Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into co_shard_for_reshape

Author: ooooo-create

.github/workflows/Slice-baseline.yml

@@ -1,9 +1,16 @@
-name: Slice-baseline
+name: Slice-baseline-paddle
 
 on:
   workflow_dispatch:
+    inputs:
+      PR_ID:
+        required: false
+        type: string
+      COMMIT_ID:
+        required: false
+        type: string
   schedule:
-    - cron: '0 20 * * *'
+    - cron: '0 20 * * 0'
 
 permissions: read-all
 
@@ -43,3 +50,5 @@ jobs:
       docker_build_image: ${{ needs.build-docker.outputs.docker_build_image }}
       slice-check: 'true'
       SLICE_TEST_MODE: insert_baseline
+      MANUALLY_PR_ID: ${{ inputs.PR_ID }}
+      MANUALLY_COMMIT_ID: ${{ inputs.COMMIT_ID }}

.github/workflows/Slice-torch.yml

@@ -1,4 +1,4 @@
-name: Slice-baseline
+name: Slice-baseline-torch
 
 on:
   workflow_dispatch:

.github/workflows/_Linux-XPU.yml

@@ -206,7 +206,7 @@ jobs:
           CCACHE_DIR: /root/.ccache
           CCACHE_MAXSIZE: 150G
           CCACHE_LIMIT_MULTIPLE: 0.8
-          IF_KUNLUN3: "OFF"
+          IF_KUNLUN3: "ON"
           GITHUB_API_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           home_dir: ${{ github.workspace }}/../../../..
           FLAGS_use_stride_kernel: "0"

.github/workflows/_Slice.yml

@@ -20,6 +20,12 @@ on:
         type: string
         required: false
         default: 'paddle'
+      MANUALLY_PR_ID:
+        type: string
+        required: false
+      MANUALLY_COMMIT_ID:
+        type: string
+        required: false
 
 env:
   PR_ID: ${{ github.event.pull_request.number || '0' }}
@@ -37,9 +43,18 @@ defaults:
     shell: bash
 
 jobs:
+  check-bypass:
+    name: Check bypass
+    uses: ./.github/workflows/check-bypass.yml
+    with:
+      workflow-name: "slice"
+    secrets:
+      github-token: ${{ secrets.GITHUB_TOKEN }}
+
   slice:
     name: Slice test
-    if: ${{ inputs.can-skip != 'true' }}
+    needs: check-bypass
+    if: ${{ inputs.can-skip != 'true' && needs.check-bypass.outputs.can-skip != 'true' }}
     runs-on:
       group: slice
     steps:
@@ -97,7 +112,11 @@ jobs:
           if [[ "${{ inputs.SLICE_BENCHMARK_FRAMEWORKS }}" == "torch" ]];then
             python3.10 -m pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu118
           else
-            python3.10 -m pip install $wheel_link
+            if [[ "${{ inputs.MANUALLY_PR_ID }}" == "" ]]; then
+              python3.10 -m pip install $wheel_link
+            else
+              python3.10 -m pip install https://paddle-github-action.bj.bcebos.com/PR/build/${{ inputs.MANUALLY_PR_ID }}/${{ inputs.MANUALLY_COMMIT_ID }}/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
+            fi
           fi
           python3.10 -m pip install -r PaddleTest/framework/e2e/api_benchmark/requirement.txt
           cd PaddleTest/framework/slice_benchmark

CMakeLists.txt

@@ -322,7 +322,7 @@ option(WITH_ARM "Compile PaddlePaddle with arm support" OFF)
 option(WITH_SW "Compile PaddlePaddle with sw support" OFF)
 option(WITH_MIPS "Compile PaddlePaddle with mips support" OFF)
 option(WITH_LOONGARCH "Compile PaddlePaddle with loongarch support" OFF)
-option(WITH_MUSL "Compile with musl libc instead of gblic" OFF)
+option(WITH_MUSL "Compile with musl libc instead of glibc" OFF)
 option(WITH_UNITY_BUILD "Compile with UnityBuild mode" OFF)
 option(WITH_STRIP "Strip so files of Whl packages" OFF)
 option(NEW_RELEASE_PYPI

ci/check_approval.sh

@@ -309,6 +309,12 @@ if [ "${ALL_PADDLE_ENFORCE}" != "" ] && [ "${PR_ID}" != "" ]; then
     check_approval 1 luotao1 zhangbo9674 phlrain
 fi
 
+CHINESE_CHECK=$(git diff -U0 upstream/$BRANCH |grep "^+" |grep -P '[\p{Han}]')
+if [ "${CHINESE_CHECK}" != "" ] && [ "${PR_ID}" != "" ]; then
+	echo_line="Not recommended to use Chinese. You must have one RD (tianshuo78520a or swgu98 or zhangbo9674 or risemeup1) approval."
+    check_approval 1 tianshuo78520a swgu98 zhangbo9674 risemeup1
+fi
+
 ALL_ADDED_LINES=$(git diff -U0 upstream/$BRANCH |grep "^+" || true)
 ALL_PADDLE_CHECK=$(echo $ALL_ADDED_LINES |grep -zoE "(PADDLE_ENFORCE[A-Z_]{0,9}|PADDLE_THROW)\(.[^,\);]*.[^;]*\);\s" || true)
 VALID_PADDLE_CHECK=$(echo "$ALL_PADDLE_CHECK" | grep -zoE '(PADDLE_ENFORCE[A-Z_]{0,9}|PADDLE_THROW)\(([^,;]+,)*[^";]*errors::.[^"]*".[^";]{20,}.[^;]*\);\s' || true)

ci/kunlun_test.sh

@@ -159,9 +159,10 @@ set +x
             git clone --depth 1000 https://gitee.com/paddlepaddle/PaddleX.git
             cd PaddleX
             pip install -e .
+            pip install numpy==1.24.4 pypdfium2
 
             #install paddle x dependency
-            paddlex --install PaddleClas
+            paddlex --install PaddleClas -y
 
             #download paddle dataset
             wget -q https://paddle-model-ecology.bj.bcebos.com/paddlex/data/cls_flowers_examples.tar -P ./dataset

cmake/cinn.cmake

@@ -26,12 +26,21 @@ else()
 endif()
 
 if(NOT DEFINED ENV{runtime_include_dir})
-  message(
-    STATUS
-      "set runtime_include_dir: ${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
-  set(ENV{runtime_include_dir} "${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
-  add_definitions(
-    -DRUNTIME_INCLUDE_DIR="${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
+  if(WITH_GPU)
+    message(
+      STATUS
+        "set runtime_include_dir: ${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
+    set(ENV{runtime_include_dir} "${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
+    add_definitions(
+      -DRUNTIME_INCLUDE_DIR="${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/cuda")
+  elseif(WITH_ROCM)
+    message(
+      STATUS
+        "set runtime_include_dir: ${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/hip")
+    set(ENV{runtime_include_dir} "${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/hip")
+    add_definitions(
+      -DRUNTIME_INCLUDE_DIR="${CMAKE_SOURCE_DIR}/paddle/cinn/runtime/hip")
+  endif()
 endif()
 
 if(WITH_TESTING)
@@ -118,6 +127,10 @@ if(WITH_ROCM)
     add_definitions(-DCINN_WITH_HIP)
   endif()
   link_libraries(${ROCM_HIPRTC_LIB})
+
+  message(
+    STATUS "copy paddle/cinn/common/float16.h to $ENV{runtime_include_dir}")
+  file(COPY paddle/cinn/common/float16.h DESTINATION $ENV{runtime_include_dir})
 endif()
 
 set(cinnapi_src CACHE INTERNAL "" FORCE)

cmake/external/xpu.cmake

@@ -28,14 +28,15 @@ set(XPU_XPTI_LIB_NAME "libxpti.so")
 set(XPU_XBLAS_LIB_NAME "libxpu_blas.so")
 set(XPU_XFA_LIB_NAME "libxpu_flash_attention.so")
 set(XPU_XPUDNN_LIB_NAME "libxpu_dnn.so")
+set(XPU_XPUDNN_OMP_LIB_NAME "libomp.so")
 set(XPU_FFT_LIB_NAME "libcufft.so")
 # Avoid deprecated int32 apis:
 add_compile_definitions(XPUAPI_NOT_INCLUDE_DEPRECATED)
 
 if(NOT DEFINED XPU_XHPC_BASE_DATE)
-  set(XPU_XHPC_BASE_DATE "dev/20990602")
+  set(XPU_XHPC_BASE_DATE "dev/20250722")
 endif()
-set(XPU_XCCL_BASE_VERSION "3.0.2.7") # For XRE5
+set(XPU_XCCL_BASE_VERSION "3.0.3.1") # For XRE5
 if(NOT DEFINED XPU_XFT_BASE_VERSION)
   set(XPU_XFT_BASE_VERSION "20250507/xpu3")
 endif()
@@ -183,6 +184,7 @@ set(XPU_CUDA_RT_LIB "${XPU_LIB_DIR}/${XPU_CUDA_RT_LIB_NAME}")
 set(XPU_ML_LIB "${XPU_LIB_DIR}/${XPU_ML_LIB_NAME}")
 set(XPU_XFA_LIB "${XPU_LIB_DIR}/${XPU_XFA_LIB_NAME}")
 set(XPU_XPUDNN_LIB "${XPU_LIB_DIR}/${XPU_XPUDNN_LIB_NAME}")
+set(XPU_XPUDNN_OMP_LIB "${XPU_LIB_DIR}/${XPU_XPUDNN_OMP_LIB_NAME}")
 
 set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${XPU_INSTALL_DIR}/lib")
 
@@ -251,8 +253,9 @@ if(WITH_XPU_XRE5)
     DOWNLOAD_COMMAND
       bash ${CMAKE_SOURCE_DIR}/tools/xpu/pack_paddle_dependence.sh
       ${XPU_XRE_URL} ${XPU_XRE_DIR_NAME} ${XPU_XHPC_URL} ${XPU_XHPC_DIR_NAME}
-      ${XPU_XCCL_URL} ${XPU_XCCL_DIR_NAME} 1 && wget ${XPU_XFT_GET_DEPENCE_URL}
-      && bash ${XFT_COMMAND} ${XPU_XFT_URL} ${XPU_XFT_DIR_NAME} && bash
+      ${XPU_XCCL_URL} ${XPU_XCCL_DIR_NAME} 1 ${WITH_MKL} "${CMAKE_BINARY_DIR}"
+      && wget ${XPU_XFT_GET_DEPENCE_URL} && bash ${XFT_COMMAND} ${XPU_XFT_URL}
+      ${XPU_XFT_DIR_NAME} && bash
       ${CMAKE_SOURCE_DIR}/tools/xpu/get_xpti_dependence.sh ${XPU_XPTI_URL}
       ${XPU_XPTI_DIR_NAME} && bash
       ${CMAKE_SOURCE_DIR}/tools/xpu/get_xpufft_dependence.sh ${XPU_FFT_URL}
@@ -263,6 +266,7 @@ if(WITH_XPU_XRE5)
     BUILD_BYPRODUCTS ${XPU_API_LIB}
     BUILD_BYPRODUCTS ${XPU_XBLAS_LIB}
     BUILD_BYPRODUCTS ${XPU_XPUDNN_LIB}
+    BUILD_BYPRODUCTS ${XPU_XPUDNN_OMP_LIB}
     BUILD_BYPRODUCTS ${XPU_XFA_LIB}
     BUILD_BYPRODUCTS ${XPU_RT_LIB}
     BUILD_BYPRODUCTS ${XPU_CUDA_RT_LIB}
@@ -360,6 +364,10 @@ if(WITH_XPU_XRE5)
     ${XPU_XFA_LIB}
     ${XPU_XPUDNN_LIB}
     ${XPU_ML_LIB})
+
+  if(NOT WITH_MKL)
+    target_link_libraries(xpulib ${XPU_XPUDNN_OMP_LIB})
+  endif()
 else()
   target_link_libraries(xpulib ${XPU_RT_LIB} ${XPU_API_LIB})
 endif()

doc/README_cn.md

@@ -0,0 +1,6 @@
+# 致读者和开发者
+感谢您阅读 PaddlePaddle 文档。
+
+自 **2018年9月17日** 起，**0.15.0 及 develop** 分支的文档源码已迁移至 [FluidDoc Repo](https://github.com/PaddlePaddle/FluidDoc) 仓库 ，并将在该仓库中持续更新。
+
+请前往 FluidDoc 仓库获取最新文档。