request prealloc_y sync after quantization

Author: 0cc4m

ggml/src/ggml-vulkan/ggml-vulkan.cpp

@@ -7478,7 +7478,7 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context&
     if (x_non_contig || qx_needs_dequant) {
         ctx->prealloc_x_need_sync = true;
     }
-    if (y_non_contig) {
+    if (y_non_contig || quantize_y) {
         ctx->prealloc_y_need_sync = true;
     }
 }
@@ -7504,7 +7504,7 @@ static void ggml_vk_mul_mat_vec_id_q_f16(ggml_backend_vk_context * ctx, vk_conte
 
     const uint64_t ne10 = src1->ne[0];
     const uint64_t ne11 = src1->ne[1];
-    // const uint64_t ne12 = src1->ne[2];
+    const uint64_t ne12 = src1->ne[2];
     // const uint64_t ne13 = src1->ne[3];
 
     const uint64_t nei0 = ids->ne[0];
@@ -7700,7 +7700,7 @@ static void ggml_vk_mul_mat_vec_id_q_f16(ggml_backend_vk_context * ctx, vk_conte
     if (x_non_contig) {
         ctx->prealloc_x_need_sync = true;
     }
-    if (y_non_contig) {
+    if (y_non_contig || quantize_y) {
         ctx->prealloc_y_need_sync = true;
     }
 }