use inplace multiply in grad_clip

Author: Waynezee

python/paddle/nn/clip.py

@@ -824,8 +824,8 @@ def async_add_n(var_list):
                         clip_input = paddle.distributed.reshard(
                             clip_input, g.process_mesh, clip_input.placements
                         )
-                new_grad = paddle.multiply(g, clip_input)
-                params_and_grads.append((p, new_grad))
+                g.multiply_(clip_input)
+                params_and_grads.append((p, g))
             else:
                 params_and_grads.append((p, g))
 
@@ -1089,8 +1089,8 @@ def async_add_n(var_list):
                         ),
                     )
 
-                new_grad = paddle.multiply(g, clip_input)
-                params_and_grads.append((p, new_grad))
+                g.multiply_(clip_input)
+                params_and_grads.append((p, g))
             else:
                 params_and_grads.append((p, g))