Add support for VK_EXT_debug_utils to add labels to Vulkan objects. (ggml-org#13792)

* Add support for VK_EXT_debug_utils to add labels to Vulkan objects. In step 1 compute pipelines are getting labeled.

* remove #ifdef for debug utils and add queue marker.

Author: mtavenrath

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

@@ -3675,6 +3675,9 @@ static void ggml_vk_instance_init() {
 
     }
 
+    size_t num_available_devices = vk_instance.instance.enumeratePhysicalDevices().size();
+    vk_perf_logger_enabled = getenv("GGML_VK_PERF_LOGGER") != nullptr;
+
     // Emulate behavior of CUDA_VISIBLE_DEVICES for Vulkan
     char * devices_env = getenv("GGML_VK_VISIBLE_DEVICES");
     if (devices_env != nullptr) {
@@ -10154,15 +10157,9 @@ static bool ggml_backend_vk_device_supports_op(ggml_backend_dev_t dev, const ggm
                 ggml_type src0_type = op->src[0]->type;
                 ggml_backend_vk_device_context * ctx = (ggml_backend_vk_device_context *)dev->context;
                 const vk_device& device = ggml_vk_get_device(ctx->device);
-                if (op->op == GGML_OP_MUL_MAT_ID) {
-                    if (!device->mul_mat_id_s[src0_type] && !device->mul_mat_id_m[src0_type] && !device->mul_mat_id_l[src0_type]) {
-                        // If there's not enough shared memory for row_ids and the result tile, fallback to CPU
-                        return false;
-                    }
-                    // Check against size of shared memory variable
-                    if (op->src[2]->ne[0] > 4096) {
-                        return false;
-                    }
+                if (op->op == GGML_OP_MUL_MAT_ID && !device->mul_mat_id_s[src0_type] && !device->mul_mat_id_m[src0_type] && !device->mul_mat_id_l[src0_type]) {
+                    // If there's not enough shared memory for row_ids and the result tile, fallback to CPU
+                    return false;
                 }
                 switch (src0_type) {
                     case GGML_TYPE_F32: