
||||
- BIN
|
||||
|
||||
- name: UseTensorrtOssParser
|
||||
|
|
@ -23,7 +23,7 @@ parameters:
|
|||
|
||||
- name: ModelGroups
|
||||
type: object
|
||||
default:
|
||||
default:
|
||||
- "onnx-zoo-models"
|
||||
- "partner-models"
|
||||
|
||||
|
|
@ -55,7 +55,7 @@ jobs:
|
|||
clean: all
|
||||
timeoutInMinutes: 3000
|
||||
pool: 'Onnxruntime-Linux-GPU-TensorRT-Perf'
|
||||
|
||||
|
||||
variables:
|
||||
- name: branchName
|
||||
${{ if not(eq(variables['ortBranchOverride'], '')) }}:
|
||||
|
|
@ -82,7 +82,7 @@ jobs:
|
|||
|
||||
- name: image
|
||||
value: ort-image-$(Build.BuildId)
|
||||
|
||||
|
||||
- name: parser
|
||||
${{ if eq(parameters.UseTensorrtOssParser, true) }}:
|
||||
value: --use_tensorrt_oss_parser $(parameters.UseTensorrtOssParser) }}
|
||||
|
|
@ -98,13 +98,13 @@ jobs:
|
|||
- script: 'python3 $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build/build_image.py -r $(Build.SourcesDirectory) -i $(image) -b $(branchName) -t $(trtVersion) -a 75 --install_bin --tar_cuda_version=$(tarCudaVersion) --trt_bins_dir=.'
|
||||
displayName: 'Install TensorRT $(tarTrtVersion) from binaries and build latest ORT Image'
|
||||
workingDirectory: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build'
|
||||
|
||||
# Build ORT with TensorRT built-in parser
|
||||
|
||||
# Build ORT with TensorRT built-in parser
|
||||
- ${{ if and(ne(parameters.TrtVersion, 'BIN'), eq(parameters.UseTensorrtOssParser, false)) }}:
|
||||
- script: 'python3 $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build/build_image.py -r $(Build.SourcesDirectory) -i $(image) -b $(branchName) -t $(trtVersion) -a 75'
|
||||
displayName: 'Build latest ORT Image with TensorRT built-in parser'
|
||||
workingDirectory: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build'
|
||||
|
||||
|
||||
# Build ORT with certain version of TensorRT OSS parser
|
||||
- ${{ if and(ne(parameters.TrtVersion, 'BIN'), eq(parameters.UseTensorrtOssParser, true), ne(variables.onnxTensorrtCommitId, '')) }}:
|
||||
- script: 'python3 $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build/build_image.py -r $(Build.SourcesDirectory) -i $(image) -b $(branchName) -t $(trtVersion) -a 75 -o $(onnxTensorrtCommitId) --use_tensorrt_oss_parser'
|
||||
|
|
@ -116,7 +116,7 @@ jobs:
|
|||
- script: 'python3 $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build/build_image.py -r $(Build.SourcesDirectory) -i $(image) -b $(branchName) -t $(trtVersion) -a 75 --use_tensorrt_oss_parser'
|
||||
displayName: 'Build latest ORT Image with TensorRT OSS parser'
|
||||
workingDirectory: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/build'
|
||||
|
||||
|
||||
- ${{ if eq(parameters.MemTest, true) }}:
|
||||
- script: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/mem_test/run_mem_test_docker.sh -d $(image) -p $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/mem_test/ -w /code/ -l false -c ${{ parameters.ConcurrencyTest }}'
|
||||
displayName: 'Run Memory Test and Concurrency Test'
|
||||
|
|
@ -127,12 +127,12 @@ jobs:
|
|||
displayName: '${{option}} perf'
|
||||
workingDirectory: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/'
|
||||
|
||||
# Prepare and Publish Artifacts
|
||||
|
||||
# Prepare and Publish Artifacts
|
||||
|
||||
- script: 'mkdir $(Build.SourcesDirectory)/Artifact'
|
||||
displayName: 'Prepare Artifacts Directory'
|
||||
workingDirectory: '$(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/'
|
||||
condition: always()
|
||||
condition: always()
|
||||
|
||||
- ${{ if not(eq(length(parameters.ModelGroups), 0)) }}:
|
||||
- script: 'cp -r $(Build.SourcesDirectory)/onnxruntime/python/tools/tensorrt/perf/result/ $(Build.SourcesDirectory)/Artifact'
|
||||
|
|
@ -154,7 +154,7 @@ jobs:
|
|||
curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash
|
||||
}
|
||||
displayName: 'Check and Install Azure CLI'
|
||||
|
||||
|
||||
- task: AzureCLI@2
|
||||
displayName: 'Parse Memory & Concurrency Test Records and Sync'
|
||||
inputs:
|
||||
|
|
@ -197,5 +197,5 @@ jobs:
|
|||
- template: templates/component-governance-component-detection-steps.yml
|
||||
parameters :
|
||||
condition : 'succeeded'
|
||||
|
||||
|
||||
- template: templates/clean-agent-build-directory-step.yml
|
||||
|
|
|
|||
|
|
@ -61,7 +61,7 @@ stages:
|
|||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
DockerBuildArgs: "
|
||||
--build-arg BASEIMAGE=nvidia/cuda:12.2.2-devel-ubuntu20.04
|
||||
--build-arg TRT_VERSION=10.2.0.19-1+cuda12.5
|
||||
--build-arg TRT_VERSION=10.3.0.26-1+cuda12.5
|
||||
--build-arg BUILD_UID=$( id -u )
|
||||
"
|
||||
${{ else }}:
|
||||
|
|
|
|||
|
|
@ -226,7 +226,7 @@ stages:
|
|||
BuildConfig: 'RelWithDebInfo'
|
||||
EnvSetupScript: setup_env_trt.bat
|
||||
buildArch: x64
|
||||
additionalBuildFlags: --enable_pybind --build_java --build_nodejs --use_cuda --cuda_home="$(Agent.TempDirectory)\v11.8" --enable_cuda_profiling --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=86
|
||||
additionalBuildFlags: --enable_pybind --build_java --build_nodejs --use_cuda --cuda_home="$(Agent.TempDirectory)\v11.8" --enable_cuda_profiling --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cmake_extra_defines CMAKE_CUDA_ARCHITECTURES=86
|
||||
msbuildPlatform: x64
|
||||
isX86: false
|
||||
job_name_suffix: x64_RelWithDebInfo
|
||||
|
|
|
|||
|
|
@ -55,7 +55,7 @@ stages:
|
|||
python_wheel_suffix: '_gpu'
|
||||
timeout: 480
|
||||
docker_base_image: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20240531.1
|
||||
trt_version: '10.2.0.19-1.cuda11.8'
|
||||
trt_version: '10.3.0.26-1.cuda11.8'
|
||||
cuda_version: '11.8'
|
||||
|
||||
|
||||
|
|
|
|||
|
|
@ -49,9 +49,9 @@ jobs:
|
|||
value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20240719.1
|
||||
- name: linux_trt_version
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: 10.2.0.19-1.cuda11.8
|
||||
value: 10.3.0.26-1.cuda11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: 10.2.0.19-1.cuda12.5
|
||||
value: 10.3.0.26-1.cuda12.5
|
||||
pool: ${{ parameters.machine_pool }}
|
||||
steps:
|
||||
- checkout: self
|
||||
|
|
@ -99,7 +99,7 @@ jobs:
|
|||
Dockerfile: tools/ci_build/github/linux/docker/Dockerfile.manylinux2_28_cuda
|
||||
Context: tools/ci_build/github/linux/docker
|
||||
DockerBuildArgs: "
|
||||
--network=host
|
||||
--network=host
|
||||
--build-arg BASEIMAGE=${{ variables.docker_base_image }}
|
||||
--build-arg TRT_VERSION=${{ variables.linux_trt_version }}
|
||||
--build-arg BUILD_UID=$( id -u )
|
||||
|
|
@ -115,4 +115,4 @@ jobs:
|
|||
arguments: -d GPU -c ${{parameters.cmake_build_type}} -i onnxruntimecuda${{ replace(parameters.CudaVersion, '.', '') }}xtrt86buildx86_64 -u 12.2
|
||||
- template: ../../templates/component-governance-component-detection-steps.yml
|
||||
parameters:
|
||||
condition: 'succeeded'
|
||||
condition: 'succeeded'
|
||||
|
|
|
|||
|
|
@ -80,9 +80,9 @@ stages:
|
|||
|
||||
- name: linux_trt_version
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: 10.2.0.19-1.cuda11.8
|
||||
value: 10.3.0.26-1.cuda11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: 10.2.0.19-1.cuda12.5
|
||||
value: 10.3.0.26-1.cuda12.5
|
||||
steps:
|
||||
- checkout: self
|
||||
clean: true
|
||||
|
|
@ -149,9 +149,9 @@ stages:
|
|||
value: '12'
|
||||
- name: linux_trt_version
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: 10.2.0.19-1.cuda11.8
|
||||
value: 10.3.0.26-1.cuda11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: 10.2.0.19-1.cuda12.5
|
||||
value: 10.3.0.26-1.cuda12.5
|
||||
steps:
|
||||
- checkout: self # due to checkout multiple repos, the root directory is $(Build.SourcesDirectory)/onnxruntime
|
||||
submodules: false
|
||||
|
|
|
|||
|
|
@ -65,9 +65,9 @@ stages:
|
|||
SpecificArtifact: ${{ parameters.SpecificArtifact }}
|
||||
BuildId: ${{ parameters.BuildId }}
|
||||
${{ if eq(parameters.cuda_version, '11.8') }}:
|
||||
EP_BUILD_FLAGS: --enable_lto --use_tensorrt --tensorrt_home=$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8 --cuda_home=$(Agent.TempDirectory)\v11.8 --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --enable_lto --use_tensorrt --tensorrt_home=$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8 --cuda_home=$(Agent.TempDirectory)\v11.8 --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
${{ if eq(parameters.cuda_version, '12.2') }}:
|
||||
EP_BUILD_FLAGS: --enable_lto --use_tensorrt --tensorrt_home=$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5 --cuda_home=$(Agent.TempDirectory)\v12.2 --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --enable_lto --use_tensorrt --tensorrt_home=$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5 --cuda_home=$(Agent.TempDirectory)\v12.2 --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
|
||||
- ${{ if eq(parameters.enable_linux_gpu, true) }}:
|
||||
- template: ../templates/py-linux-gpu.yml
|
||||
|
|
@ -79,7 +79,7 @@ stages:
|
|||
cuda_version: ${{ parameters.cuda_version }}
|
||||
${{ if eq(parameters.cuda_version, '11.8') }}:
|
||||
docker_base_image: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20240531.1
|
||||
trt_version: 10.2.0.19-1.cuda11.8
|
||||
trt_version: 10.3.0.26-1.cuda11.8
|
||||
${{ if eq(parameters.cuda_version, '12.2') }}:
|
||||
docker_base_image: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda12_x64_ubi8_gcc12:20240719.1
|
||||
trt_version: 10.2.0.19-1.cuda12.5
|
||||
trt_version: 10.3.0.26-1.cuda12.5
|
||||
|
|
|
|||
|
|
@ -11,7 +11,7 @@ steps:
|
|||
packageType: upack
|
||||
feed: '/7424c8e4-5c62-490e-95c4-79446f31017c'
|
||||
definition: '517c4f6f-5437-4392-a70d-4f15ec5be2f0'
|
||||
version: 1.0.175
|
||||
version: 1.0.178
|
||||
downloadPath: $(Build.BinariesDirectory)/deps
|
||||
|
||||
# The private ADO project
|
||||
|
|
@ -22,7 +22,7 @@ steps:
|
|||
packageType: upack
|
||||
feed: '/4c7631f5-24c0-4307-8822-1aa8f180c325'
|
||||
definition: 'fd9dd5ad-b73e-4678-890e-edcf680dbc1a'
|
||||
version: 1.0.175
|
||||
version: 1.0.178
|
||||
downloadPath: $(Build.BinariesDirectory)/deps
|
||||
|
||||
# You can add more ADO accounts at here.
|
||||
|
|
|
|||
|
|
@ -13,10 +13,10 @@ parameters:
|
|||
- 12.2
|
||||
- name: TrtVersion
|
||||
type: string
|
||||
default: '10.2.0.19'
|
||||
default: '10.3.0.26'
|
||||
values:
|
||||
- 8.6.1.6
|
||||
- 10.2.0.19
|
||||
- 10.3.0.26
|
||||
|
||||
steps:
|
||||
- ${{ if eq(parameters.DownloadCUDA, true) }}:
|
||||
|
|
@ -42,7 +42,7 @@ steps:
|
|||
- powershell: |
|
||||
Write-Host "##vso[task.setvariable variable=trtCudaVersion;]12.0"
|
||||
displayName: Set trtCudaVersion
|
||||
- ${{ if and(eq(parameters.CudaVersion, '12.2'), eq(parameters.TrtVersion, '10.2.0.19')) }}:
|
||||
- ${{ if and(eq(parameters.CudaVersion, '12.2'), eq(parameters.TrtVersion, '10.3.0.26')) }}:
|
||||
- powershell: |
|
||||
Write-Host "##vso[task.setvariable variable=trtCudaVersion;]12.5"
|
||||
displayName: Set trtCudaVersion
|
||||
|
|
|
|||
|
|
@ -24,12 +24,12 @@ steps:
|
|||
displayName: 'Download Secondary CUDA SDK v${{ parameters.SecondaryCUDAVersion }}'
|
||||
- ${{ if eq(parameters.DownloadTRT, 'true') }}:
|
||||
- powershell: |
|
||||
azcopy.exe cp --recursive "https://lotusscus.blob.core.windows.net/models/local/TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" $(Agent.TempDirectory)
|
||||
displayName: 'Download TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8'
|
||||
azcopy.exe cp --recursive "https://lotusscus.blob.core.windows.net/models/local/TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" $(Agent.TempDirectory)
|
||||
displayName: 'Download TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8'
|
||||
- powershell: |
|
||||
azcopy.exe cp --recursive "https://lotusscus.blob.core.windows.net/models/local/TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5" $(Agent.TempDirectory)
|
||||
displayName: 'Download TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5'
|
||||
|
||||
azcopy.exe cp --recursive "https://lotusscus.blob.core.windows.net/models/local/TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5" $(Agent.TempDirectory)
|
||||
displayName: 'Download TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5'
|
||||
|
||||
- task: BatchScript@1
|
||||
displayName: 'setup env'
|
||||
inputs:
|
||||
|
|
|
|||
|
|
@ -22,10 +22,10 @@ parameters:
|
|||
|
||||
- name: trt_version
|
||||
type: string
|
||||
default: '10.2.0.19-1.cuda11.8'
|
||||
default: '10.3.0.26-1.cuda11.8'
|
||||
values:
|
||||
- 10.2.0.19-1.cuda11.8
|
||||
- 10.2.0.19-1.cuda12.5
|
||||
- 10.3.0.26-1.cuda11.8
|
||||
- 10.3.0.26-1.cuda12.5
|
||||
- name: cuda_version
|
||||
type: string
|
||||
default: '11.8'
|
||||
|
|
|
|||
|
|
@ -18,10 +18,10 @@ parameters:
|
|||
|
||||
- name: trt_version
|
||||
type: string
|
||||
default: '10.2.0.19-1.cuda11.8'
|
||||
default: '10.3.0.26-1.cuda11.8'
|
||||
values:
|
||||
- 10.2.0.19-1.cuda11.8
|
||||
- 10.2.0.19-1.cuda12.5
|
||||
- 10.3.0.26-1.cuda11.8
|
||||
- 10.3.0.26-1.cuda12.5
|
||||
- name: cuda_version
|
||||
type: string
|
||||
default: '11.8'
|
||||
|
|
|
|||
|
|
@ -381,7 +381,7 @@ stages:
|
|||
variables:
|
||||
CUDA_VERSION: '11.8'
|
||||
buildArch: x64
|
||||
EpBuildFlags: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_version=$(CUDA_VERSION) --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v$(CUDA_VERSION)" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
|
||||
EpBuildFlags: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_version=$(CUDA_VERSION) --cuda_home="C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v$(CUDA_VERSION)" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=37;50;52;60;61;70;75;80"
|
||||
EnvSetupScript: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
VSGenerator: 'Visual Studio 17 2022'
|
||||
|
|
|
|||
|
|
@ -266,7 +266,7 @@ stages:
|
|||
SymbolServerType: TeamServices
|
||||
SymbolExpirationInDays: 3650
|
||||
SymbolsArtifactName: 'win_cpu_$(PythonVersion)_$(buildArch)_$(Build.BuildNumber)'
|
||||
|
||||
|
||||
- task: TSAUpload@2
|
||||
displayName: 'TSA upload'
|
||||
condition: and(and (succeeded(), and(eq(variables['buildArch'], 'x64'), eq(variables['PythonVersion'], '3.8'))), eq(variables['Build.SourceBranch'], 'refs/heads/main'))
|
||||
|
|
@ -288,7 +288,7 @@ stages:
|
|||
parameters:
|
||||
MACHINE_POOL: 'onnxruntime-Win2022-GPU-A10'
|
||||
PYTHON_VERSION: '3.8'
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
ENV_SETUP_SCRIPT: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
publish_symbols: ${{ parameters.publish_symbols }}
|
||||
|
|
@ -298,7 +298,7 @@ stages:
|
|||
parameters:
|
||||
MACHINE_POOL: 'onnxruntime-Win2022-GPU-A10'
|
||||
PYTHON_VERSION: '3.9'
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
ENV_SETUP_SCRIPT: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
publish_symbols: ${{ parameters.publish_symbols }}
|
||||
|
|
@ -308,7 +308,7 @@ stages:
|
|||
parameters:
|
||||
MACHINE_POOL: 'onnxruntime-Win2022-GPU-A10'
|
||||
PYTHON_VERSION: '3.10'
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
ENV_SETUP_SCRIPT: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
publish_symbols: ${{ parameters.publish_symbols }}
|
||||
|
|
@ -318,7 +318,7 @@ stages:
|
|||
parameters:
|
||||
MACHINE_POOL: 'onnxruntime-Win2022-GPU-A10'
|
||||
PYTHON_VERSION: '3.11'
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
ENV_SETUP_SCRIPT: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
publish_symbols: ${{ parameters.publish_symbols }}
|
||||
|
|
@ -328,7 +328,7 @@ stages:
|
|||
parameters:
|
||||
MACHINE_POOL: 'onnxruntime-Win2022-GPU-A10'
|
||||
PYTHON_VERSION: '3.12'
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
EP_BUILD_FLAGS: --use_tensorrt --tensorrt_home="$(Agent.TempDirectory)\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8" --cuda_home="$(Agent.TempDirectory)\v11.8" --cmake_extra_defines "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80"
|
||||
ENV_SETUP_SCRIPT: setup_env_gpu.bat
|
||||
EP_NAME: gpu
|
||||
publish_symbols: ${{ parameters.publish_symbols }}
|
||||
|
|
@ -485,7 +485,7 @@ stages:
|
|||
- template: py-linux.yml
|
||||
parameters:
|
||||
arch: 'x86_64'
|
||||
machine_pool: 'onnxruntime-Ubuntu2204-AMD-CPU'
|
||||
machine_pool: 'onnxruntime-Ubuntu2204-AMD-CPU'
|
||||
extra_build_arg: ${{ parameters.build_py_parameters }}
|
||||
cmake_build_type: ${{ parameters.cmake_build_type }}
|
||||
|
||||
|
|
@ -498,7 +498,7 @@ stages:
|
|||
docker_base_image: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20240531.1
|
||||
extra_build_arg: ${{ parameters.build_py_parameters }}
|
||||
cmake_build_type: ${{ parameters.cmake_build_type }}
|
||||
trt_version: '10.2.0.19-1.cuda11.8'
|
||||
trt_version: '10.3.0.26-1.cuda11.8'
|
||||
cuda_version: '11.8'
|
||||
|
||||
- ${{ if eq(parameters.enable_windows_arm64_qnn, true) }}:
|
||||
|
|
|
|||
|
|
@ -38,9 +38,9 @@ parameters:
|
|||
variables:
|
||||
- name: win_trt_folder
|
||||
${{ if eq(parameters.CudaVersion, '11.8') }}:
|
||||
value: TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8
|
||||
value: TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8
|
||||
${{ if eq(parameters.CudaVersion, '12.2') }}:
|
||||
value: TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5
|
||||
value: TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5
|
||||
|
||||
jobs:
|
||||
- job: 'build'
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@
|
|||
|
||||
# Build base image with required system packages
|
||||
ARG BASEIMAGE=nvidia/cuda:12.5.1-cudnn-devel-ubi8
|
||||
ARG TRT_VERSION=10.2.0.19-1.cuda12.4
|
||||
ARG TRT_VERSION=10.3.0.26-1.cuda12.4
|
||||
FROM $BASEIMAGE AS base
|
||||
ARG TRT_VERSION
|
||||
ENV PATH /opt/python/cp38-cp38/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/src/tensorrt/bin:${PATH}
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@
|
|||
|
||||
# Build base image with required system packages
|
||||
ARG BASEIMAGE=nvidia/cuda:11.8.0-cudnn8-devel-ubi8
|
||||
ARG TRT_VERSION=10.2.0.19-1.cuda11.8
|
||||
ARG TRT_VERSION=10.3.0.26-1.cuda11.8
|
||||
FROM $BASEIMAGE AS base
|
||||
ARG TRT_VERSION
|
||||
ENV PATH /opt/python/cp38-cp38/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/src/tensorrt/bin:${PATH}
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@
|
|||
|
||||
# Build base image with required system packages
|
||||
ARG BASEIMAGE=nvidia/cuda:11.8.0-cudnn8-devel-ubuntu20.04
|
||||
ARG TRT_VERSION=10.2.0.19-1+cuda11.8
|
||||
ARG TRT_VERSION=10.3.0.26-1+cuda11.8
|
||||
ARG LD_LIBRARY_PATH_ARG=/usr/local/lib64:/usr/local/cuda/lib64
|
||||
FROM $BASEIMAGE AS base
|
||||
ARG TRT_VERSION
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@
|
|||
|
||||
# Build base image with required system packages
|
||||
ARG BASEIMAGE=nvidia/cuda:11.8.0-cudnn8-devel-ubuntu20.04
|
||||
ARG TRT_VERSION=10.2.0.19-1+cuda11.8
|
||||
ARG TRT_VERSION=10.3.0.26-1+cuda11.8
|
||||
ARG LD_LIBRARY_PATH_ARG=/usr/local/lib64:/usr/local/cuda/lib64
|
||||
FROM $BASEIMAGE AS base
|
||||
ARG TRT_VERSION
|
||||
|
|
|
|||
|
|
@ -27,11 +27,11 @@ RUN apt-get install -y --no-install-recommends \
|
|||
ln -s /usr/bin/python3 python &&\
|
||||
ln -s /usr/bin/pip3 pip;
|
||||
|
||||
RUN pip install --upgrade pip
|
||||
RUN pip install --upgrade pip
|
||||
RUN pip install psutil setuptools>=68.2.2
|
||||
|
||||
# Install TensorRT
|
||||
RUN version="10.2.0.19-1+cuda11.8" &&\
|
||||
RUN version="10.3.0.26-1+cuda11.8" &&\
|
||||
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/7fa2af80.pub &&\
|
||||
apt-get update &&\
|
||||
apt-get install -y \
|
||||
|
|
@ -98,7 +98,7 @@ RUN if [ -z "$ONNXRUNTIME_COMMIT_ID" ] ; then echo "Building branch ${ONNXRUNTIM
|
|||
git reset --hard ${ONNXRUNTIME_COMMIT_ID} && git submodule update --recursive ; fi
|
||||
|
||||
# Build ORT
|
||||
ENV CUDA_MODULE_LOADING "LAZY"
|
||||
ENV CUDA_MODULE_LOADING "LAZY"
|
||||
ARG PARSER_CONFIG=""
|
||||
RUN /bin/sh build.sh ${PARSER_CONFIG} --parallel --build_shared_lib --cuda_home /usr/local/cuda --cudnn_home /usr/lib/x86_64-linux-gnu/ --use_tensorrt --tensorrt_home /usr/lib/x86_64-linux-gnu/ --config Release --build_wheel --skip_tests --skip_submodule_sync --cmake_extra_defines '"CMAKE_CUDA_ARCHITECTURES='${CMAKE_CUDA_ARCHITECTURES}'"'
|
||||
|
||||
|
|
@ -106,4 +106,4 @@ RUN /bin/sh build.sh ${PARSER_CONFIG} --parallel --build_shared_lib --cuda_home
|
|||
USER root
|
||||
|
||||
# Intall ORT wheel
|
||||
RUN pip install ${ONNXRUNTIME_LOCAL_CODE_DIR}/onnxruntime/build/Linux/Release/dist/*.whl
|
||||
RUN pip install ${ONNXRUNTIME_LOCAL_CODE_DIR}/onnxruntime/build/Linux/Release/dist/*.whl
|
||||
|
|
|
|||
|
|
@ -27,11 +27,11 @@ RUN apt-get install -y --no-install-recommends \
|
|||
ln -s /usr/bin/python3 python &&\
|
||||
ln -s /usr/bin/pip3 pip;
|
||||
|
||||
RUN pip install --upgrade pip
|
||||
RUN pip install --upgrade pip
|
||||
RUN pip install setuptools>=68.2.2 psutil
|
||||
|
||||
# Install TensorRT
|
||||
RUN version="10.2.0.19-1+cuda12.5" &&\
|
||||
RUN version="10.3.0.26-1+cuda12.5" &&\
|
||||
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/7fa2af80.pub &&\
|
||||
apt-get update &&\
|
||||
apt-get install -y \
|
||||
|
|
@ -98,7 +98,7 @@ RUN if [ -z "$ONNXRUNTIME_COMMIT_ID" ] ; then echo "Building branch ${ONNXRUNTIM
|
|||
git reset --hard ${ONNXRUNTIME_COMMIT_ID} && git submodule update --recursive ; fi
|
||||
|
||||
# Build ORT
|
||||
ENV CUDA_MODULE_LOADING "LAZY"
|
||||
ENV CUDA_MODULE_LOADING "LAZY"
|
||||
ARG PARSER_CONFIG=""
|
||||
RUN /bin/sh build.sh ${PARSER_CONFIG} --parallel --build_shared_lib --cuda_home /usr/local/cuda --cudnn_home /usr/lib/x86_64-linux-gnu/ --use_tensorrt --tensorrt_home /usr/lib/x86_64-linux-gnu/ --config Release --build_wheel --skip_tests --skip_submodule_sync --cmake_extra_defines '"CMAKE_CUDA_ARCHITECTURES='${CMAKE_CUDA_ARCHITECTURES}'"'
|
||||
|
||||
|
|
@ -106,4 +106,4 @@ RUN /bin/sh build.sh ${PARSER_CONFIG} --parallel --build_shared_lib --cuda_home
|
|||
USER root
|
||||
|
||||
# Intall ORT wheel
|
||||
RUN pip install ${ONNXRUNTIME_LOCAL_CODE_DIR}/onnxruntime/build/Linux/Release/dist/*.whl
|
||||
RUN pip install ${ONNXRUNTIME_LOCAL_CODE_DIR}/onnxruntime/build/Linux/Release/dist/*.whl
|
||||
|
|
|
|||
|
|
@ -5,7 +5,7 @@
|
|||
ARG BASEIMAGE=nvidia/cuda:11.8.0-cudnn8-devel-ubi8
|
||||
|
||||
FROM $BASEIMAGE
|
||||
ARG TRT_VERSION=10.2.0.19-1.cuda11.8
|
||||
ARG TRT_VERSION=10.3.0.26-1.cuda11.8
|
||||
|
||||
#Install TensorRT only if TRT_VERSION is not empty
|
||||
RUN if [ -n "${TRT_VERSION}" ]; then \
|
||||
|
|
@ -43,4 +43,3 @@ ARG BUILD_USER=onnxruntimedev
|
|||
RUN adduser --uid $BUILD_UID $BUILD_USER
|
||||
WORKDIR /home/$BUILD_USER
|
||||
USER $BUILD_USER
|
||||
|
||||
|
|
|
|||
|
|
@ -6,10 +6,10 @@ if exist PATH=%AGENT_TEMPDIRECTORY%\v12.2\ (
|
|||
) else (
|
||||
set PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\bin;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\extras\CUPTI\lib64;%PATH%
|
||||
)
|
||||
set PATH=%AGENT_TEMPDIRECTORY%\TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5\lib;%PATH%
|
||||
set PATH=%AGENT_TEMPDIRECTORY%\TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5\lib;%PATH%
|
||||
|
||||
@REM The default version is still cuda v12.2, because set cuda v11.8 after it
|
||||
set PATH=%PATH%;%AGENT_TEMPDIRECTORY%\TensorRT-10.2.0.19.Windows10.x86_64.cuda-11.8\lib
|
||||
set PATH=%PATH%;%AGENT_TEMPDIRECTORY%\TensorRT-10.3.0.26.Windows10.x86_64.cuda-11.8\lib
|
||||
if exist PATH=%AGENT_TEMPDIRECTORY%\v11.8\ (
|
||||
set PATH=%PATH%;%AGENT_TEMPDIRECTORY%\v11.8\bin;%AGENT_TEMPDIRECTORY%\v11.8\extras\CUPTI\lib64
|
||||
) else (
|
||||
|
|
|
|||
|
|
@ -6,6 +6,6 @@ if exist PATH=%AGENT_TEMPDIRECTORY%\v12.2\ (
|
|||
) else (
|
||||
set PATH=%PATH%;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\bin;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.2\extras\CUPTI\lib64
|
||||
)
|
||||
set PATH=%AGENT_TEMPDIRECTORY%\TensorRT-10.2.0.19.Windows10.x86_64.cuda-12.5\lib;%PATH%
|
||||
set PATH=%AGENT_TEMPDIRECTORY%\TensorRT-10.3.0.26.Windows10.x86_64.cuda-12.5\lib;%PATH%
|
||||
set GRADLE_OPTS=-Dorg.gradle.daemon=false
|
||||
set CUDA_MODULE_LOADING=LAZY
|
||||
set CUDA_MODULE_LOADING=LAZY
|
||||
|
|
|
|||
Loading…
Reference in a new issue