Add support inference on SYCL devices (#9800)

---------

Co-authored-by: Dmitry Razdoburdin <>
Co-authored-by: Nikolay Petrov <nikolay.a.petrov@intel.com>
Co-authored-by: Alexandra <alexandra.epanchinzeva@intel.com>

tests/ci_build/conda_env/linux_sycl_test.yml

@@ -0,0 +1,20 @@
+name: linux_sycl_test
+channels:
+- conda-forge
+- intel
+dependencies:
+- python=3.8
+- cmake
+- c-compiler
+- cxx-compiler
+- pip
+- wheel
+- numpy
+- scipy
+- scikit-learn
+- pandas
+- hypothesis>=6.46
+- pytest
+- pytest-timeout
+- pytest-cov
+- dpcpp_linux-64

tests/ci_build/lint_cpp.py

@@ -138,7 +138,7 @@ def main():
         "path",
         nargs="*",
         help="Path to traverse",
-        default=["src", "include", os.path.join("R-package", "src"), "python-package"],
+        default=["src", "include", os.path.join("R-package", "src"), "python-package", "plugin/sycl"],
     )
     parser.add_argument(
         "--exclude_path",

tests/ci_build/lint_python.py

@@ -33,6 +33,7 @@ class LintersPaths:
         "tests/python-gpu/test_gpu_pickling.py",
         "tests/python-gpu/test_gpu_eval_metrics.py",
         "tests/python-gpu/test_gpu_with_sklearn.py",
+        "tests/python-sycl/test_sycl_prediction.py",
         "tests/test_distributed/test_with_spark/",
         "tests/test_distributed/test_gpu_with_spark/",
         # demo