Enable special strided CPY for conv state writeback

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

@@ -118,16 +118,6 @@ bool is_conv_states_all_tensor(const ggml_tensor * tensor) {
     return tensor != nullptr && strncmp(tensor->name, "conv_states_all", strlen("conv_states_all")) == 0;
 }
 
-// CPY writing the tail of conv_input (the concat of the previous conv state and the new tokens)
-// back into a slot block of the recurrent state cache. Detected structurally because the rollback
-// variant (cparams.n_rs_seq > 0) emits one such CPY per snapshot slot without naming them.
-bool is_conv_state_writeback(const ggml_tensor * node) {
-    return node->op == GGML_OP_CPY && node->view_src != nullptr && GgmlOvDecoder::is_kvcache(node->view_src, nullptr) &&
-           node->src[0] != nullptr && node->src[0]->op == GGML_OP_VIEW && node->src[0]->src[0] != nullptr &&
-           node->src[0]->src[0]->op == GGML_OP_CONCAT && node->src[1] != nullptr && node->src[1]->op == GGML_OP_VIEW &&
-           node->src[1]->view_src == node->view_src;
-}
-
 // MoE expert aggregation (build_moe_ffn in llama-graph.cpp): each expert plane is
 // `ggml_view_2d(experts, n_embd, n_tokens, experts->nb[2], i*experts->nb[1])` and the planes
 // are summed with a chain of ADDs: moe_out = ((view_0 + view_1) + view_2) + ... + view_{n-1}.
@@ -453,7 +443,7 @@ int GgmlOvDecoder::compute_op_case(const ggml_tensor * node) const {
         if (node->src[0]->op == GGML_OP_VIEW) {
             if (node->src[0]->src[0]->op == GGML_OP_GATED_DELTA_NET) {
                 op_case = 1;
-            } else if (is_conv_state_writeback(node)) {
+            } else if (GgmlOvDecoder::is_conv_state_writeback(node)) {
                 op_case = 2;
                 break;
             } else if (is_conv_states_all_tensor(node->view_src) && node->src[1] != nullptr &&

ggml/src/ggml-openvino/ggml-decoder.h

@@ -385,6 +385,13 @@ class GgmlOvDecoder : public ov::frontend::ggml::GgmlDecoder {
                (op != nullptr && op->op == GGML_OP_SET_ROWS && op->src[2] == tensor);
     }
 
+    inline static bool is_conv_state_writeback(const ggml_tensor * node) {
+        return node->op == GGML_OP_CPY && node->view_src != nullptr && is_kvcache(node->view_src, nullptr) &&
+               node->src[0] != nullptr && node->src[0]->op == GGML_OP_VIEW && node->src[0]->src[0] != nullptr &&
+               node->src[0]->src[0]->op == GGML_OP_CONCAT && node->src[1] != nullptr &&
+               node->src[1]->op == GGML_OP_VIEW && node->src[1]->view_src == node->view_src;
+    }
+
     inline static bool is_kv_idx(const ggml_tensor * tensor, const ggml_tensor * op) {
         return op->op == GGML_OP_SET_ROWS && op->src[1] == tensor;
     }

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

@@ -1059,7 +1059,7 @@ static bool cpy_output_view_is_supported(const ggml_tensor * op) {
         return false;
     }
 
-    return ggml_nbytes(op) == 0 || ggml_is_contiguous(op);
+    return ggml_nbytes(op) == 0 || ggml_is_contiguous(op) || GgmlOvDecoder::is_conv_state_writeback(op);
 }
 
 static bool mul_mat_id_requires_large_tmp(const ggml_tensor * op) {