upload_from_view()

template<typename T >

void MayaFlux::Buffers::upload_from_view	(	const Kakshya::DataAccess &	accessor,
		const std::shared_ptr< VKBuffer > &	target,
		const std::shared_ptr< VKBuffer > &	staging = nullptr
	)

Upload data from DataAccess view to GPU buffer (precision-preserving)

Template Parameters

T	View type (double, glm::dvec2, glm::dvec3, glm::vec3, float, etc.)

Parameters

accessor	DataAccess instance providing the view
target	Target GPU buffer
staging	Optional staging buffer (auto-created if needed)

Zero-copy when types match, automatic conversion cache when they don't. For AUDIO modalities, defaults to DOUBLE precision to preserve accuracy.

Definition at line 187 of file StagingUtils.hpp.

{
    auto view = accessor.view<T>();
 
    const void* data_ptr = view.data();
    size_t data_bytes = view.size() * sizeof(T);
 
    if constexpr (std::is_same_v<T, double>) {
        if (target->get_format() != vk::Format::eR64Sfloat) {
            MF_WARN(Journal::Component::Buffers, Journal::Context::BufferProcessing,
                "Uploading double precision to buffer with format {}. Consider using R64Sfloat for audio.",
                vk::to_string(target->get_format()));
        }
    }
 
    upload_to_gpu(data_ptr, data_bytes, target, staging);
}

References MayaFlux::Journal::BufferProcessing, MayaFlux::Journal::Buffers, MF_WARN, upload_to_gpu(), and MayaFlux::Kakshya::DataAccess::view().

Here is the call graph for this function: