Implement Qwen text core

native/metal/ds4_gpu.h

@@ -128,6 +128,30 @@ int ds4_gpu_set_aux_model_map_range(const void *model_map,
                                     uint64_t map_offset,
                                     uint64_t map_size);
 int ds4_gpu_set_model_map_spans(const void *model_map, uint64_t model_size, const uint64_t *offsets, const uint64_t *sizes, uint32_t count, uint64_t max_tensor_bytes);
+
+typedef struct {
+    const void *map;
+    uint64_t size;
+    uint64_t offset;
+    uint64_t bytes;
+} ds4_gpu_qwen_weight_view;
+
+typedef struct {
+    uint32_t u[16];
+    float f[8];
+} ds4_gpu_qwen_kernel_args;
+
+int ds4_gpu_qwen_dispatch(
+        const char *kernel,
+        ds4_gpu_tensor *out,
+        const ds4_gpu_tensor *a,
+        const ds4_gpu_tensor *b,
+        const ds4_gpu_tensor *c,
+        const ds4_gpu_qwen_weight_view *weights,
+        uint32_t weight_count,
+        const ds4_gpu_qwen_kernel_args *args,
+        uint32_t grid_x,
+        uint32_t grid_y);
 int ds4_gpu_cache_model_range(const void *model_map, uint64_t model_size, uint64_t offset, uint64_t bytes, const char *label);
 int ds4_gpu_cache_q8_f16_range(const void *model_map, uint64_t model_size, uint64_t offset, uint64_t bytes, uint64_t in_dim, uint64_t out_dim, const char *label);
 int ds4_gpu_q8_cache_suppressed(void);