[BE] Upgrade XPU support package to 2025.2 (pytorch#158733)

Including below changes,

- Add XPU support package 2025.2 build and test in CI for both Linux and Windows
- Keep XPU support package 2025.1 build in CI to ensure no break issue until PyTorch 2.9 release
- Upgrade XPU support package from 2025.1 to 2025.2 in CD for both Linux and Windows
- Rename Linux CI job name & image name to n & n-1
- Update XPU runtime pypi packages dependencies of CD wheels
- Remove deprecated support package version docker image build

Pull Request resolved: pytorch#158733
Approved by: https://github.com/EikanWang, https://github.com/atalman

Author: chuanqi129

.ci/docker/build.sh

@@ -209,19 +209,19 @@ case "$tag" in
     UCC_COMMIT=${_UCC_COMMIT}
     PYTORCH_ROCM_ARCH="gfx90a;gfx942;gfx950"
     ;;
-  pytorch-linux-jammy-xpu-2025.0-py3)
+  pytorch-linux-jammy-xpu-n-1-py3)
     ANACONDA_PYTHON_VERSION=3.9
     GCC_VERSION=11
     VISION=yes
-    XPU_VERSION=2025.0
+    XPU_VERSION=2025.1
     NINJA_VERSION=1.9.0
     TRITON=yes
     ;;
-  pytorch-linux-jammy-xpu-2025.1-py3)
+  pytorch-linux-jammy-xpu-n-py3)
     ANACONDA_PYTHON_VERSION=3.9
     GCC_VERSION=11
     VISION=yes
-    XPU_VERSION=2025.1
+    XPU_VERSION=2025.2
     NINJA_VERSION=1.9.0
     TRITON=yes
     ;;

.ci/docker/common/install_xpu.sh

@@ -146,11 +146,11 @@ if [[ "${XPU_DRIVER_TYPE,,}" == "lts" ]]; then
     XPU_DRIVER_VERSION="/lts/2350"
 fi
 
-# Default use Intel® oneAPI Deep Learning Essentials 2025.0
-if [[ "$XPU_VERSION" == "2025.1" ]]; then
-    XPU_PACKAGES="intel-deep-learning-essentials-2025.1"
+# Default use Intel® oneAPI Deep Learning Essentials 2025.1
+if [[ "$XPU_VERSION" == "2025.2" ]]; then
+    XPU_PACKAGES="intel-deep-learning-essentials-2025.2"
 else
-    XPU_PACKAGES="intel-deep-learning-essentials-2025.0"
+    XPU_PACKAGES="intel-deep-learning-essentials-2025.1"
 fi
 
 # The installation depends on the base OS

.ci/docker/manywheel/Dockerfile_2_28

@@ -175,6 +175,6 @@ ENV XPU_DRIVER_TYPE ROLLING
 RUN python3 -m pip install --upgrade pip && \
     python3 -mpip install cmake==3.28.4
 ADD ./common/install_xpu.sh install_xpu.sh
-ENV XPU_VERSION 2025.1
+ENV XPU_VERSION 2025.2
 RUN bash ./install_xpu.sh && rm install_xpu.sh
 RUN pushd /opt/_internal && tar -xJf static-libs-for-embedding-only.tar.xz && popd

.ci/pytorch/windows/internal/xpu_install.bat

@@ -13,9 +13,9 @@ if not exist "%SRC_DIR%\temp_build" mkdir "%SRC_DIR%\temp_build"
 :xpu_bundle_install_start
 
 set XPU_BUNDLE_PARENT_DIR=C:\Program Files (x86)\Intel\oneAPI
-set XPU_BUNDLE_URL=https://registrationcenter-download.intel.com/akdlm/IRC_NAS/9d6d6c17-ca2d-4735-9331-99447e4a1280/intel-deep-learning-essentials-2025.0.1.28_offline.exe
+set XPU_BUNDLE_URL=https://registrationcenter-download.intel.com/akdlm/IRC_NAS/75d4eb97-914a-4a95-852c-7b9733d80f74/intel-deep-learning-essentials-2025.1.3.8_offline.exe
 set XPU_BUNDLE_PRODUCT_NAME=intel.oneapi.win.deep-learning-essentials.product
-set XPU_BUNDLE_VERSION=2025.0.1+20
+set XPU_BUNDLE_VERSION=2025.1.3+5
 set XPU_BUNDLE_INSTALLED=0
 set XPU_BUNDLE_UNINSTALL=0
 set XPU_EXTRA_URL=NULL
@@ -24,9 +24,9 @@ set XPU_EXTRA_VERSION=2025.0.1+1226
 set XPU_EXTRA_INSTALLED=0
 set XPU_EXTRA_UNINSTALL=0
 
-if not [%XPU_VERSION%]==[] if [%XPU_VERSION%]==[2025.1] (
-    set XPU_BUNDLE_URL=https://registrationcenter-download.intel.com/akdlm/IRC_NAS/75d4eb97-914a-4a95-852c-7b9733d80f74/intel-deep-learning-essentials-2025.1.3.8_offline.exe
-    set XPU_BUNDLE_VERSION=2025.1.3+5
+if not [%XPU_VERSION%]==[] if [%XPU_VERSION%]==[2025.2] (
+    set XPU_BUNDLE_URL=https://registrationcenter-download.intel.com/akdlm/IRC_NAS/24751ead-ddc5-4479-b9e6-f9fe2ff8b9f2/intel-deep-learning-essentials-2025.2.1.25_offline.exe
+    set XPU_BUNDLE_VERSION=2025.2.1+20
 )
 
 :: Check if XPU bundle is target version or already installed
@@ -90,14 +90,3 @@ if errorlevel 1 exit /b 1
 del xpu_extra.exe
 
 :xpu_install_end
-
-if not "%XPU_ENABLE_KINETO%"=="1" goto install_end
-:: Install Level Zero SDK
-set XPU_EXTRA_LZ_URL=https://github.com/oneapi-src/level-zero/releases/download/v1.14.0/level-zero-sdk_1.14.0.zip
-curl -k -L %XPU_EXTRA_LZ_URL% --output "%SRC_DIR%\temp_build\level_zero_sdk.zip"
-echo "Installing level zero SDK..."
-7z x "%SRC_DIR%\temp_build\level_zero_sdk.zip" -o"%SRC_DIR%\temp_build\level_zero"
-set "INCLUDE=%SRC_DIR%\temp_build\level_zero\include;%INCLUDE%"
-del "%SRC_DIR%\temp_build\level_zero_sdk.zip"
-
-:install_end

.circleci/scripts/binary_windows_build.sh

@@ -15,8 +15,7 @@ fi
 if [[ "$DESIRED_CUDA" == 'xpu' ]]; then
     export VC_YEAR=2022
     export USE_SCCACHE=0
-    export XPU_VERSION=2025.1
-    export XPU_ENABLE_KINETO=1
+    export XPU_VERSION=2025.2
 fi
 
 echo "Free space on filesystem before build:"

.circleci/scripts/binary_windows_test.sh

@@ -8,7 +8,7 @@ export VC_YEAR=2022
 
 if [[ "$DESIRED_CUDA" == 'xpu' ]]; then
     export VC_YEAR=2022
-    export XPU_VERSION=2025.1
+    export XPU_VERSION=2025.2
 fi
 
 pushd "$PYTORCH_ROOT/.ci/pytorch/"

.github/scripts/generate_binary_build_matrix.py

@@ -113,26 +113,26 @@
         "nvidia-cufile==1.15.0.42; platform_system == 'Linux' and platform_machine == 'x86_64'"
     ),
     "xpu": (
-        "intel-cmplr-lib-rt==2025.1.1 | "
-        "intel-cmplr-lib-ur==2025.1.1 | "
-        "intel-cmplr-lic-rt==2025.1.1 | "
-        "intel-sycl-rt==2025.1.1 | "
-        "oneccl-devel==2021.15.2; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "oneccl==2021.15.2; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "impi-rt==2021.15.0; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "onemkl-sycl-blas==2025.1.0 | "
-        "onemkl-sycl-dft==2025.1.0 | "
-        "onemkl-sycl-lapack==2025.1.0 | "
-        "onemkl-sycl-rng==2025.1.0 | "
-        "onemkl-sycl-sparse==2025.1.0 | "
-        "dpcpp-cpp-rt==2025.1.1 | "
-        "intel-opencl-rt==2025.1.1 | "
-        "mkl==2025.1.0 | "
-        "intel-openmp==2025.1.1 | "
-        "tbb==2022.1.0 | "
-        "tcmlib==1.3.0 | "
-        "umf==0.10.0 | "
-        "intel-pti==0.12.3"
+        "intel-cmplr-lib-rt==2025.2.1 | "
+        "intel-cmplr-lib-ur==2025.2.1 | "
+        "intel-cmplr-lic-rt==2025.2.1 | "
+        "intel-sycl-rt==2025.2.1 | "
+        "oneccl-devel==2021.16.1; platform_system == 'Linux' and platform_machine == 'x86_64' | "
+        "oneccl==2021.16.1; platform_system == 'Linux' and platform_machine == 'x86_64' | "
+        "impi-rt==2021.16.1; platform_system == 'Linux' and platform_machine == 'x86_64' | "
+        "onemkl-sycl-blas==2025.2.0 | "
+        "onemkl-sycl-dft==2025.2.0 | "
+        "onemkl-sycl-lapack==2025.2.0 | "
+        "onemkl-sycl-rng==2025.2.0 | "
+        "onemkl-sycl-sparse==2025.2.0 | "
+        "dpcpp-cpp-rt==2025.2.1 | "
+        "intel-opencl-rt==2025.2.1 | "
+        "mkl==2025.2.0 | "
+        "intel-openmp==2025.2.1 | "
+        "tbb==2022.2.0 | "
+        "tcmlib==1.4.0 | "
+        "umf==0.11.0 | "
+        "intel-pti==0.13.1"
     ),
 }
 

.github/templates/common.yml.j2

@@ -4,7 +4,7 @@
 {%- set download_artifact_action = "actions/[email protected]" -%}
 
 {%- set timeout_minutes = 240 -%}
-{%- set timeout_minutes_windows_binary = 300 -%}
+{%- set timeout_minutes_windows_binary = 360 -%}
 
 {%- macro concurrency(build_environment) -%}
 concurrency:

.github/workflows/docker-builds.yml

@@ -66,8 +66,8 @@ jobs:
           pytorch-linux-jammy-py3.10-gcc11,
           pytorch-linux-jammy-py3.9-gcc11-inductor-benchmarks,
           pytorch-linux-jammy-py3.12-halide,
-          pytorch-linux-jammy-xpu-2025.0-py3,
-          pytorch-linux-jammy-xpu-2025.1-py3,
+          pytorch-linux-jammy-xpu-n-1-py3,
+          pytorch-linux-jammy-xpu-n-py3,
           pytorch-linux-jammy-py3-clang18-asan,
           pytorch-linux-jammy-py3-clang12-onnx,
           pytorch-linux-jammy-linter,