mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-07-13 18:08:13 +00:00
4dfb89b3ad
Implemented "lock-free" spinlock to save CPU usage on context switching. The change has been tested on queene service of Ads team, the lock-free version of ort (40 threads) saves CPU usage on gen8 (128 logical processors on 8 numa nodes) windows by nearly half, from 65% to 35%. For 32 cores, the curve is flat: Anubis, 32 vCPU, windows, hugging face models, 95 percentile E2E latency in ms: model | mutex(ms) | mutex-free --- | --- | --- alvert_base_v2 | 34.21 | 34.09 bert_large_uncased | 116.27| 117.84 bart_base | 72.06 | 71.99 distilgpt2 | 25.43 | 25.02 vit_base_patch16_224 | 37.33 | 37.76 Anubis, 32 vCPU win, Linux, 1st party models, 95 percentile E2E latency in ms: model | mutex(ms) | mutex-free --- | --- | --- deepthink_v2 | 24.35 | 22.95 bing_feeds | 36.96 | 36.48 deep_writes | 14.46 | 14.32 keypoints | 9.34 | 7.69 model11 | 1.71 | 1.66 model12 | 1.82 | 1.44 model2 | 4.21 | 3.95 model6 | 1.08 | 1.05 agiencoder | 0.99 | 0.93 geminet_transformer | 5.32 | 5.24 --------- Co-authored-by: Randy Shuai <rashuai@microsoft.com>
214 lines
No EOL
6.9 KiB
YAML
214 lines
No EOL
6.9 KiB
YAML
parameters:
|
|
- name: RunOnnxRuntimeTests
|
|
displayName: Run Tests?
|
|
type: boolean
|
|
default: true
|
|
|
|
stages:
|
|
- stage: x64_debug
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'Debug'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --build_java --build_nodejs --build_wheel --disable_memleak_checker
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_debug
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: false
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022'
|
|
|
|
- stage: x64_release
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
# Compare to our Nuget packaging pipeline, this job has "--build_wheel" but doesn't have "--enable_lto --disable_rtti --use_telemetry --enable_wcos"
|
|
# Python bindings use typeid so I can't disable RTTI here. If it causes a problem, we will need to split this job to two jobs.
|
|
additionalBuildFlags: --build_wheel --build_java --build_nodejs
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_release
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: false
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022'
|
|
|
|
- stage: x64_release_dnnl
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --build_wheel --use_dnnl --build_java
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_release
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: false
|
|
ORT_EP_NAME: DNNL
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
# Intel EPs require Intel CPUs
|
|
MachinePool: 'onnxruntime-Win2022-Intel-CPU'
|
|
|
|
# Tests doesn't work on AMD CPUs
|
|
- stage: x64_release_xnnpack
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --build_wheel --use_xnnpack
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_release
|
|
RunOnnxRuntimeTests: false
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: false
|
|
ORT_EP_NAME: XNNPACK
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022'
|
|
|
|
- stage: x64_release_winml
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --use_winml --enable_wcos --disable_rtti --cmake_extra_defines CMAKE_SYSTEM_VERSION=10.0.22000.0
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_release_winml
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
# WinML has many warnings
|
|
RunStaticCodeAnalysis: false
|
|
EnablePython: false
|
|
isTraining: false
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022'
|
|
|
|
- stage: x86_release
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env_x86.bat
|
|
buildArch: x86
|
|
additionalBuildFlags: --build_wheel
|
|
msbuildPlatform: Win32
|
|
isX86: true
|
|
job_name_suffix: x86_release
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: false
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022'
|
|
|
|
- stage: training_x64_debug
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'Debug'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --enable_training --build_wheel --disable_memleak_checker
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: training_x64_debug
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
isTraining: true
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win2022-CPU-training-AMD'
|
|
|
|
- stage: training_x64_release
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --enable_training --build_wheel
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: training_x64_release
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: true
|
|
isTraining: true
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win2022-CPU-training-AMD'
|
|
|
|
- stage: ort_training_apis_x64_release
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --enable_training_apis
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: ort_training_apis_x64_release
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
EnablePython: false
|
|
isTraining: true
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win2022-CPU-training-AMD'
|
|
|
|
- stage: x64_release_azure
|
|
dependsOn: []
|
|
jobs:
|
|
- template: templates/jobs/win-ci-vs-2022-job.yml
|
|
parameters:
|
|
BuildConfig: 'RelWithDebInfo'
|
|
EnvSetupScript: setup_env_azure.bat
|
|
buildArch: x64
|
|
additionalBuildFlags: --use_azure --use_lock_free_queue
|
|
msbuildPlatform: x64
|
|
isX86: false
|
|
job_name_suffix: x64_release_azure
|
|
RunOnnxRuntimeTests: ${{ parameters.RunOnnxRuntimeTests }}
|
|
RunStaticCodeAnalysis: false
|
|
EnablePython: false
|
|
isTraining: false
|
|
ORT_EP_NAME: CPU
|
|
GenerateDocumentation: false
|
|
WITH_CACHE: true
|
|
MachinePool: 'onnxruntime-Win-CPU-2022' |