TimeScope: How Long Can Your Video Large Multimodal Model Go?