Skip to content

DeepRec-Compile-And-Install.md

Latest commit

 

History

History
152 lines (108 loc) · 5.87 KB

DeepRec-Compile-And-Install.md

File metadata and controls

152 lines (108 loc) · 5.87 KB

DeepRec源代码编译&安装

开发环境准备

CPU Base Docker Image

GCC Version Python Version IMAGE
7.5.0 3.6.9 alideeprec/deeprec-base:deeprec-base-cpu-py36-ubuntu18.04
11.2.0 3.8.6 alideeprec/deeprec-base:deeprec-base-cpu-py38-ubuntu22.04

GPU Base Docker Image

GCC Version Python Version CUDA VERSION IMAGE
7.5.0 3.6.9 CUDA 11.0.3 alideeprec/deeprec-base:deeprec-base-gpu-py36-cu110-ubuntu18.04
7.5.0 3.6.9 CUDA 11.2.2 alideeprec/deeprec-base:deeprec-base-gpu-py36-cu112-ubuntu18.04
7.5.0 3.6.9 CUDA 11.4.2 alideeprec/deeprec-base:deeprec-base-gpu-py36-cu114-ubuntu18.04
7.5.0 3.6.9 CUDA 11.6.1 alideeprec/deeprec-base:deeprec-base-gpu-py36-cu116-ubuntu18.04
7.5.0 3.6.9 CUDA 11.7.1 alideeprec/deeprec-base:deeprec-base-gpu-py36-cu117-ubuntu18.04
11.2.0 3.8.6 CUDA 11.7.1 alideeprec/deeprec-base:deeprec-base-gpu-py38-cu117-ubuntu22.04

CPU Dev Docker

GCC Version Python Version IMAGE
7.5.0 3.6.9 alideeprec/deeprec-dev:deeprec-cpu-py36-ubuntu18.04
11.2.0 3.8.6 alideeprec/deeprec-dev:deeprec-cpu-py38-ubuntu22.04

GPU Dev Docker

GCC Version Python Version CUDA VERSION IMAGE
7.5.0 3.6.9 CUDA 11.6.1 alideeprec/deeprec-dev:deeprec-gpu-py36-cu116-ubuntu18.04
11.2.0 3.8.6 CUDA 11.7.1 alideeprec/deeprec-dev:deeprec-gpu-py38-cu117-ubuntu22.04

CPU Dev Docker (with bazel cache)

GCC Version Python Version IMAGE
7.5.0 3.6.9 alideeprec/deeprec-cicd:deeprec-dev-cpu-cibuild-py36-ubuntu18.04
11.2.0 3.8.6 alideeprec/deeprec-cicd:deeprec-dev-cpu-cibuild-py38-ubuntu22.04

GPU Dev Docker (with bazel cache)

GCC Version Python Version CUDA VERSION IMAGE
7.5.0 3.6.9 CUDA 11.6.1 alideeprec/deeprec-cicd:deeprec-dev-gpu-cibuild-py36-cu116-ubuntu18.04
11.2.0 3.8.6 CUDA 11.7.1 alideeprec/deeprec-cicd:deeprec-dev-gpu-cibuild-py38-cu117-ubuntu22.04

代码编译

GPU Environment 为了更好的发挥GPU性能，根据编译/运行的GPU卡，配置不同的TF_CUDA_COMPUTE_CAPABILITIES

GPU architecture TF_CUDA_COMPUTE_CAPABILITIES
Pascal (P100) 6.0+6.1
Volta (V100) 7.0
Turing (T4) 7.5
Ampere (A10, A100) 8.0+8.6

如果希望编译出支持不同GPU卡上执行的版本，可以配置多个值，比如DeepRec中默认配置为"6.0,6.1,7.0,7.5,8.0"

比如配置环境变量TF_CUDA_COMPUTE_CAPABILITIES方法：

export TF_CUDA_COMPUTE_CAPABILITIES="7.5,8.0"
./configure

GPU/CPU版本编译

bazel build -c opt --config=opt //tensorflow/tools/pip_package:build_pip_package

GPU/CPU版本编译+ABI=0

bazel build --cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0" --host_cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0" -c opt --config=opt //tensorflow/tools/pip_package:build_pip_package

编译开启OneDNN + Eigen Threadpool工作线程池版本（CPU）

bazel build  -c opt --config=opt  --config=mkl_threadpool --define build_with_mkl_dnn_v1_only=true //tensorflow/tools/pip_package:build_pip_package

编译开启OneDNN + Eigen Threadpool工作线程池版本+ABI=0的版本 （CPU）

bazel build --cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0" --host_cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0" -c opt --config=opt --config=mkl_threadpool --define build_with_mkl_dnn_v1_only=true //tensorflow/tools/pip_package:build_pip_package

生成Whl包

./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

安装Whl包

pip3 install /tmp/tensorflow_pkg/tensorflow-1.15.5+${version}-cp36-cp36m-linux_x86_64.whl

最新Release镜像

CPU镜像

alideeprec/deeprec-release:deeprec2208-cpu-py36-ubuntu18.04

GPU CUDA11.6镜像

alideeprec/deeprec-release:deeprec2208-gpu-py36-cu116-ubuntu18.04

DeepRec Processor编译打包

配置.bazelrc（注意如果编译DeepRec请重新进行configure配置）

./configure serving

增加MKL相关配置

./configure serving --mkl
./configure serving --mkl_open_source_v1_only
./configure serving --mkl_threadpool
./configure serving --mkl --cuda ...

更多细节请查看: serving/configure.py

编译processor库，会生成libserving_processor.so，用户可以加载该so，并且调用示例中的serving API进行predict

bazel build //serving/processor/serving:libserving_processor.so

单元测试

bazel test -- //serving/processor/... -//serving/processor/framework:lookup_manual_test

E2E测试细节请查看: serving/processor/tests/end2end/README