[CustomDevice]add custom place supports (#43813)

* [CustomDevice]add custom place supports

* sync format

paddle/fluid/operators/array_operator.h

@@ -52,7 +52,8 @@ class ArrayOp : public framework::OperatorBase {
     size_t offset;
     if (platform::is_gpu_place(i_tensor.place()) ||
         platform::is_xpu_place(i_tensor.place()) ||
-        platform::is_npu_place(i_tensor.place())) {
+        platform::is_npu_place(i_tensor.place()) ||
+        platform::is_custom_place(i_tensor.place())) {
       // FIXME: Avoid copy from GPU to CPU
       framework::Tensor t;
       framework::TensorCopy(i_tensor, platform::CPUPlace(), dev_ctx, &t);

paddle/fluid/operators/reduce_ops/reduce_op.h

@@ -566,7 +566,8 @@ class ReduceOp : public framework::OperatorWithKernel {
       PADDLE_ENFORCE_EQ(
           platform::is_gpu_place(ctx.GetPlace()) ||
               platform::is_npu_place(ctx.GetPlace()) ||
-              platform::is_mlu_place(ctx.GetPlace()),
+              platform::is_mlu_place(ctx.GetPlace()) ||
+              platform::is_custom_place(ctx.GetPlace()),
           true,
           platform::errors::InvalidArgument(
               "float16 can only be used on GPU or NPU or MLU place"));

paddle/fluid/pybind/pybind.cc

@@ -2801,6 +2801,7 @@ All parameter, weight, gradient are variables in Paddle.
       .def("_equals", &IsSamePlace<platform::Place, platform::IPUPlace>)
       .def("_equals", &IsSamePlace<platform::Place, platform::CUDAPinnedPlace>)
       .def("_equals", &IsSamePlace<platform::Place, platform::MLUPlace>)
+      .def("_equals", &IsSamePlace<platform::Place, platform::CustomPlace>)
       .def("is_gpu_place",
            [](platform::Place &self) { return platform::is_gpu_place(self); })
       .def("is_cpu_place",

python/paddle/device/__init__.py

@@ -349,6 +349,10 @@ def get_device():
     elif isinstance(place, core.MLUPlace):
         device_id = place.get_device_id()
         device = 'mlu:' + str(device_id)
+    elif isinstance(place, core.CustomPlace):
+        device_id = place.get_device_id()
+        device_type = place.get_device_type()
+        device = device_type + ':' + str(device_id)
     else:
         raise ValueError("The device specification {} is invalid".format(place))
 

python/paddle/fluid/tests/unittests/op_test.py

@@ -1685,7 +1685,8 @@ class OpTest(unittest.TestCase):
         # Currently not support ParallelExecutor on XPUPlace.
         if not paddle.is_compiled_with_xpu(
         ) and not paddle.is_compiled_with_npu(
-        ) and not paddle.is_compiled_with_mlu():
+        ) and not paddle.is_compiled_with_mlu() and not isinstance(
+                place, core.CustomPlace):
             self.check_inplace_output_with_place(place,
                                                  no_check_set=no_check_set,
                                                  inplace_atol=inplace_atol)