[CUDA] Added a cudaMemcpy2DAsync fast path to ggml_cuda_cpy (#25057)

* [CUDA] Added a cudaMemcpy2DAsync fast path to ggml_cuda_cpy

Add a CUDA ggml_cpy fast path for same-type, same-shape strided copies that are just 2D pitched block copies.
When tensors are not fully contiguous but each row is contiguous, it now uses cudaMemcpy2DAsync instead of the slow element-wise scalar copy kernel.

This fixes the GDN recurrent snapshot update with -np 4, where rollback slots are separated by cache stride gaps.

* Add new tests that execute the new optimized strided copy path

* Return unsupported for strided copy in OpenVINO, as new tests are failing

ggml/src/ggml-openvino/ggml-openvino.cpp

@@ -1053,6 +1053,10 @@ static bool is_op_unsupported_case(const ggml_tensor * op) {
             (op->ne[0] == 2 && op->ne[1] == 4 && op->ne[2] == 3 && op->ne[3] == 2)) {
             return true;
         }
+        // CPY into a strided view of a larger buffer (recurrent-state snapshots) not supported
+        if (op->view_src && ggml_nbytes(op) != ggml_nbytes(op->view_src)) {
+            return true;
+        }
         break;
     }
     case GGML_OP_MUL_MAT: {