Compare commits

..

84 Commits

Author SHA1 Message Date
Leo Chen 463617d757 Revert "[NPU] add npu kernel for mean Op (#31562)"
5 years ago
OleNet 468ac6993b
[NPU] add npu kernel for mean Op (#31562)
5 years ago
Leo Chen 5118968d80
[NPU] add npu kernel for softmax_with_cross_entropy (#31656)
5 years ago
zhang wenhui 925432d85e
【NPU】Support npu kernel for mul op (#31584)
5 years ago
Leo Chen 1e956001ec
[NPU] add npu kernel for adam (#31644)
5 years ago
pangyoki 795b0f92d3
【NPU】Support NPU kernel for reduce_sum op v2 (#31620)
5 years ago
Leo Chen b541ca8795
[NPU] add npu kernel for sgd (#31639)
5 years ago
oyxuan-11 57220f594d
[NPU] Support NPU kernel cast op (#31635)
5 years ago
Leo Chen 3ca4bc1004
[NPU] fix allocator min chunk size (#31632)
5 years ago
oyxuan-11 11f788771c
[NPU] Support npu kernel scatter op (#31624)
5 years ago
Meiyim e3e15792a4
[NPU] support npu kernel for `less_than` (#31327)
5 years ago
Meiyim a3cc4a4a69
[NPU] Support npu op table_lookup_v2 and table_lookup_v2_grad (#31399)
5 years ago
oyjxer f250416029
[NPU] Support npu op elementwise_pow (#31576)
5 years ago
oyjxer 7241bc2210
[NPU] Support npu op elementwise_min (#31575)
5 years ago
oyjxer 9606a86b18
[NPU] Support npu op logicalnot_op (#31534)
5 years ago
oyjxer 47860ce20d
[NPU] Support npu op log, log_grad, sqrt, sqrt_grad, square, tanh and tanh_grad (#31600)
5 years ago
oyjxer de65486c19
【NPU】Support npu op elementwise_div and elementwise_div_grad (#31573)
5 years ago
OleNet ec2160a622
[NPU] add range op (#31560)
5 years ago
Leo Chen 0234693040
fix gather_grad bug (#31607)
5 years ago
Leo Chen 5e851bff42
[NPU] fix assgin cmake (#31595)
5 years ago
oyjxer 382fc31f89
【NPU】Support npu op gelu and gelu_grad (#31530)
5 years ago
oyjxer 5d29a27c2e
[NPU] fix npu op elementwise_mul_grad (#31592)
5 years ago
OleNet 09bf2cfc0e
[NPU] add Assign OP (#31561)
5 years ago
xiayanming f1fdddfdc8
[NPU] Support npu kernel for c sync stream op (#31386)
5 years ago
yinhaofeng e1c33a6d69
[NPU] accuracy op (#31492)
5 years ago
xiayanming 3bf8a34c69
[NPU] Support npu kernel for amp_check_finite_and_unscale_npu op (#31457)
5 years ago
xiayanming d746197398
[NPU] Support npu kernel for gather op fix bug (#31541)
5 years ago
zhang wenhui 5d22e15b6e
【NPU】Suppert npu kernel for reshape2 op (#31524)
5 years ago
zhang wenhui 581e5460a0
【NPU】add relu op for npu (#31515)
5 years ago
oyjxer cfeeb4bc95
[NPU] Support npu op elementwise_max (#31574)
5 years ago
oyjxer e15ccafb84
[NPU] Support npu op elementwise_mul and elementwise_mul_grad (#31571)
5 years ago
zhang wenhui 29d50d2049
【NPU】Support npu kernel for matmul op (#31544)
5 years ago
xiayanming f400ce9f51
[NPU] Support npu kernel for reduceany op (#31422)
5 years ago
zhang wenhui 7524ac9345
【NPU】support npu kernel for fill_constant op (#31521)
5 years ago
Leo Chen 3f206e97c4
Support TensorFormVector, TensorToVector of bool type (#31518)
5 years ago
zhang wenhui 9df84bd693
【NPU】add scale op for npu (#31499)
5 years ago
xiayanming e19195f795
Support npu kernel for gather op (#31458)
5 years ago
lw921014 15823bb0df
[NPU] add npu kernel for communication op (#31437)
5 years ago
Reventon_L 388c69f27d
[NPU] squeeze and unsqueeze op for ascend (#31452)
5 years ago
Leo Chen c956c035dc
fix cmake of cryptopp to avoid downloading every time (#31451)
5 years ago
Leo Chen 83f81eb573
Fix pow, refine code (#31440)
5 years ago
Leo Chen 5fe3d596e4
Fix pow, use fillD instead of broadcast (#31433)
5 years ago
zhang wenhui ecc6e213d7
fix endif (#31431)
5 years ago
zhang wenhui b3c88e961c
[NPU] Support npu kernel for shape op (#31427)
5 years ago
Leo Chen ac3d821bc0
[NPU] add npu kernel for equal op (#31393)
5 years ago
Leo Chen 0310945f5c
[NPU] Support npu op layer_norm and layer_norm_grad (#31310)
5 years ago
Void Main 45765d6eb6
Refactor HCCLCommContext to be compatible with Paddle (#31359)
5 years ago
Leo Chen 8497e2aad3
[NPU] add npu kernel for elementwise_add_grad (#31347)
5 years ago
lw921014 9fcdaeba5e
add allreduce and broadcast without test (#31024)
5 years ago
Leo Chen 5618f14047
fix reading flags from env (#31329)
5 years ago
liym27 a1ddff81e3
[NPU] Support npu op: (1) slice (2) slice_grad (#31275)
5 years ago
Leo Chen d23bf89cf6
support list of list attribute for NPU (#31299)
5 years ago
liym27 77a0c41cb2
Fix pow npu fp16 test (#31256)
5 years ago
liym27 187248f568
[NPU] Support npu op pow and pow grad (#31247)
5 years ago
xiayanming 821c2f4ef8
add ascend unittest (#31249)
5 years ago
Leo Chen d45f5d787e
Fix typo of selected_npus (#31230)
5 years ago
xiayanming 387c1db4f1
Ascendrc (#31065)
5 years ago
Leo Chen ff4654e216
refactor npu device manager (#31154)
5 years ago
liym27 1435b4c096
[NPU] Support executor with NPU (#31057)
5 years ago
Leo Chen 678a3e8fed
support adding correct npu op in pybind.h (#31143)
5 years ago
Leo Chen 85cbd55648
Fix compilation problem (#31100)
5 years ago
Leo Chen 5cb20f30fc
add npu kernel for elementwise_sub and elementwise_sub_grad (#30973)
5 years ago
gongweibao c687edecd8
Fix reshape on GE graph. (#31084)
5 years ago
xiayanming a6edbc478b
support parsing ascend rank table file (#31000)
5 years ago
Leo Chen 1201cd2ef2
[feature] support npu allocator, part 2 (#30972)
5 years ago
Leo Chen 7e049108c5
[feature] support npu operator (#30951)
5 years ago
Leo Chen 81138239db
[feature] support npu allocator (#30840)
5 years ago
gongweibao ebef6601d5
Destroy session first. (#30954)
5 years ago
Leo Chen 500f28ec37
pass cxx_flags to gloo cmake (#30857)
5 years ago
gongweibao de42d19336
Add paddle ascend distribution training supported (#30796)
5 years ago
OleNet ebb5d181e8
Ascendrc add converted op : [range/equal/range/uniform_random/expand/squeeze], fix cast op bug (#30797)
5 years ago
dingsiyu 4a26729540
Merge ascend_optimizer and ascend_parser. (#30776)
5 years ago
gongweibao 636fefd9f8
code style (#30781)
5 years ago
Leo Chen 88dfd067bf
Dev/fix ascend string (#30749)
5 years ago
Leo Chen 6eabbc8076
fix compilation on ascend-20.1 (#30722)
5 years ago
Void Main 904cc44349
[Feature] Build parser to support distributed training (#30658)
5 years ago
gongweibao 5b77b259d8
cleanup (#30646)
5 years ago
gongweibao 7158061a29
Add startup bash files of test_ascend_group. (#30645)
5 years ago
gongweibao e4287ca60b
Add Hccl program group (#30642)
5 years ago
gongweibao f5aca8fbb4
Pass device_ids info from launch to trainer. (#30632)
5 years ago
Void Main d2404da768
Build praser for Hcom* operators (#30627)
5 years ago
gongweibao f9c97dd728
Add distribution supported (#30578)
5 years ago
gongweibao 1882f2ce2d
Fix compilcation on CANN20.1 and older (#30494)
5 years ago
hutuxian 6dd52c5b25
Ascend rc (#30483)
5 years ago

@ -12,8 +12,7 @@
# See the License for the specific language governing permissions and
# limitations under the License
cmake_minimum_required(VERSION 3.15)
cmake_policy(VERSION 3.10)
cmake_minimum_required(VERSION 3.10)
set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
set(PADDLE_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR})
set(PADDLE_BINARY_DIR ${CMAKE_CURRENT_BINARY_DIR})
@ -33,12 +32,21 @@ option(WITH_TENSORRT "Compile PaddlePaddle with NVIDIA TensorRT" OFF)
option(WITH_XPU "Compile PaddlePaddle with BAIDU KUNLUN XPU" OFF)
option(WITH_WIN_DUMP_DBG "Compile with windows core dump debug mode" OFF)
option(WITH_ASCEND "Compile PaddlePaddle with ASCEND" OFF)
# NOTE(zhiqiu): WITH_ASCEND_CL can be compile on x86_64, so we can set WITH_ASCEND=OFF and WITH_ASCEND_CL=ON
# to develop some acl related functionality on x86
option(WITH_ASCEND_CL "Compile PaddlePaddle with ASCEND CL" ${WITH_ASCEND})
option(WITH_ASCEND_CXX11 "Compile PaddlePaddle with ASCEND and CXX11 ABI" OFF)
if (WITH_GPU AND WITH_XPU)
message(FATAL_ERROR "Error when compile GPU and XPU at the same time")
endif()
if (WITH_GPU AND WITH_ASCEND)
if (WITH_GPU AND WITH_ASCEND)
message(FATAL_ERROR "Error when compile GPU and ASCEND at the same time")
endif()
# cmake 3.12, 3.13, 3.14 will append gcc link options to nvcc, and nvcc doesn't recognize them.
if(WITH_GPU AND (${CMAKE_VERSION} VERSION_GREATER_EQUAL 3.12) AND (${CMAKE_VERSION} VERSION_LESS 3.15))
message(FATAL_ERROR "cmake ${CMAKE_VERSION} is not supported when WITH_GPU=ON because of bug https://cmake.org/pipermail/cmake/2018-September/068195.html. "
"You can use cmake 3.16 (recommended), 3.10, 3.11, 3.15 or 3.17. Please refer to the install document: https://cmake.org/install/")
endif()
if(WITH_GPU AND NOT APPLE)
enable_language(CUDA)
@ -57,6 +65,9 @@ if(WITH_MUSL)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-error=deprecated-declarations -Wno-deprecated-declarations -Wno-error=pessimizing-move -Wno-error=deprecated-copy")
endif()
if(WITH_ASCEND AND NOT WITH_ASCEND_CXX11)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_GLIBCXX_USE_CXX11_ABI=0")
endif()
if(WIN32)
option(MSVC_STATIC_CRT "use static C Runtime library by default" ON)
@ -64,24 +75,12 @@ if(WIN32)
set(CMAKE_SUPPRESS_REGENERATION ON)
set(CMAKE_STATIC_LIBRARY_PREFIX lib)
set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} /bigobj")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /bigobj")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /bigobj")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /bigobj")
if("${CMAKE_GENERATOR}" STREQUAL "Ninja")
set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} /Zc:inline")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /Zc:inline")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Zc:inline")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Zc:inline")
endif()
if (MSVC_STATIC_CRT)
message(STATUS "Use static C runtime time, refer to https://docs.microsoft.com/en-us/cpp/c-runtime-library/crt-library-features?view=vs-2019")
set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} /MTd")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /MT")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /MTd")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /MT")
set(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} /bigobj /MTd")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} /bigobj /MT")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /bigobj /MTd")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /bigobj /MT")
foreach(flag_var
CMAKE_CXX_FLAGS CMAKE_CXX_FLAGS_DEBUG CMAKE_CXX_FLAGS_RELEASE
CMAKE_CXX_FLAGS_MINSIZEREL CMAKE_CXX_FLAGS_RELWITHDEBINFO
@ -92,9 +91,7 @@ if(WIN32)
endif()
endforeach(flag_var)
endif()
# NOTE(Avin0323): Less parallel count result in faster compilation.
math(EXPR PROCESS_MAX "${CPU_CORES} * 2 / 3")
# windows build turn off warnings, use parallel compiling.
foreach(flag_var
CMAKE_CXX_FLAGS CMAKE_CXX_FLAGS_DEBUG CMAKE_CXX_FLAGS_RELEASE
@ -102,7 +99,13 @@ if(WIN32)
CMAKE_C_FLAGS CMAKE_C_FLAGS_DEBUG CMAKE_C_FLAGS_RELEASE
CMAKE_C_FLAGS_MINSIZEREL CMAKE_C_FLAGS_RELWITHDEBINFO)
string(REGEX REPLACE "/W[1-4]" " /W0 " ${flag_var} "${${flag_var}}")
set(${flag_var} "${${flag_var}} /MP${PROCESS_MAX}")
# NOTE(Avin0323): Less parallel count result in faster compilation with
# Unity Build on GPU.
if(WITH_UNITY_BUILD AND WITH_GPU)
set(${flag_var} "${${flag_var}} /MP8")
else()
set(${flag_var} "${${flag_var}} /MP")
endif()
endforeach(flag_var)
foreach(flag_var CMAKE_CXX_FLAGS CMAKE_C_FLAGS)
set(${flag_var} "${${flag_var}} /w")
@ -120,10 +123,6 @@ if(WIN32)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} /wd4068 /wd4129 /wd4244 /wd4267 /wd4297 /wd4530 /wd4577 /wd4819 /wd4838")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /wd4068 /wd4129 /wd4244 /wd4267 /wd4297 /wd4530 /wd4577 /wd4819 /wd4838")
foreach(flag_var CMAKE_SHARED_LINKER_FLAGS CMAKE_STATIC_LINKER_FLAGS CMAKE_EXE_LINKER_FLAGS CMAKE_LINKER_FLAGS)
set(${flag_var} "${${flag_var}} /ignore:4049 /ignore:4217 /ignore:4006 /ignore:4221")
endforeach(flag_var)
if (WITH_WIN_DUMP_DBG)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} /Zi")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi")
@ -161,9 +160,8 @@ option(WITH_DISTRIBUTE "Compile with distributed support" OFF)
option(WITH_BRPC_RDMA "Use brpc rdma as the rpc protocal" OFF)
option(ON_INFER "Turn on inference optimization and inference-lib generation" OFF)
################################ Internal Configurations #######################################
option(WITH_ROCM "Compile PaddlePaddle with ROCM platform" OFF)
option(WITH_RCCL "Compile PaddlePaddle with RCCL support" OFF)
option(WITH_NV_JETSON "Compile PaddlePaddle with NV JETSON" OFF)
option(WITH_ROCM_PLATFORM "Compile PaddlePaddle with ROCM platform" OFF)
option(WITH_NV_JETSON "Compile PaddlePaddle with NV JETSON" OFF)
option(WITH_PROFILER "Compile PaddlePaddle with GPU profiler and gperftools" OFF)
option(WITH_COVERAGE "Compile PaddlePaddle with code coverage" OFF)
option(WITH_INCREMENTAL_COVERAGE "Generate coverage reports only for incremental code" OFF)
@ -292,25 +290,19 @@ include(configure) # add paddle env configuration
include_directories("${PADDLE_SOURCE_DIR}")
if(WITH_ROCM)
include(hip)
endif(WITH_ROCM)
if (NOT WITH_ROCM AND WITH_RCCL)
MESSAGE(WARNING
"Disable RCCL when compiling without GPU. Force WITH_RCCL=OFF.")
set(WITH_NCCL OFF CACHE STRING
"Disable RCCL when compiling without GPU" FORCE)
endif()
if(WITH_RCCL)
add_definitions("-DPADDLE_WITH_RCCL")
include(rccl)
if(NOT DEFINED ENV{ROCM_PATH})
set(ROCM_PATH "/opt/rocm" CACHE PATH "Path to which ROCm has been installed")
set(HIP_PATH ${ROCM_PATH}/hip CACHE PATH "Path to which HIP has been installed")
else()
if(WITH_ROCM)
MESSAGE(WARNING "If the environment is multi-card, the WITH_RCCL option needs to be turned on, otherwise only a single card can be used.")
endif()
set(ROCM_PATH $ENV{ROCM_PATH} CACHE PATH "Path to which ROCm has been installed")
set(HIP_PATH ${ROCM_PATH}/hip CACHE PATH "Path to which HIP has been installed")
endif()
set(CMAKE_MODULE_PATH "${HIP_PATH}/cmake" ${CMAKE_MODULE_PATH})
if(WITH_ROCM_PLATFORM)
find_package(HIP)
include(hip)
endif(WITH_ROCM_PLATFORM)
if(WITH_NV_JETSON)
set(WITH_ARM ON CACHE STRING "Set WITH_ARM=ON when compiling WITH_NV_JETSON=ON." FORCE)
@ -343,8 +335,6 @@ set(PADDLE_PYTHON_BUILD_DIR "${CMAKE_CURRENT_BINARY_DIR}/python/build")
set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
set(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
add_definitions(-DPADDLE_DLL_EXPORT)
if(ON_INFER)
# you can trun off the paddle fluid and inference lib by set ON_INFER=OFF
message(STATUS "On inference mode, will take place some specific optimization.")

@ -8,8 +8,8 @@
English | [简体中文](./README_cn.md)
[![Build Status](https://travis-ci.org/PaddlePaddle/Paddle.svg?branch=develop)](https://travis-ci.org/PaddlePaddle/Paddle)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](https://paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](https://paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://www.paddlepaddle.org.cn/documentation/docs/en/1.8/beginners_guide/index_en.html)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://www.paddlepaddle.org.cn/documentation/docs/zh/1.8/beginners_guide/index_cn.html)
[![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle.svg)](https://github.com/PaddlePaddle/Paddle/releases)
[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
@ -22,21 +22,23 @@ PaddlePaddle is originated from industrial practices with dedication and commitm
## Installation
### Latest PaddlePaddle Release: [v2.0](https://github.com/PaddlePaddle/Paddle/tree/release/2.0)
### Latest PaddlePaddle Release: [v1.8](https://github.com/PaddlePaddle/Paddle/tree/release/1.8)
Our vision is to enable deep learning for everyone via PaddlePaddle.
Please refer to our [release announcement](https://github.com/PaddlePaddle/Paddle/releases) to track the latest features of PaddlePaddle.
### Install Latest Stable Release:
```
# CPU
# Linux CPU
pip install paddlepaddle
# GPU
# Linux GPU cuda10cudnn7
pip install paddlepaddle-gpu
# Linux GPU cuda9cudnn7
pip install paddlepaddle-gpu==1.8.5.post97
```
More infomation about installation, please view [Quick Install](https://www.paddlepaddle.org.cn/install/quick)
It is recommended to read [this doc](https://www.paddlepaddle.org.cn/documentation/docs/en/beginners_guide/install/index_en.html) on our website.
Now our developers can acquire Tesla V100 online computing resources for free. If you create a program by AI Studio, you will obtain 10 hours to train models online per day. [Click here to start](https://aistudio.baidu.com/aistudio/index).
Now our developers can acquire Tesla V100 online computing resources for free. If you create a program by AI Studio, you will obtain 12 hours to train models online per day. If you can insist on that for five consecutive days, then you will receive an extra 48 hours. [Click here to start](http://ai.baidu.com/support/news?action=detail&id=981).
## FOUR LEADING TECHNOLOGIES
@ -65,30 +67,38 @@ Now our developers can acquire Tesla V100 online computing resources for free. I
## Documentation
We provide [English](https://www.paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html) and
[Chinese](https://www.paddlepaddle.org.cn/documentation/docs/zh/guide/index_cn.html) documentation.
We provide [English](http://www.paddlepaddle.org.cn/documentation/docs/en/1.8/beginners_guide/index_en.html) and
[Chinese](http://www.paddlepaddle.org.cn/documentation/docs/zh/1.8/beginners_guide/index_cn.html) documentation.
- [Guides](https://www.paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html)
- [Basic Deep Learning Models](https://www.paddlepaddle.org.cn/documentation/docs/en/beginners_guide/index_en.html)
You might want to start from how to implement deep learning basics with PaddlePaddle.
- [Practice](https://www.paddlepaddle.org.cn/documentation/docs/zh/tutorial/index_cn.html)
- [User Guides](https://www.paddlepaddle.org.cn/documentation/docs/en/user_guides/index_en.html)
You might have got the hang of Beginners Guide, and wish to model practical problems and build your original networks.
- [Advanced User Guides](https://www.paddlepaddle.org.cn/documentation/docs/en/advanced_guide/index_en.html)
So far you have already been familiar with Fluid. And the next step should be building a more efficient model or inventing your original Operator.
- [API Reference](https://www.paddlepaddle.org.cn/documentation/docs/en/api/index_en.html)
Our new API enables much shorter programs.
- [How to Contribute](https://www.paddlepaddle.org.cn/documentation/docs/en/guides/08_contribution/index_en.html)
- [How to Contribute](https://www.paddlepaddle.org.cn/documentation/docs/en/advanced_guide/addon_development/contribute_code/index_en.html)
We appreciate your contributions!
## Communication
- [Github Issues](https://github.com/PaddlePaddle/Paddle/issues): bug reports, feature requests, install issues, usage issues, etc.
- QQ discussion group: 778260830 (PaddlePaddle).
- [Forums](https://ai.baidu.com/forum/topic/list/168?pageNo=1): discuss implementations, research, etc.
- QQ discussion group: 796771754 (PaddlePaddle).
- [Forums](http://ai.baidu.com/forum/topic/list/168?pageNo=1): discuss implementations, research, etc.
## Copyright and License
PaddlePaddle is provided under the [Apache-2.0 license](LICENSE).

@ -8,8 +8,8 @@
[English](./README.md) | 简体中文
[![Build Status](https://travis-ci.org/PaddlePaddle/Paddle.svg?branch=develop)](https://travis-ci.org/PaddlePaddle/Paddle)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](https://paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](https://paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://www.paddlepaddle.org.cn/documentation/docs/en/1.8/beginners_guide/index_en.html)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://www.paddlepaddle.org.cn/documentation/docs/zh/1.8/beginners_guide/index_cn.html)
[![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle.svg)](https://github.com/PaddlePaddle/Paddle/releases)
[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
@ -19,20 +19,23 @@
## 安装
### PaddlePaddle最新版本: [v2.0](https://github.com/PaddlePaddle/Paddle/tree/release/2.0)
### PaddlePaddle最新版本: [v1.8](https://github.com/PaddlePaddle/Paddle/tree/release/1.8)
跟进PaddlePaddle最新特性请参考我们的[版本说明](https://github.com/PaddlePaddle/Paddle/releases)
### 安装最新稳定版本:
```
# CPU
# Linux CPU
pip install paddlepaddle
# GPU
# Linux GPU cuda10cudnn7
pip install paddlepaddle-gpu
# Linux GPU cuda9cudnn7
pip install paddlepaddle-gpu==1.8.5.post97
```
更多安装信息详见官网 [安装说明](https://www.paddlepaddle.org.cn/install/quick)
PaddlePaddle用户可领取**免费Tesla V100在线算力资源**,训练模型更高效。**每日登陆即送10小时**[前往使用免费算力](https://aistudio.baidu.com/aistudio/index)。
PaddlePaddle用户可领取**免费Tesla V100在线算力资源**,训练模型更高效。**每日登陆即送12小时****连续五天运行再加送48小时**[前往使用免费算力](https://ai.baidu.com/support/news?action=detail&id=981)。
## 四大领先技术
@ -61,30 +64,38 @@ PaddlePaddle用户可领取**免费Tesla V100在线算力资源**,训练模型
## 文档
我们提供 [英文](https://www.paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html) 和
[中文](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html) 文档
我们提供 [英文](http://www.paddlepaddle.org.cn/documentation/docs/en/1.8/beginners_guide/index_en.html) 和
[中文](http://www.paddlepaddle.org.cn/documentation/docs/zh/1.8/beginners_guide/index_cn.html) 文档
- [使用指南](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html)
- [深度学习基础教程](https://www.paddlepaddle.org.cn/documentation/docs/zh/beginners_guide/index_cn.html)
或许您想从深度学习基础开始学习飞桨
- [应用实践](https://www.paddlepaddle.org.cn/documentation/docs/zh/tutorial/index_cn.html)
- [典型案例](https://www.paddlepaddle.org.cn/documentation/docs/zh/user_guides/index_cn.html)
或许您已经掌握了新手入门阶段的内容,期望可以针对实际问题建模、搭建自己网络
- [进阶指南](https://www.paddlepaddle.org.cn/documentation/docs/zh/advanced_guide/index_cn.html)
或许您已比较熟练使用PaddlePaddle来完成常规任务期望获得更高效的模型或者定义自己的Operator
- [API Reference](https://www.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html)
- [API Reference](https://www.paddlepaddle.org.cn/documentation/docs/zh/api_cn/index_cn.html)
新的API支持代码更少更简洁的程序
- [贡献方式](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/08_contribution/index_cn.html)
- [贡献方式](https://www.paddlepaddle.org.cn/documentation/docs/zh/advanced_guide/addon_development/contribute_code/index_cn.html)
欢迎您的贡献!
## 交流与反馈
- 欢迎您通过[Github Issues](https://github.com/PaddlePaddle/Paddle/issues)来提交问题、报告与建议
- QQ群: 778260830 (PaddlePaddle)
- [论坛](https://ai.baidu.com/forum/topic/list/168): 欢迎大家在PaddlePaddle论坛分享在使用PaddlePaddle中遇到的问题和经验, 营造良好的论坛氛围
- QQ群: 796771754 (PaddlePaddle)
- [论坛](http://ai.baidu.com/forum/topic/list/168): 欢迎大家在PaddlePaddle论坛分享在使用PaddlePaddle中遇到的问题和经验, 营造良好的论坛氛围
## 版权和许可证
PaddlePaddle由[Apache-2.0 license](LICENSE)提供

@ -82,6 +82,10 @@ if(WITH_ASCEND)
add_definitions(-DPADDLE_WITH_ASCEND)
endif()
if(WITH_ASCEND_CL)
add_definitions(-DPADDLE_WITH_ASCEND_CL)
endif()
if(WITH_XPU)
message(STATUS "Compile with XPU!")
add_definitions(-DPADDLE_WITH_XPU)
@ -130,10 +134,14 @@ if(WITH_GPU)
endif()
include_directories(${TENSORRT_INCLUDE_DIR})
endif()
elseif(WITH_ROCM)
elseif(WITH_ROCM_PLATFORM)
add_definitions(-DPADDLE_WITH_HIP)
add_definitions(-DEIGEN_USE_GPU)
add_definitions(-DEIGEN_USE_HIP)
add_definitions(-D__HIP_PLATFORM_HCC__)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -D__HIP_PLATFORM_HCC__")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D__HIP_PLATFORM_HCC__")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_HIP")
set(THRUST_DEVICE_SYSTEM THRUST_DEVICE_SYSTEM_HIP)
else()
add_definitions(-DHPPL_STUB_FUNC)
list(APPEND CMAKE_CXX_SOURCE_FILE_EXTENSIONS cu)

@ -63,10 +63,7 @@ endfunction()
if(WITH_COVERAGE)
if (WITH_INCREMENTAL_COVERAGE)
# if *.h changed, generate coverage report totaly.
# if pybind.cc changed, generate coverage report totaly.
# Because if pybind.cc add '-g -O0 -fprofile-arcs -ftest-coverage' only, some testcase will fail.
if ( (NOT ("$ENV{PADDLE_GIT_DIFF_H_FILE}" STREQUAL "")) OR ("$ENV{PADDLE_GIT_DIFF_CC_FILE}" MATCHES "pybind.cc") )
if (NOT ("$ENV{PADDLE_GIT_DIFF_H_FILE}" STREQUAL ""))
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -g -O0 -fprofile-arcs -ftest-coverage")
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -g -O0 -fprofile-arcs -ftest-coverage")
endif()

@ -74,7 +74,7 @@ endfunction()
# select_nvcc_arch_flags(out_variable)
function(select_nvcc_arch_flags out_variable)
# List of arch names
set(archs_names "Kepler" "Maxwell" "Pascal" "Volta" "Turing" "Ampere" "All" "Manual")
set(archs_names "Kepler" "Maxwell" "Pascal" "Volta" "Turing" "All" "Manual")
set(archs_name_default "Auto")
list(APPEND archs_names "Auto")
@ -91,7 +91,7 @@ function(select_nvcc_arch_flags out_variable)
if(${CUDA_ARCH_NAME} STREQUAL "Manual")
set(CUDA_ARCH_BIN ${paddle_known_gpu_archs} CACHE STRING "Specify 'real' GPU architectures to build binaries for, BIN(PTX) format is supported")
set(CUDA_ARCH_PTX "" CACHE STRING "Specify 'virtual' PTX architectures to build PTX intermediate code for")
set(CUDA_ARCH_PTX "50" CACHE STRING "Specify 'virtual' PTX architectures to build PTX intermediate code for")
mark_as_advanced(CUDA_ARCH_BIN CUDA_ARCH_PTX)
else()
unset(CUDA_ARCH_BIN CACHE)
@ -108,8 +108,6 @@ function(select_nvcc_arch_flags out_variable)
set(cuda_arch_bin "70")
elseif(${CUDA_ARCH_NAME} STREQUAL "Turing")
set(cuda_arch_bin "75")
elseif(${CUDA_ARCH_NAME} STREQUAL "Ampere")
set(cuda_arch_bin "80")
elseif(${CUDA_ARCH_NAME} STREQUAL "All")
set(cuda_arch_bin ${paddle_known_gpu_archs})
elseif(${CUDA_ARCH_NAME} STREQUAL "Auto")
@ -177,22 +175,14 @@ elseif (${CMAKE_CUDA_COMPILER_VERSION} LESS 10.0) # CUDA 9.x
set(paddle_known_gpu_archs ${paddle_known_gpu_archs9})
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D_MWAITXINTRIN_H_INCLUDED")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D__STRICT_ANSI__")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
elseif (${CMAKE_CUDA_COMPILER_VERSION} LESS 11.0) # CUDA 10.x
set(paddle_known_gpu_archs ${paddle_known_gpu_archs10})
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D_MWAITXINTRIN_H_INCLUDED")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D__STRICT_ANSI__")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
elseif (${CMAKE_CUDA_COMPILER_VERSION} LESS 11.2) # CUDA 11.0/11.1
elseif (${CMAKE_CUDA_COMPILER_VERSION} LESS 12.0) # CUDA 11.x
set(paddle_known_gpu_archs ${paddle_known_gpu_archs11})
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D_MWAITXINTRIN_H_INCLUDED")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D__STRICT_ANSI__")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
elseif (${CMAKE_CUDA_COMPILER_VERSION} LESS 12.0) # CUDA 11.2+
set(paddle_known_gpu_archs "${paddle_known_gpu_archs11} 86")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D_MWAITXINTRIN_H_INCLUDED")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -D__STRICT_ANSI__")
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Wno-deprecated-gpu-targets")
endif()
if (NOT ${CMAKE_CUDA_COMPILER_VERSION} LESS 10.0)
@ -208,11 +198,14 @@ select_nvcc_arch_flags(NVCC_FLAGS_EXTRA)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} ${NVCC_FLAGS_EXTRA}")
message(STATUS "NVCC_FLAGS_EXTRA: ${NVCC_FLAGS_EXTRA}")
# Set C++14 support
# Set C++11 support
set(CUDA_PROPAGATE_HOST_FLAGS OFF)
# Release/Debug flags set by cmake. Such as -O3 -g -DNDEBUG etc.
# So, don't set these flags here.
set(CMAKE_CUDA_STANDARD 14)
if (NOT WIN32) # windows msvc2015 support c++11 natively.
# -std=c++11 -fPIC not recoginize by msvc, -Xcompiler will be added by cmake.
set(CMAKE_CUDA_STANDARD 11)
endif(NOT WIN32)
# (Note) For windows, if delete /W[1-4], /W1 will be added defaultly and conflic with -w
# So replace /W[1-4] with /W0
@ -223,8 +216,6 @@ endif(WIN32)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -w")
# Set :expt-relaxed-constexpr to suppress Eigen warnings
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --expt-relaxed-constexpr")
# Set :expt-extended-lambda to enable HOSTDEVICE annotation on lambdas
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} --expt-extended-lambda")
if(WIN32)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Xcompiler \"/wd4244 /wd4267 /wd4819 \"")

@ -94,7 +94,7 @@ macro(find_cudnn_version cudnn_header_file)
"${CUDNN_MAJOR_VERSION} * 1000 +
${CUDNN_MINOR_VERSION} * 100 + ${CUDNN_PATCHLEVEL_VERSION}")
message(STATUS "Current cuDNN header is ${cudnn_header_file} "
"Current cuDNN version is v${CUDNN_MAJOR_VERSION}.${CUDNN_MINOR_VERSION}.${CUDNN_PATCHLEVEL_VERSION}. ")
"Current cuDNN version is v${CUDNN_MAJOR_VERSION}.${CUDNN_MINOR_VERSION}. ")
endif()
endif()
endmacro()

@ -12,50 +12,73 @@
# See the License for the specific language governing permissions and
# limitations under the License.
INCLUDE(ExternalProject)
SET(ASCEND_PROJECT "extern_ascend")
IF((NOT DEFINED ASCEND_VER) OR (NOT DEFINED ASCEND_URL))
MESSAGE(STATUS "use pre defined download url")
SET(ASCEND_VER "0.1.1" CACHE STRING "" FORCE)
SET(ASCEND_NAME "ascend" CACHE STRING "" FORCE)
SET(ASCEND_URL "http://paddle-ascend.bj.bcebos.com/ascend.tar.gz" CACHE STRING "" FORCE)
ENDIF()
MESSAGE(STATUS "ASCEND_NAME: ${ASCEND_NAME}, ASCEND_URL: ${ASCEND_URL}")
SET(ASCEND_SOURCE_DIR "${THIRD_PARTY_PATH}/ascend")
SET(ASCEND_DOWNLOAD_DIR "${ASCEND_SOURCE_DIR}/src/${ASCEND_PROJECT}")
SET(ASCEND_DST_DIR "ascend")
SET(ASCEND_INSTALL_ROOT "${THIRD_PARTY_PATH}/install")
SET(ASCEND_INSTALL_DIR ${ASCEND_INSTALL_ROOT}/${ASCEND_DST_DIR})
SET(ASCEND_ROOT ${ASCEND_INSTALL_DIR})
SET(ASCEND_INC_DIR ${ASCEND_ROOT}/include)
SET(ASCEND_LIB_DIR ${ASCEND_ROOT}/lib)
SET(ASCEND_LIB ${ASCEND_LIB_DIR}/libge_runner.so)
SET(ASCEND_GRAPH_LIB ${ASCEND_LIB_DIR}/libgraph.so)
SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${ASCEND_ROOT}/lib")
INCLUDE_DIRECTORIES(${ASCEND_INC_DIR})
FILE(WRITE ${ASCEND_DOWNLOAD_DIR}/CMakeLists.txt
"PROJECT(ASCEND)\n"
"cmake_minimum_required(VERSION 3.0)\n"
"install(DIRECTORY ${ASCEND_NAME}/include ${ASCEND_NAME}/lib \n"
" DESTINATION ${ASCEND_DST_DIR})\n")
ExternalProject_Add(
${ASCEND_PROJECT}
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${ASCEND_SOURCE_DIR}
DOWNLOAD_DIR ${ASCEND_DOWNLOAD_DIR}
DOWNLOAD_COMMAND wget --no-check-certificate ${ASCEND_URL} -c -q -O ${ASCEND_NAME}.tar.gz
&& tar zxvf ${ASCEND_NAME}.tar.gz
DOWNLOAD_NO_PROGRESS 1
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${ASCEND_INSTALL_ROOT}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${ASCEND_INSTALL_ROOT}
)
ADD_LIBRARY(ascend SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascend PROPERTY IMPORTED_LOCATION ${ASCEND_LIB})
ADD_LIBRARY(ascend_graph SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascend_graph PROPERTY IMPORTED_LOCATION ${ASCEND_GRAPH_LIB})
ADD_DEPENDENCIES(ascend ascend_graph ${ASCEND_PROJECT})
#NOTE: Logic is from
# https://github.com/mindspore-ai/graphengine/blob/master/CMakeLists.txt
if(DEFINED ENV{ASCEND_CUSTOM_PATH})
set(ASCEND_DIR $ENV{ASCEND_CUSTOM_PATH})
else()
set(ASCEND_DIR /usr/local/Ascend)
endif()
if(WITH_ASCEND)
set(ASCEND_DRIVER_DIR ${ASCEND_DIR}/driver/lib64)
set(ASCEND_DRIVER_COMMON_DIR ${ASCEND_DIR}/driver/lib64/common)
set(ASCEND_DRIVER_SHARE_DIR ${ASCEND_DIR}/driver/lib64/share)
set(ASCEND_RUNTIME_DIR ${ASCEND_DIR}/fwkacllib/lib64)
set(ASCEND_ATC_DIR ${ASCEND_DIR}/atc/lib64)
set(ASCEND_ACL_DIR ${ASCEND_DIR}/acllib/lib64)
set(STATIC_ACL_LIB ${ASCEND_ACL_DIR})
set(ASCEND_MS_RUNTIME_PATH ${ASCEND_RUNTIME_DIR} ${ASCEND_ACL_DIR} ${ASCEND_ATC_DIR})
set(ASCEND_MS_DRIVER_PATH ${ASCEND_DRIVER_DIR} ${ASCEND_DRIVER_COMMON_DIR})
set(ATLAS_RUNTIME_DIR ${ASCEND_DIR}/ascend-toolkit/latest/fwkacllib/lib64)
set(ATLAS_RUNTIME_INC_DIR ${ASCEND_DIR}/ascend-toolkit/latest/fwkacllib/include)
set(ATLAS_ACL_DIR ${ASCEND_DIR}/ascend-toolkit/latest/acllib/lib64)
set(ATLAS_ATC_DIR ${ASCEND_DIR}/ascend-toolkit/latest/atc/lib64)
set(ATLAS_MS_RUNTIME_PATH ${ATLAS_RUNTIME_DIR} ${ATLAS_ACL_DIR} ${ATLAS_ATC_DIR})
set(atlas_graph_lib ${ATLAS_RUNTIME_DIR}/libgraph.so)
set(atlas_ge_runner_lib ${ATLAS_RUNTIME_DIR}/libge_runner.so)
set(atlas_acl_lib ${ATLAS_RUNTIME_DIR}/libascendcl.so)
INCLUDE_DIRECTORIES(${ATLAS_RUNTIME_INC_DIR})
if(EXISTS ${ATLAS_RUNTIME_INC_DIR}/graph/ascend_string.h)
add_definitions(-DPADDLE_WITH_ASCEND_STRING)
endif()
ADD_LIBRARY(ascend_ge SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascend_ge PROPERTY IMPORTED_LOCATION ${atlas_ge_runner_lib})
ADD_LIBRARY(ascend_graph SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascend_graph PROPERTY IMPORTED_LOCATION ${atlas_graph_lib})
ADD_LIBRARY(atlas_acl SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET atlas_acl PROPERTY IMPORTED_LOCATION ${atlas_acl_lib})
add_custom_target(extern_ascend DEPENDS ascend_ge ascend_graph atlas_acl)
endif()
if(WITH_ASCEND_CL)
set(ASCEND_CL_DIR ${ASCEND_DIR}/ascend-toolkit/latest/fwkacllib/lib64)
set(ascend_hccl_lib ${ASCEND_CL_DIR}/libhccl.so)
set(ascendcl_lib ${ASCEND_CL_DIR}/libascendcl.so)
set(acl_op_compiler_lib ${ASCEND_CL_DIR}/libacl_op_compiler.so)
set(ASCEND_CL_INC_DIR ${ASCEND_DIR}/ascend-toolkit/latest/fwkacllib/include)
message(STATUS "ASCEND_CL_INC_DIR ${ASCEND_CL_INC_DIR}")
message(STATUS "ASCEND_CL_DIR ${ASCEND_CL_DIR}")
INCLUDE_DIRECTORIES(${ASCEND_CL_INC_DIR})
ADD_LIBRARY(ascendcl SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascendcl PROPERTY IMPORTED_LOCATION ${ascendcl_lib})
ADD_LIBRARY(ascend_hccl SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET ascend_hccl PROPERTY IMPORTED_LOCATION ${ascend_hccl_lib})
ADD_LIBRARY(acl_op_compiler SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET acl_op_compiler PROPERTY IMPORTED_LOCATION ${acl_op_compiler_lib})
add_custom_target(extern_ascend_cl DEPENDS ascendcl acl_op_compiler)
endif()

@ -22,13 +22,6 @@ SET(CRYPTOPP_TAG CRYPTOPP_8_2_0)
IF(WIN32)
SET(CRYPTOPP_LIBRARIES "${CRYPTOPP_INSTALL_DIR}/lib/cryptopp-static.lib" CACHE FILEPATH "cryptopp library." FORCE)
# There is a compilation parameter 'FI\"winapifamily.h\"' can't be used correctly
# with Ninja on Windows. The only difference between the patch file and original
# file is that the compilation parameters are changed to 'FIwinapifamily.h'. This
# patch command can be removed when upgrading to a higher version.
if("${CMAKE_GENERATOR}" STREQUAL "Ninja")
set(CRYPTOPP_PATCH_COMMAND ${CMAKE_COMMAND} -E copy_if_different "${PADDLE_SOURCE_DIR}/patches/cryptopp/CMakeLists.txt" "<SOURCE_DIR>/")
endif()
ELSE(WIN32)
SET(CRYPTOPP_LIBRARIES "${CRYPTOPP_INSTALL_DIR}/lib/libcryptopp.a" CACHE FILEPATH "cryptopp library." FORCE)
ENDIF(WIN32)
@ -66,7 +59,6 @@ ExternalProject_Add(
COMMAND git clone ${GIT_URL}/noloader/cryptopp-cmake "<SOURCE_DIR>/cmake"
COMMAND cd "<SOURCE_DIR>/cmake" && git checkout tags/${CRYPTOPP_TAG} -b ${CRYPTOPP_TAG}
COMMAND ${CMAKE_COMMAND} -E copy_directory "<SOURCE_DIR>/cmake/" "<SOURCE_DIR>/"
COMMAND ${CRYPTOPP_PATCH_COMMAND}
INSTALL_DIR ${CRYPTOPP_INSTALL_DIR}
CMAKE_ARGS ${CRYPTOPP_CMAKE_ARGS}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${CRYPTOPP_INSTALL_DIR}

@ -20,28 +20,28 @@ set(EIGEN_SOURCE_DIR ${THIRD_PARTY_PATH}/eigen3/src/extern_eigen3)
set(EIGEN_REPOSITORY https://gitlab.com/libeigen/eigen.git)
set(EIGEN_TAG 4da2c6b1974827b1999bab652a3d4703e1992d26)
# the recent version of eigen will cause compilation error on windows
if(WIN32)
set(EIGEN_REPOSITORY ${GIT_URL}/eigenteam/eigen-git-mirror.git)
set(EIGEN_TAG 917060c364181f33a735dc023818d5a54f60e54c)
endif()
# eigen on cuda9.1 missing header of math_funtions.hpp
# https://stackoverflow.com/questions/43113508/math-functions-hpp-not-found-when-using-cuda-with-eigen
if(WITH_ROCM_PLATFORM)
set(EIGEN_REPOSITORY ${GIT_URL}/sabreshao/hipeigen.git)
set(EIGEN_TAG 7cb2b6e5a4b4a1efe658abb215cd866c6fb2275e)
endif()
cache_third_party(extern_eigen3
REPOSITORY ${EIGEN_REPOSITORY}
TAG ${EIGEN_TAG}
DIR EIGEN_SOURCE_DIR)
if(WIN32)
add_definitions(-DEIGEN_STRONG_INLINE=inline)
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/Half.h native_src)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/Eigen/src/Core/arch/CUDA/Half.h native_dst)
# For Windows
# which will cause a compilation error in Tensor:74:
# "can not open file 'unistd.h'"
# so use following patch to solve compilation error On Windows.
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/Tensor native_src2)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/unsupported/Eigen/CXX11/Tensor native_dst2)
# For VS2015
# which will cause a compilation error in TensorBlock.h:1028:
# "syntax error"
# so use following patch to solve compilation error On Windows.
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/TensorBlock.h native_src3)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/unsupported/Eigen/CXX11/src/Tensor/TensorBlock.h native_dst3)
set(EIGEN_PATCH_COMMAND copy ${native_src} ${native_dst} /Y && copy ${native_src2} ${native_dst2} /Y && copy ${native_src3} ${native_dst3} /Y)
set(EIGEN_PATCH_COMMAND copy ${native_src} ${native_dst} /Y)
elseif(LINUX)
# For gxx=4.8, __GXX_ABI_VERSION is less than 1004
# which will cause a compilation error in Geometry_SSE.h:38:
@ -56,38 +56,43 @@ elseif(LINUX)
# add patch to avoid compilation error in c++11
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/MathFunctions.h native_src2)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/Eigen/src/Core/MathFunctions.h native_dst2)
if(WITH_ROCM)
# For HIPCC Eigen::internal::device::numeric_limits is not EIGEN_DEVICE_FUNC
# which will cause compiler error of using __host__ funciont in __host__ __device__
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/Meta.h native_src3)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/Eigen/src/Core/util/Meta.h native_dst3)
# For HIPCC Eigen::internal::scalar_sum_op<bool,bool> is not EIGEN_DEVICE_FUNC
# which will cause compiler error of using __host__ funciont in __host__ __device__
file(TO_NATIVE_PATH ${PADDLE_SOURCE_DIR}/patches/eigen/BinaryFunctors.h native_src4)
file(TO_NATIVE_PATH ${EIGEN_SOURCE_DIR}/Eigen/src/Core/functors/BinaryFunctors.h native_dst4)
set(EIGEN_PATCH_COMMAND cp ${native_src1} ${native_dst1} && cp ${native_src2} ${native_dst2} && cp ${native_src3} ${native_dst3} && cp ${native_src4} ${native_dst4})
else()
set(EIGEN_PATCH_COMMAND cp ${native_src1} ${native_dst1} && cp ${native_src2} ${native_dst2})
endif()
set(EIGEN_PATCH_COMMAND cp ${native_src1} ${native_dst1} && cp ${native_src2} ${native_dst2})
endif()
set(EIGEN_INCLUDE_DIR ${EIGEN_SOURCE_DIR})
INCLUDE_DIRECTORIES(${EIGEN_INCLUDE_DIR})
ExternalProject_Add(
extern_eigen3
${EXTERNAL_PROJECT_LOG_ARGS}
${SHALLOW_CLONE}
"${EIGEN_DOWNLOAD_CMD}"
PREFIX ${EIGEN_PREFIX_DIR}
SOURCE_DIR ${EIGEN_SOURCE_DIR}
UPDATE_COMMAND ""
PATCH_COMMAND ${EIGEN_PATCH_COMMAND}
CONFIGURE_COMMAND ""
BUILD_COMMAND ""
INSTALL_COMMAND ""
TEST_COMMAND ""
)
if(WITH_AMD_GPU)
ExternalProject_Add(
extern_eigen3
${EXTERNAL_PROJECT_LOG_ARGS}
${SHALLOW_CLONE}
"${EIGEN_DOWNLOAD_CMD}"
PREFIX ${EIGEN_PREFIX_DIR}
SOURCE_DIR ${EIGEN_SOURCE_DIR}
UPDATE_COMMAND ""
PATCH_COMMAND ${EIGEN_PATCH_COMMAND}
CONFIGURE_COMMAND ""
BUILD_COMMAND ""
INSTALL_COMMAND ""
TEST_COMMAND ""
)
else()
ExternalProject_Add(
extern_eigen3
${EXTERNAL_PROJECT_LOG_ARGS}
${SHALLOW_CLONE}
"${EIGEN_DOWNLOAD_CMD}"
PREFIX ${EIGEN_PREFIX_DIR}
SOURCE_DIR ${EIGEN_SOURCE_DIR}
UPDATE_COMMAND ""
PATCH_COMMAND ${EIGEN_PATCH_COMMAND}
CONFIGURE_COMMAND ""
BUILD_COMMAND ""
INSTALL_COMMAND ""
TEST_COMMAND ""
)
endif()
add_library(eigen3 INTERFACE)

@ -30,6 +30,8 @@ ENDIF(WIN32)
INCLUDE_DIRECTORIES(${GFLAGS_INCLUDE_DIR})
set(GFLAGS_NAMESPACE "paddle_gflags")
cache_third_party(extern_gflags
REPOSITORY ${GFLAGS_REPOSITORY}
TAG ${GFLAGS_TAG}
@ -57,6 +59,7 @@ ExternalProject_Add(
-DCMAKE_POSITION_INDEPENDENT_CODE=ON
-DBUILD_TESTING=OFF
-DCMAKE_BUILD_TYPE=${THIRD_PARTY_BUILD_TYPE}
-DGFLAGS_NAMESPACE=${GFLAGS_NAMESPACE}
${EXTERNAL_OPTIONAL_ARGS}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${GFLAGS_INSTALL_DIR}
-DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON

@ -42,7 +42,7 @@ ExternalProject_Add(
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
BUILD_COMMAND mkdir -p ${GLOO_SOURCE_DIR}/build
&& cd ${GLOO_SOURCE_DIR}/build && cmake .. && make
&& cd ${GLOO_SOURCE_DIR}/build && cmake .. -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS} && make
&& mkdir -p ${GLOO_LIBRARY_DIR} ${GLOO_INCLUDE_DIR}/gloo
INSTALL_COMMAND ${CMAKE_COMMAND} -E copy ${GLOO_SOURCE_DIR}/build/gloo/libgloo.a ${GLOO_LIBRARY_DIR}
COMMAND ${CMAKE_COMMAND} -E copy_directory "${GLOO_SOURCE_DIR}/gloo/" "${GLOO_INCLUDE_DIR}/gloo"

@ -58,7 +58,6 @@ if (NOT LITE_SOURCE_DIR OR NOT LITE_BINARY_DIR)
-DLITE_BUILD_EXTRA=ON
-DLITE_WITH_XPU=${LITE_WITH_XPU}
-DXPU_SDK_ROOT=${XPU_SDK_ROOT}
-DLITE_WITH_CODE_META_INFO=OFF
-DLITE_WITH_ARM=ON)
ExternalProject_Add(
${LITE_PROJECT}
@ -100,7 +99,6 @@ if (NOT LITE_SOURCE_DIR OR NOT LITE_BINARY_DIR)
-DCUDA_ARCH_NAME=${CUDA_ARCH_NAME}
-DLITE_WITH_XPU=${LITE_WITH_XPU}
-DXPU_SDK_ROOT=${XPU_SDK_ROOT}
-DLITE_WITH_CODE_META_INFO=OFF
-DLITE_WITH_ARM=OFF)
ExternalProject_Add(

@ -20,7 +20,7 @@ SET(MKLDNN_SOURCE_DIR ${THIRD_PARTY_PATH}/mkldnn/src/extern_mkldnn)
SET(MKLDNN_INSTALL_DIR ${THIRD_PARTY_PATH}/install/mkldnn)
SET(MKLDNN_INC_DIR "${MKLDNN_INSTALL_DIR}/include" CACHE PATH "mkldnn include directory." FORCE)
SET(MKLDNN_REPOSITORY ${GIT_URL}/oneapi-src/oneDNN.git)
SET(MKLDNN_TAG 72efa005effb49595933e033cc732f215ef0445a)
SET(MKLDNN_TAG a18f78f1f058437e9efee403655d671633360f98)
# Introduce variables:
# * CMAKE_INSTALL_LIBDIR

@ -198,8 +198,19 @@ FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
"-Dprotobuf_MSVC_STATIC_RUNTIME=${MSVC_STATIC_CRT}")
ENDIF()
SET(PROTOBUF_REPOSITORY ${GIT_URL}/protocolbuffers/protobuf.git)
SET(PROTOBUF_TAG 9f75c5aa851cd877fb0d93ccc31b8567a6706546)
if(WITH_ASCEND AND NOT WITH_ASCEND_CXX11)
SET(PROTOBUF_REPOSITORY https://gitee.com/tianjianhe/protobuf.git)
SET(PROTOBUF_TAG v3.8.0)
elseif(WITH_ASCEND_CL AND NOT WITH_ASCEND_CXX11)
SET(PROTOBUF_REPOSITORY https://gitee.com/tianjianhe/protobuf.git)
SET(PROTOBUF_TAG v3.8.0)
else()
SET(PROTOBUF_REPOSITORY https://gitee.com/tianjianhe/protobuf.git)
SET(PROTOBUF_TAG v3.8.0)
# SET(PROTOBUF_REPOSITORY ${GIT_URL}/protocolbuffers/protobuf.git)
# SET(PROTOBUF_TAG 9f75c5aa851cd877fb0d93ccc31b8567a6706546)
endif()
cache_third_party(${TARGET_NAME}
REPOSITORY ${PROTOBUF_REPOSITORY}
@ -234,7 +245,7 @@ FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
)
ENDFUNCTION()
SET(PROTOBUF_VERSION 3.1.0)
SET(PROTOBUF_VERSION 3.8.0)
IF(NOT PROTOBUF_FOUND)
build_protobuf(extern_protobuf FALSE)
@ -250,8 +261,5 @@ IF(NOT PROTOBUF_FOUND)
SET(PROTOBUF_PROTOC_EXECUTABLE ${extern_protobuf_PROTOC_EXECUTABLE}
CACHE FILEPATH "protobuf executable." FORCE)
# `EXTERN_PROTOBUF_DEPEND` used in cmake function `proto_library` to ensure
# `protoc.exe` existed before calling it.
set(EXTERN_PROTOBUF_DEPEND extern_protobuf)
PROMPT_PROTOBUF_LIB(extern_protobuf)
ENDIF(NOT PROTOBUF_FOUND)

@ -16,7 +16,7 @@ INCLUDE(ExternalProject)
SET(THREADPOOL_PREFIX_DIR ${THIRD_PARTY_PATH}/threadpool)
SET(THREADPOOL_SOURCE_DIR ${THIRD_PARTY_PATH}/threadpool/src/extern_threadpool)
SET(THREADPOOL_REPOSITORY ${GIT_URL}/progschj/ThreadPool.git)
SET(THREADPOOL_REPOSITORY https://gitee.com/tianjianhe/ThreadPool.git)
SET(THREADPOOL_TAG 9a42ec1329f259a5f4881a291db1dcb8f2ad9040)
cache_third_party(extern_threadpool

@ -14,15 +14,12 @@
INCLUDE(ExternalProject)
IF(WITH_ROCM)
add_definitions(-DWARPCTC_WITH_HIP)
ENDIF()
SET(WARPCTC_PREFIX_DIR ${THIRD_PARTY_PATH}/warpctc)
SET(WARPCTC_SOURCE_DIR ${THIRD_PARTY_PATH}/warpctc/src/extern_warpctc)
SET(WARPCTC_INSTALL_DIR ${THIRD_PARTY_PATH}/install/warpctc)
set(WARPCTC_REPOSITORY https://gitee.com/tianjianhe/warp-ctc.git)
set(WARPCTC_TAG 95a461eddeabd51099ef059dcfada1117eb1bfb8)
set(WARPCTC_REPOSITORY ${GIT_URL}/baidu-research/warp-ctc.git)
set(WARPCTC_TAG c690fc5755abbdbdc98ef78d51ec10a6748a8cd1)
SET(WARPCTC_INCLUDE_DIR "${WARPCTC_INSTALL_DIR}/include"
CACHE PATH "Warp-ctc Directory" FORCE)
@ -53,15 +50,14 @@ ExternalProject_Add(
BUILD_ALWAYS 1
CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
-DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
-DCMAKE_C_FLAGS=$<FILTER:${CMAKE_C_FLAGS},EXCLUDE,/Zc:inline>
-DCMAKE_C_FLAGS_DEBUG=$<FILTER:${CMAKE_C_FLAGS_DEBUG},EXCLUDE,/Zc:inline>
-DCMAKE_C_FLAGS_RELEASE=$<FILTER:${CMAKE_C_FLAGS_RELEASE},EXCLUDE,/Zc:inline>
-DCMAKE_CXX_FLAGS=$<FILTER:${CMAKE_CXX_FLAGS},EXCLUDE,/Zc:inline>
-DCMAKE_CXX_FLAGS_RELEASE=$<FILTER:${CMAKE_CXX_FLAGS_RELEASE},EXCLUDE,/Zc:inline>
-DCMAKE_CXX_FLAGS_DEBUG=$<FILTER:${CMAKE_CXX_FLAGS_DEBUG},EXCLUDE,/Zc:inline>
-DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
-DCMAKE_C_FLAGS_DEBUG=${CMAKE_C_FLAGS_DEBUG}
-DCMAKE_C_FLAGS_RELEASE=${CMAKE_C_FLAGS_RELEASE}
"-DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}"
-DCMAKE_CXX_FLAGS_RELEASE=${CMAKE_CXX_FLAGS_RELEASE}
-DCMAKE_CXX_FLAGS_DEBUG=${CMAKE_CXX_FLAGS_DEBUG}
-DCMAKE_INSTALL_PREFIX=${WARPCTC_INSTALL_DIR}
-DWITH_GPU=${WITH_GPU}
-DWITH_ROCM=${WITH_ROCM}
-DWITH_OMP=${USE_OMP}
-DWITH_TORCH=OFF
-DCMAKE_DISABLE_FIND_PACKAGE_Torch=ON

@ -20,7 +20,7 @@ SET(XBYAK_SOURCE_DIR ${THIRD_PARTY_PATH}/xbyak/src/extern_xbyak)
set(XBYAK_INSTALL_ROOT ${THIRD_PARTY_PATH}/install/xbyak)
set(XBYAK_INC_DIR ${XBYAK_INSTALL_ROOT}/include)
set(XBYAK_REPOSITORY ${GIT_URL}/herumi/xbyak.git)
set(XBYAK_TAG v5.81) # Dec 19, 2019
set(XBYAK_TAG v5.661) # Jul 26th
include_directories(${XBYAK_INC_DIR})
include_directories(${XBYAK_INC_DIR}/xbyak)
@ -44,7 +44,7 @@ ExternalProject_Add(
DEPENDS ""
PREFIX ${XBYAK_PREFIX_DIR}
SOURCE_DIR ${XBYAK_SOURCE_DIR}
# UPDATE_COMMAND ""
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${XBYAK_INSTALL_ROOT}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${XBYAK_INSTALL_ROOT}
)

@ -4,55 +4,49 @@ endif()
INCLUDE(ExternalProject)
SET(XPU_PROJECT "extern_xpu")
SET(XPU_API_LIB_NAME "libxpuapi.so")
SET(XPU_RT_LIB_NAME "libxpurt.so")
if(NOT XPU_SDK_ROOT)
if (WITH_AARCH64)
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/aarch64/xpu_2021_01_13.tar.gz" CACHE STRING "" FORCE)
elseif(WITH_SUNWAY)
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/sunway/xpu_2021_01_13.tar.gz" CACHE STRING "" FORCE)
else()
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/xpu_2021_03_30.tar.gz" CACHE STRING "" FORCE)
endif()
if (WITH_AARCH64)
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/aarch64/xpu_2020_1229.tar.gz" CACHE STRING "" FORCE)
elseif(WITH_SUNWAY)
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/sunway/xpu_2020_1227.tar.gz" CACHE STRING "" FORCE)
else()
SET(XPU_URL "https://baidu-kunlun-public.su.bcebos.com/paddle_depence/xpu_2021_01_13.tar.gz" CACHE STRING "" FORCE)
endif()
SET(XPU_SOURCE_DIR "${THIRD_PARTY_PATH}/xpu")
SET(XPU_DOWNLOAD_DIR "${XPU_SOURCE_DIR}/src/${XPU_PROJECT}")
SET(XPU_INSTALL_DIR "${THIRD_PARTY_PATH}/install/xpu")
SET(XPU_API_INC_DIR "${THIRD_PARTY_PATH}/install/xpu/include")
SET(XPU_LIB_DIR "${THIRD_PARTY_PATH}/install/xpu/lib")
SET(XPU_SOURCE_DIR "${THIRD_PARTY_PATH}/xpu")
SET(XPU_DOWNLOAD_DIR "${XPU_SOURCE_DIR}/src/${XPU_PROJECT}")
SET(XPU_INSTALL_DIR "${THIRD_PARTY_PATH}/install/xpu")
SET(XPU_API_INC_DIR "${THIRD_PARTY_PATH}/install/xpu/include")
SET(XPU_LIB_DIR "${THIRD_PARTY_PATH}/install/xpu/lib")
SET(XPU_API_LIB_NAME "libxpuapi.so")
SET(XPU_RT_LIB_NAME "libxpurt.so")
SET(XPU_API_LIB "${XPU_LIB_DIR}/${XPU_API_LIB_NAME}")
SET(XPU_RT_LIB "${XPU_LIB_DIR}/${XPU_RT_LIB_NAME}")
SET(XPU_API_LIB "${XPU_LIB_DIR}/${XPU_API_LIB_NAME}")
SET(XPU_RT_LIB "${XPU_LIB_DIR}/${XPU_RT_LIB_NAME}")
SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${XPU_INSTALL_DIR}/lib")
SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${XPU_INSTALL_DIR}/lib")
INCLUDE_DIRECTORIES(${XPU_API_INC_DIR})
FILE(WRITE ${XPU_DOWNLOAD_DIR}/CMakeLists.txt
"PROJECT(XPU)\n"
"cmake_minimum_required(VERSION 3.0)\n"
"install(DIRECTORY xpu/include xpu/lib \n"
" DESTINATION ${XPU_INSTALL_DIR})\n")
FILE(WRITE ${XPU_DOWNLOAD_DIR}/CMakeLists.txt
"PROJECT(XPU)\n"
"cmake_minimum_required(VERSION 3.0)\n"
"install(DIRECTORY xpu/include xpu/lib \n"
" DESTINATION ${XPU_INSTALL_DIR})\n")
ExternalProject_Add(
${XPU_PROJECT}
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${XPU_SOURCE_DIR}
DOWNLOAD_DIR ${XPU_DOWNLOAD_DIR}
DOWNLOAD_COMMAND wget --no-check-certificate ${XPU_URL} -c -q -O xpu.tar.gz
&& tar xvf xpu.tar.gz
DOWNLOAD_NO_PROGRESS 1
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${XPU_INSTALL_ROOT}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${XPU_INSTALL_ROOT}
)
else()
SET(XPU_API_INC_DIR "${XPU_SDK_ROOT}/XTDK/include/")
SET(XPU_API_LIB "${XPU_SDK_ROOT}/XTDK/shlib/libxpuapi.so")
SET(XPU_RT_LIB "${XPU_SDK_ROOT}/XTDK/runtime/shlib/libxpurt.so")
SET(XPU_LIB_DIR "${XPU_SDK_ROOT}/XTDK/shlib/")
endif()
ExternalProject_Add(
${XPU_PROJECT}
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${XPU_SOURCE_DIR}
DOWNLOAD_DIR ${XPU_DOWNLOAD_DIR}
DOWNLOAD_COMMAND wget --no-check-certificate ${XPU_URL} -c -q -O xpu.tar.gz
&& tar xvf xpu.tar.gz
DOWNLOAD_NO_PROGRESS 1
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${XPU_INSTALL_ROOT}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${XPU_INSTALL_ROOT}
)
INCLUDE_DIRECTORIES(${XPU_API_INC_DIR})
ADD_LIBRARY(shared_xpuapi SHARED IMPORTED GLOBAL)
set_property(TARGET shared_xpuapi PROPERTY IMPORTED_LOCATION "${XPU_API_LIB}")
@ -75,14 +69,4 @@ else(WITH_XPU_BKCL)
TARGET_LINK_LIBRARIES(xpulib ${XPU_API_LIB} ${XPU_RT_LIB} )
endif(WITH_XPU_BKCL)
if(NOT XPU_SDK_ROOT)
ADD_DEPENDENCIES(xpulib ${XPU_PROJECT})
else()
ADD_CUSTOM_TARGET(extern_xpu DEPENDS xpulib)
endif()
# Ensure that xpu/api.h can be included without dependency errors.
file(GENERATE OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/.xpu_headers_dummy.cc CONTENT "")
add_library(xpu_headers_dummy STATIC ${CMAKE_CURRENT_BINARY_DIR}/.xpu_headers_dummy.cc)
add_dependencies(xpu_headers_dummy extern_xpu)
link_libraries(xpu_headers_dummy)
ADD_DEPENDENCIES(xpulib ${XPU_PROJECT})

@ -4,10 +4,10 @@ include(CheckCCompilerFlag)
include(CheckCXXSymbolExists)
include(CheckTypeSize)
function(CheckCompilerCXX14Flag)
function(CheckCompilerCXX11Flag)
if(CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
if(${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS 5.4)
message(FATAL_ERROR "Unsupported GCC version. GCC >= 5.4 required.")
if(${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS 4.8)
message(FATAL_ERROR "Unsupported GCC version. GCC >= 4.8 required.")
elseif(${CMAKE_CXX_COMPILER_VERSION} VERSION_GREATER 8.2)
message(WARNING "Found GCC ${CMAKE_CXX_COMPILER_VERSION} which is too high, recommended to use GCC 8.2")
endif()
@ -20,15 +20,23 @@ function(CheckCompilerCXX14Flag)
message(FATAL_ERROR "Unsupported AppleClang version. AppleClang >= 5.1 required.")
endif()
else()
if (${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS 3.4)
message(FATAL_ERROR "Unsupported Clang version. Clang >= 3.4 required.")
if (${CMAKE_CXX_COMPILER_VERSION} VERSION_LESS 3.3)
message(FATAL_ERROR "Unsupported Clang version. Clang >= 3.3 required.")
endif()
endif()
endif()
endfunction()
CheckCompilerCXX14Flag()
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++14")
CheckCompilerCXX11Flag()
if (WITH_GPU)
if (${CMAKE_CUDA_COMPILER_VERSION} GREATER_EQUAL 11.0)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++14")
else()
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
endif()
else()
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
endif()
# safe_set_flag
#
# Set a compile flag only if compiler is support
@ -143,11 +151,13 @@ set(COMMON_FLAGS
-Wno-error=int-in-bool-context # Warning in Eigen gcc 7.2
-Wimplicit-fallthrough=0 # Warning in tinyformat.h
-Wno-error=maybe-uninitialized # Warning in boost gcc 7.2
-Wno-error=nonnull-compare # Warning in boost gcc 7.2
-Wno-error=address # Warning in boost gcc 7.2
${fsanitize}
)
if(NOT APPLE)
if((${CMAKE_CXX_COMPILER_VERSION} VERSION_GREATER 8.0) OR (WITH_ROCM))
if((${CMAKE_CXX_COMPILER_VERSION} VERSION_GREATER 8.0) OR (WITH_ROCM_PLATFORM AND ${CMAKE_CXX_COMPILER_VERSION} VERSION_GREATER 7.3))
set(COMMON_FLAGS
${COMMON_FLAGS}
-Wno-format-truncation # Warning in boost gcc 8.2
@ -205,17 +215,5 @@ foreach(flag ${GPU_COMMON_FLAGS})
safe_set_nvflag(${flag})
endforeach()
if(WITH_GPU)
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} ${SAFE_GPU_COMMON_FLAGS}")
endif()
if(WITH_ROCM)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} ${SAFE_GPU_COMMON_FLAGS}")
endif()
# Disable -Werror, otherwise the compile will fail for rocblas_gemm_ex
if(WITH_ROCM)
string (REPLACE "-Werror" "-Wno-error" CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS})
string (REPLACE "-Werror" "-Wno-error" CMAKE_C_FLAGS ${CMAKE_C_FLAGS})
endif()
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} ${SAFE_GPU_COMMON_FLAGS}")

File diff suppressed because it is too large Load Diff

@ -1,87 +1,104 @@
if(NOT WITH_ROCM)
if(NOT WITH_ROCM_PLATFORM)
return()
endif()
if(NOT DEFINED ENV{ROCM_PATH})
set(ROCM_PATH "/opt/rocm" CACHE PATH "Path to which ROCm has been installed")
set(HIP_PATH ${ROCM_PATH}/hip CACHE PATH "Path to which HIP has been installed")
set(HIP_CLANG_PATH ${ROCM_PATH}/llvm/bin CACHE PATH "Path to which clang has been installed")
else()
set(ROCM_PATH $ENV{ROCM_PATH} CACHE PATH "Path to which ROCm has been installed")
set(HIP_PATH ${ROCM_PATH}/hip CACHE PATH "Path to which HIP has been installed")
set(HIP_CLANG_PATH ${ROCM_PATH}/llvm/bin CACHE PATH "Path to which clang has been installed")
include_directories("${ROCM_PATH}/include")
include_directories("${ROCM_PATH}/hip/include")
include_directories("${ROCM_PATH}/miopen/include")
include_directories("${ROCM_PATH}/hipblas/include")
include_directories("${ROCM_PATH}/rocblas/include")
include_directories("${ROCM_PATH}/hiprand/include")
include_directories("${ROCM_PATH}/rocrand/include")
include_directories("${ROCM_PATH}/rccl/include")
include_directories("${ROCM_PATH}/rocthrust/include/")
include_directories("${ROCM_PATH}/hipcub/include/")
include_directories("${ROCM_PATH}/rocprim/include/")
include_directories("${ROCM_PATH}/hipsparse/include/")
include_directories("${ROCM_PATH}/rocsparse/include/")
include_directories("${ROCM_PATH}/rocfft/include/")
set(HIP_CLANG_PARALLEL_BUILD_COMPILE_OPTIONS "")
set(HIP_CLANG_PARALLEL_BUILD_LINK_OPTIONS "")
# now default is clang
set(HIP_COMPILER "clang")
list(APPEND EXTERNAL_LIBS "-L${ROCM_PATH}/lib/ -lhip_hcc")
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -fPIC -DPADDLE_WITH_HIP -DEIGEN_USE_HIP -DEIGEN_USE_GPU -D__HIP_NO_HALF_CONVERSIONS__ -std=c++11 --amdgpu-target=gfx906" )
if(WITH_RCCL)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_WITH_RCCL")
endif()
set(CMAKE_MODULE_PATH "${HIP_PATH}/cmake" ${CMAKE_MODULE_PATH})
find_package(HIP REQUIRED)
include_directories(${ROCM_PATH}/include)
message(STATUS "HIP version: ${HIP_VERSION}")
message(STATUS "HIP_CLANG_PATH: ${HIP_CLANG_PATH}")
macro(find_package_and_include PACKAGE_NAME)
find_package("${PACKAGE_NAME}" REQUIRED)
include_directories("${ROCM_PATH}/${PACKAGE_NAME}/include")
message(STATUS "${PACKAGE_NAME} version: ${${PACKAGE_NAME}_VERSION}")
endmacro()
find_package_and_include(miopen)
find_package_and_include(rocblas)
find_package_and_include(hiprand)
find_package_and_include(rocrand)
find_package_and_include(rccl)
find_package_and_include(rocthrust)
find_package_and_include(hipcub)
find_package_and_include(rocprim)
find_package_and_include(hipsparse)
find_package_and_include(rocsparse)
find_package_and_include(rocfft)
# set CXX flags for HIP
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -D__HIP_PLATFORM_HCC__")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D__HIP_PLATFORM_HCC__")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_HIP")
set(THRUST_DEVICE_SYSTEM THRUST_DEVICE_SYSTEM_HIP)
# define HIP_CXX_FLAGS
list(APPEND HIP_CXX_FLAGS -fPIC)
list(APPEND HIP_CXX_FLAGS -D__HIP_PLATFORM_HCC__=1)
# Note(qili93): HIP has compile conflicts of float16.h as platform::float16 overload std::is_floating_point and std::is_integer
list(APPEND HIP_CXX_FLAGS -D__HIP_NO_HALF_CONVERSIONS__=1)
list(APPEND HIP_CXX_FLAGS -Wno-macro-redefined)
list(APPEND HIP_CXX_FLAGS -Wno-inconsistent-missing-override)
list(APPEND HIP_CXX_FLAGS -Wno-exceptions)
list(APPEND HIP_CXX_FLAGS -Wno-shift-count-negative)
list(APPEND HIP_CXX_FLAGS -Wno-shift-count-overflow)
list(APPEND HIP_CXX_FLAGS -Wno-unused-command-line-argument)
list(APPEND HIP_CXX_FLAGS -Wno-duplicate-decl-specifier)
list(APPEND HIP_CXX_FLAGS -Wno-implicit-int-float-conversion)
list(APPEND HIP_CXX_FLAGS -Wno-pass-failed)
list(APPEND HIP_CXX_FLAGS -DTHRUST_DEVICE_SYSTEM=THRUST_DEVICE_SYSTEM_HIP)
list(APPEND HIP_CXX_FLAGS -std=c++14)
if(CMAKE_BUILD_TYPE MATCHES Debug)
list(APPEND HIP_CXX_FLAGS -g2)
list(APPEND HIP_CXX_FLAGS -O0)
list(APPEND HIP_HIPCC_FLAGS -fdebug-info-for-profiling)
endif(CMAKE_BUILD_TYPE MATCHES Debug)
set(HIP_HCC_FLAGS ${HIP_CXX_FLAGS})
set(HIP_CLANG_FLAGS ${HIP_CXX_FLAGS})
# Ask hcc to generate device code during compilation so we can use
# host linker to link.
list(APPEND HIP_HCC_FLAGS -fno-gpu-rdc)
list(APPEND HIP_HCC_FLAGS --amdgpu-target=gfx906)
list(APPEND HIP_CLANG_FLAGS -fno-gpu-rdc)
list(APPEND HIP_CLANG_FLAGS --amdgpu-target=gfx906)
if(HIP_COMPILER STREQUAL clang)
set(hip_library_name amdhip64)
else()
set(hip_library_name hip_hcc)
if(NOT WITH_PYTHON)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_NO_PYTHON")
endif(NOT WITH_PYTHON)
if(WITH_DSO)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_USE_DSO")
endif(WITH_DSO)
if(WITH_TESTING)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_WITH_TESTING")
endif(WITH_TESTING)
if(WITH_DISTRIBUTE)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_WITH_DISTRIBUTE")
endif(WITH_DISTRIBUTE)
if(WITH_GRPC)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_WITH_GRPC")
endif(WITH_GRPC)
if(WITH_MKLDNN)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DPADDLE_WITH_MKLDNN")
endif(WITH_MKLDNN)
set(HIP_HIPCC_FLAGS "${HIP_HIPCC_FLAGS} -DANY_IMPL_ANY_CAST_MOVEABLE")
if(CMAKE_BUILD_TYPE STREQUAL "Debug")
list(APPEND HIP_HIPCC_FLAGS ${CMAKE_CXX_FLAGS_DEBUG})
elseif(CMAKE_BUILD_TYPE STREQUAL "RelWithDebInfo")
list(APPEND HIP_HIPCC_FLAGS ${CMAKE_CXX_FLAGS_RELWITHDEBINFO})
elseif(CMAKE_BUILD_TYPE STREQUAL "MinSizeRel")
list(APPEND HIP_HIPCC_FLAGS ${CMAKE_CXX_FLAGS_MINSIZEREL})
endif()
message(STATUS "HIP library name: ${hip_library_name}")
# set HIP link libs
find_library(ROCM_HIPRTC_LIB ${hip_library_name} HINTS ${HIP_PATH}/lib)
message(STATUS "ROCM_HIPRTC_LIB: ${ROCM_HIPRTC_LIB}")
if("${HIP_COMPILER}" STREQUAL "hcc")
if("x${HCC_HOME}" STREQUAL "x")
set(HCC_HOME "${ROCM_PATH}/hcc")
endif()
set(CMAKE_HIP_LINK_EXECUTABLE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HCC_HOME} <FLAGS> <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES> -ldl --amdgpu-target=gfx906 ")
set(CMAKE_HIP_CREATE_SHARED_LIBRARY "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HCC_HOME} <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES> -shared --amdgpu-target=gfx906")
set(CMAKE_HIP_CREATE_SHARED_MODULE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HCC_HOME} <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES> -shared --amdgpu-target=gfx906")
elseif("${HIP_COMPILER}" STREQUAL "clang")
if("x${HIP_CLANG_PATH}" STREQUAL "x")
set(HIP_CLANG_PATH "${ROCM_PATH}/llvm/bin")
endif()
#Number of parallel jobs by default is 1
if(NOT DEFINED HIP_CLANG_NUM_PARALLEL_JOBS)
set(HIP_CLANG_NUM_PARALLEL_JOBS 1)
endif()
#Add support for parallel build and link
if(${CMAKE_CXX_COMPILER_ID} STREQUAL "Clang")
check_cxx_compiler_flag("-parallel-jobs=1" HIP_CLANG_SUPPORTS_PARALLEL_JOBS)
endif()
if(HIP_CLANG_NUM_PARALLEL_JOBS GREATER 1)
if(${HIP_CLANG_SUPPORTS_PARALLEL_JOBS})
set(HIP_CLANG_PARALLEL_BUILD_COMPILE_OPTIONS "-parallel-jobs=${HIP_CLANG_NUM_PARALLEL_JOBS} -Wno-format-nonliteral")
set(HIP_CLANG_PARALLEL_BUILD_LINK_OPTIONS "-parallel-jobs=${HIP_CLANG_NUM_PARALLEL_JOBS}")
else()
message("clang compiler doesn't support parallel jobs")
endif()
endif()
# Set the CMake Flags to use the HIP-Clang Compiler.
set(CMAKE_HIP_CREATE_SHARED_LIBRARY "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HIP_CLANG_PATH} ${HIP_CLANG_PARALLEL_BUILD_LINK_OPTIONS} <CMAKE_SHARED_LIBRARY_CXX_FLAGS> <LANGUAGE_COMPILE_FLAGS> <LINK_FLAGS> <CMAKE_SHARED_LIBRARY_CREATE_CXX_FLAGS> <SONAME_FLAG><TARGET_SONAME> -o <TARGET> <OBJECTS> <LINK_LIBRARIES> --amdgpu-target=gfx906")
set(CMAKE_HIP_CREATE_SHARED_MODULE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HIP_CLANG_PATH} ${HIP_CLANG_PARALLEL_BUILD_LINK_OPTIONS} <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> <SONAME_FLAG><TARGET_SONAME> -o <TARGET> <LINK_LIBRARIES> -shared --amdgpu-target=gfx906" )
set(CMAKE_HIP_LINK_EXECUTABLE "${HIP_HIPCC_CMAKE_LINKER_HELPER} ${HIP_CLANG_PATH} ${HIP_CLANG_PARALLEL_BUILD_LINK_OPTIONS} <FLAGS> <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES> -ldl --amdgpu-target=gfx906")
endif()

@ -137,7 +137,7 @@ function(copy_part_of_thrid_party TARGET DST)
endfunction()
# inference library for only inference
set(inference_lib_deps third_party paddle_inference paddle_inference_c paddle_inference_shared paddle_inference_c_shared)
set(inference_lib_deps third_party paddle_fluid paddle_fluid_c paddle_fluid_shared paddle_fluid_c_shared)
add_custom_target(inference_lib_dist DEPENDS ${inference_lib_deps})
@ -164,20 +164,20 @@ copy_part_of_thrid_party(inference_lib_dist ${PADDLE_INFERENCE_INSTALL_DIR})
set(src_dir "${PADDLE_SOURCE_DIR}/paddle/fluid")
if(WIN32)
if(WITH_STATIC_LIB)
set(paddle_inference_lib $<TARGET_FILE_DIR:paddle_inference>/libpaddle_inference.lib
$<TARGET_FILE_DIR:paddle_inference>/paddle_inference.*)
set(paddle_fluid_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/libpaddle_fluid.lib
${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/paddle_fluid.*)
else()
set(paddle_inference_lib $<TARGET_FILE_DIR:paddle_inference_shared>/paddle_inference.dll
$<TARGET_FILE_DIR:paddle_inference_shared>/paddle_inference.lib)
set(paddle_fluid_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/paddle_fluid.dll
${PADDLE_BINARY_DIR}/paddle/fluid/inference/${CMAKE_BUILD_TYPE}/paddle_fluid.lib)
endif()
copy(inference_lib_dist
SRCS ${src_dir}/inference/api/paddle_*.h ${paddle_inference_lib}
SRCS ${src_dir}/inference/api/paddle_*.h ${paddle_fluid_lib}
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/lib
${PADDLE_INFERENCE_INSTALL_DIR}/paddle/lib)
else(WIN32)
set(paddle_inference_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/libpaddle_inference.*)
set(paddle_fluid_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/libpaddle_fluid.*)
copy(inference_lib_dist
SRCS ${src_dir}/inference/api/paddle_*.h ${paddle_inference_lib}
SRCS ${src_dir}/inference/api/paddle_*.h ${paddle_fluid_lib}
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/lib)
endif(WIN32)
@ -189,19 +189,6 @@ copy(inference_lib_dist
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include/crypto/)
include_directories(${CMAKE_BINARY_DIR}/../paddle/fluid/framework/io)
copy(inference_lib_dist
SRCS ${PADDLE_SOURCE_DIR}/paddle/fluid/extension/include/*
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include/experimental/)
copy(inference_lib_dist
SRCS ${PADDLE_SOURCE_DIR}/paddle/fluid/platform/complex64.h
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include/experimental/)
copy(inference_lib_dist
SRCS ${PADDLE_SOURCE_DIR}/paddle/fluid/platform/complex128.h
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include/experimental/)
copy(inference_lib_dist
SRCS ${PADDLE_SOURCE_DIR}/paddle/fluid/platform/float16.h
DSTS ${PADDLE_INFERENCE_INSTALL_DIR}/paddle/include/experimental/)
# CAPI inference library for only inference
set(PADDLE_INFERENCE_C_INSTALL_DIR "${CMAKE_BINARY_DIR}/paddle_inference_c_install_dir" CACHE STRING
"A path setting CAPI paddle inference shared")
@ -209,13 +196,13 @@ copy_part_of_thrid_party(inference_lib_dist ${PADDLE_INFERENCE_C_INSTALL_DIR})
set(src_dir "${PADDLE_SOURCE_DIR}/paddle/fluid")
if(WIN32)
set(paddle_inference_c_lib $<TARGET_FILE_DIR:paddle_inference_c>/paddle_inference_c.*)
set(paddle_fluid_c_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/capi/${CMAKE_BUILD_TYPE}/paddle_fluid_c.*)
else(WIN32)
set(paddle_inference_c_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/capi/libpaddle_inference_c.*)
set(paddle_fluid_c_lib ${PADDLE_BINARY_DIR}/paddle/fluid/inference/capi/libpaddle_fluid_c.*)
endif(WIN32)
copy(inference_lib_dist
SRCS ${src_dir}/inference/capi/paddle_c_api.h ${paddle_inference_c_lib}
SRCS ${src_dir}/inference/capi/paddle_c_api.h ${paddle_fluid_c_lib}
DSTS ${PADDLE_INFERENCE_C_INSTALL_DIR}/paddle/include ${PADDLE_INFERENCE_C_INSTALL_DIR}/paddle/lib)
# fluid library for both train and inference
@ -226,12 +213,12 @@ set(dst_dir "${PADDLE_INSTALL_DIR}/paddle/fluid")
set(module "inference")
if(WIN32)
copy(fluid_lib_dist
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/api/paddle_*.h ${paddle_inference_lib}
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/api/paddle_*.h ${paddle_fluid_lib}
DSTS ${dst_dir}/${module} ${dst_dir}/${module} ${dst_dir}/${module} ${dst_dir}/${module}
)
else()
copy(fluid_lib_dist
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/api/paddle_*.h ${paddle_inference_lib}
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/api/paddle_*.h ${paddle_fluid_lib}
DSTS ${dst_dir}/${module} ${dst_dir}/${module} ${dst_dir}/${module}
)
endif()

@ -18,10 +18,6 @@ if(NOT WIN32)
set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O2 -g -DNDEBUG")
set(CMAKE_CXX_FLAGS_MINSIZEREL "-Os -DNDEBUG")
else()
# It has not been used now, it can specify CUDA compile flag manualy,
# its use is to remvoe /Zi to reduce GPU static library size. But it's dangerous
# because CUDA will update by nvidia, then error will occur.
# Now, it's used in CUDA:[10.0, 10.2]
set(WIN_PROPS ${CMAKE_SOURCE_DIR}/cmake/paddle_win.props)
endif()

File diff suppressed because it is too large Load Diff

Some files were not shown because too many files have changed in this diff Show More

Loading…
Cancel
Save