Receiving

Heaps

Unlike the Python bindings, the C++ bindings expose three heap types: live heaps (spead2::recv::live_heap) are used for heaps being constructed, and may be missing data; frozen heaps (spead2::recv::heap) always have all their data; and incomplete heaps (spead2::recv::incomplete_heap) are frozen heaps that are missing data. Frozen heaps can be move-constructed from live heaps, which will typically be done in the callback.

class live_heap

A SPEAD heap that is in the process of being received.

Once it is fully received, it is converted to a heap for further processing.

Any SPEAD-64-* flavour can be used, but all packets in the heap must use the same flavour. It may be possible to relax this, but it hasn’t been examined, and may cause issues for decoding descriptors (whose format depends on the flavour).

A heap can be:

complete: a heap length item was found in a packet, and we have received all the payload corresponding to it. No more packets are expected.
contiguous: the payload we have received is a contiguous range from 0 up to some amount, and cover all items described in the item pointers.
incomplete: not contiguous A complete heap is also contiguous, but not necessarily the other way around. Only contiguous heaps can be frozen to heap, and only incomplete heaps can be frozen to incomplete_heap.

Public Functions

bool is_complete() const: True if the heap is complete.

bool is_contiguous() const: True if the heap is contiguous.

bool is_end_of_stream() const: True if an end-of-stream heap control item was found.

inline s_item_pointer_t get_cnt() const: Retrieve the heap ID.

inline bug_compat_mask get_bug_compat() const: Get protocol bug compatibility flags.

class heap : public spead2::recv::heap_base

Received heap that has been finalised.

Public Functions

explicit heap(live_heap &&h)

Freeze a heap, which must satisfy live_heap::is_contiguous.

The original heap is destroyed.

descriptor to_descriptor() const

Extract descriptor fields from the heap.

Any missing fields are default-initialized. This should be used on a heap constructed from the content of a descriptor item.

The original PySPEAD package (version 0.5.2) does not follow the specification here. The macros in common_defines.h can be used to control whether to interpret the specification or be bug-compatible.

The protocol allows descriptors to use immediate-mode items, but the decoding of these into variable-length strings is undefined. This implementation will discard such descriptor fields.

std::vector<descriptor> get_descriptors() const: Extract and decode descriptors from this heap.

inline s_item_pointer_t get_cnt() const: Get heap ID.

inline const flavour &get_flavour() const: Get protocol flavour used.

inline const std::vector<item> &get_items() const

Get the items from the heap.

This includes descriptors, but excludes any items with ID <= 4.

bool is_ctrl_item(ctrl_mode value) const: Convenience function to check whether any of the items is a STREAM_CTRL_ID item with value value.

bool is_start_of_stream() const: Convenience function to check whether any of the items is a CTRL_STREAM_START.

bool is_end_of_stream() const: Convenience function to check whether any of the items is a CTRL_STREAM_STOP.

inline const memory_allocator::pointer &get_payload() const

Get the payload pointer.

This will return an empty pointer unless keep_payload was set in the constructor. This is not normally needed, but has applications in some advanced use cases.

class incomplete_heap : public spead2::recv::heap_base

Received heap that has been finalised, but which is missing data.

The payload and any items that refer to the payload are discarded.

Public Functions

incomplete_heap(live_heap &&h, bool keep_payload, bool keep_payload_ranges)

Freeze a heap.

The original heap is destroyed.

Parameters:

h – The heap to freeze.
keep_payload – If true, transfer the payload memory allocation from the live heap to this object. If false, discard it.
keep_payload_ranges – If true, store information that allows get_payload_ranges to work.

inline s_item_pointer_t get_heap_length() const: Heap payload length encoded in packets (-1 for unknown)

inline s_item_pointer_t get_received_length() const: Number of bytes of payload received.

std::vector<std::pair<s_item_pointer_t, s_item_pointer_t>> get_payload_ranges() const

Return a list of contiguous ranges of payload that were received.

This is intended for special cases where a custom memory allocator was used to channel the payload into a caller-managed area, so that the caller knows which parts of that area have been filled in.

If keep_payload_ranges was false in the constructor, returns an empty list.

inline s_item_pointer_t get_cnt() const: Get heap ID.

inline const flavour &get_flavour() const: Get protocol flavour used.

inline const std::vector<item> &get_items() const

Get the items from the heap.

This includes descriptors, but excludes any items with ID <= 4.

bool is_ctrl_item(ctrl_mode value) const: Convenience function to check whether any of the items is a STREAM_CTRL_ID item with value value.

bool is_start_of_stream() const: Convenience function to check whether any of the items is a CTRL_STREAM_START.

bool is_end_of_stream() const: Convenience function to check whether any of the items is a CTRL_STREAM_STOP.

inline const memory_allocator::pointer &get_payload() const

Get the payload pointer.

This will return an empty pointer unless keep_payload was set in the constructor. This is not normally needed, but has applications in some advanced use cases.

struct item

An item extracted from a heap.

Subclassed by spead2::recv::item_wrapper

Public Members

s_item_pointer_t id: Item ID.

std::uint8_t *ptr: Start of memory containing value.

std::size_t length: Length of memory.

item_pointer_t immediate_value: The immediate interpreted as an integer (undefined if not immediate)

bool is_immediate: Whether the item is immediate.

struct descriptor

An unpacked descriptor.

If numpy_header is non-empty, it overrides format and shape.

Public Members

s_item_pointer_t id = 0: SPEAD ID.

std::string name: Short name.

std::string description: Long description.

std::vector<std::pair<char, s_item_pointer_t>> format

Legacy format.

Each element is a specifier character (e.g. ‘u’ for unsigned) and a bit width.

std::vector<s_item_pointer_t> shape

Shape.

Elements are either non-negative, or -1 is used to indicate a variable-length size. At most one dimension may be variable-length.

std::string numpy_header: Description in the format used in .npy files.

Streams

At the lowest level, heaps are given to the application via a callback to a virtual function. While this callback is running, no new packets can be received from the network socket, so this function needs to complete quickly to avoid data loss when using UDP. To use this interface, subclass spead2::recv::stream and implement heap_ready() and optionally override stop_received().

Note that some public functions are incorrectly listed as protected below due to limitations of the documentation tools.

class stream_config

Parameters for a receive stream.

Public Functions

stream_config &set_max_heaps(std::size_t max_heaps)

Set maximum number of partial heaps that can be live at one time (per substream).

This affects how intermingled heaps can be (due to out-of-order packet delivery) before heaps get dropped.

inline std::size_t get_max_heaps() const: Get maximum number of partial heaps that can be live at one time.

stream_config &set_substreams(std::size_t substreams)

Set number of substreams.

The substream is determined by taking the heap cnt modulo the number of substreams. The value set by set_max_heaps applies independently for each substream.

inline std::size_t get_substreams() const: Get number of substreams.

stream_config &set_memory_allocator(std::shared_ptr<memory_allocator> allocator): Set an allocator to use for allocating heap memory.

inline const std::shared_ptr<memory_allocator> &get_memory_allocator() const: Get allocator for allocating heap memory.

stream_config &set_memcpy(packet_memcpy_function memcpy): Set an alternative memcpy function for copying heap payload.

stream_config &set_memcpy(memcpy_function memcpy): Set an alternative memcpy function for copying heap payload.

stream_config &set_memcpy(memcpy_function_id id): Set builtin memcpy function to use for copying heap payload.

inline const packet_memcpy_function &get_memcpy() const: Get memcpy function for copying heap payload.

stream_config &set_stop_on_stop_item(bool stop): Set whether to stop the stream when a stop item is received.

inline bool get_stop_on_stop_item() const: Get whether to stop the stream when a stop item is received.

stream_config &set_allow_unsized_heaps(bool allow): Set whether to allow heaps without HEAP_LENGTH.

inline bool get_allow_unsized_heaps() const: Get whether to allow heaps without HEAP_LENGTH.

stream_config &set_allow_out_of_order(bool allow): Set whether to allow out-of-order packets within a heap.

inline bool get_allow_out_of_order() const: Get whether to allow out-of-order packets within a heap.

stream_config &set_bug_compat(bug_compat_mask bug_compat): Set bug compatibility flags.

inline bug_compat_mask get_bug_compat() const: Get bug compatibility flags.

stream_config &set_stream_id(std::uintptr_t stream_id): Set a stream ID.

inline std::uintptr_t get_stream_id() const: Get the stream ID.

stream_config &set_explicit_start(bool explicit_start)

Set the explicit start control.

If explicit start is enabled, no data will be received on the stream until stream::start is called, and no readers can be added afterwards.

inline bool get_explicit_start() const: Get the explicit start flag.

std::size_t add_stat(std::string name, stream_stat_config::mode mode = stream_stat_config::mode::COUNTER)

Add a new custom statistic.

Returns the index to use with stream_stats.

Throws:: std::invalid_argument – if name already exists.

inline const std::vector<stream_stat_config> &get_stats() const: Get the stream statistics (including the core ones)

std::size_t get_stat_index(const std::string &name) const

Helper to get the index of a specific statistic.

Throws:: std::out_of_range – if name is not a known statistic.

inline std::size_t next_stat_index() const: The index that will be returned by the next call to add_stat.

class stream : protected spead2::recv::stream_base

Stream that is fed by subclasses of reader.

The public interface to this class is thread-safe.

Subclassed by spead2::recv::chunk_stream, spead2::recv::chunk_stream_group_member, spead2::recv::ring_stream_base

Public Functions

template<typename T, typename ...Args> inline void emplace_reader(Args&&... args): Add a new reader by passing its constructor arguments, excluding the initial io_service and owner arguments.

void start()

Start the stream.

This is only needed if the config specifies explicit start (see stream_config::set_explicit_start). In that case, no new readers can be added after starting the stream.

virtual void stop()

Stop the stream.

After this returns, the io_service may still have outstanding completion handlers, but they should be no-ops when they’re called.

In most cases subclasses should override stop_received rather than this function. However, if heap_ready can block indefinitely, this function should be overridden to unblock it before calling the base implementation.

inline const stream_config &get_config() const: Get the stream’s configuration.

stream_stats get_stats() const

Return statistics about the stream.

See the Python documentation.

Protected Functions

inline const stream_config &get_config() const: Get the stream’s configuration.

void flush(): Flush the collection of live heaps, passing them to heap_ready.

stream_stats get_stats() const

Return statistics about the stream.

See the Python documentation.

A potentially more convenient interface is spead2::recv::ring_stream<Ringbuffer>, which places received heaps into a fixed-size thread-safe ring buffer. Another thread can then pull from this ring buffer in a loop. The template parameter selects the ringbuffer implementation. The default is a good light-weight choice, but if you need to use select()-like functions to wait for data, you can use spead2::ringbuffer<spead2::recv::live_heap, spead2::semaphore_fd, spead2::semaphore>.

class ring_stream_config

Parameters for configuring ring_stream.

Subclassed by spead2::recv::ring_stream_config_wrapper

Public Functions

ring_stream_config &set_heaps(std::size_t heaps): Set capacity of the ring buffer.

inline std::size_t get_heaps() const: Get capacity of the ring buffer.

ring_stream_config &set_contiguous_only(bool contiguous_only): Set whether only contiguous heaps are pushed to the ring buffer.

inline bool get_contiguous_only() const: Get whether only contiguous heaps are pushed to the ring buffer.

template<typename Ringbuffer = ringbuffer<live_heap>> class ring_stream : public spead2::recv::ring_stream_base

Specialisation of stream that pushes its results into a ringbuffer.

The ringbuffer class may be replaced, but must provide the same interface as ringbuffer. If the ring buffer fills up, add_packet will block the reader.

On the consumer side, heaps are automatically frozen as they are extracted.

This class is thread-safe.

Public Functions

explicit ring_stream(io_service_ref io_service, const stream_config &config = stream_config(), const ring_stream_config &ring_config = ring_stream_config())

Constructor.

Parameters:

io_service – I/O service (also used by the readers).
config – Stream configuration
ring_config – Ringbuffer configuration

heap pop()

Wait until a contiguous heap is available, freeze it, and return it; or until the stream is stopped.

Throws:: ringbuffer_stopped – if stop has been called and there are no more contiguous heaps.

template<typename ...SemArgs> live_heap pop_live(SemArgs&&... sem_args)

Wait until a heap is available and return it; or until the stream is stopped.

Parameters:: sem_args – Arbitrary arguments to pass to the data semaphore
Throws:: ringbuffer_stopped – if stop has been called and there are no more heaps.

heap try_pop()

Like pop, but if no contiguous heap is available, throws spead2::ringbuffer_empty.

Throws:

ringbuffer_empty – if there is no contiguous heap available, but the stream has not been stopped
ringbuffer_stopped – if stop has been called and there are no more contiguous heaps.

live_heap try_pop_live()

Like pop_live, but if no heap is available, throws spead2::ringbuffer_empty.

Throws:

ringbuffer_empty – if there is no heap available, but the stream has not been stopped
ringbuffer_stopped – if stop has been called and there are no more heaps.

spead2::detail::ringbuffer_iterator<ring_stream> begin()

Start iteration over the heaps in the ringbuffer.

This does not return a full-fledged iterator. It is intended only to enable a range-based for loop over the stream, and any other use of the iterator is unsupported.

For example: for (spead2::recv::heap heap : stream) { ... }

spead2::detail::ringbuffer_sentinel end(): End iterator (see begin).

inline const ring_stream_config &get_ring_config() const: Get the ringbuffer configuration passed to the constructor.

Readers

Reader classes are constructed inside a stream by calling spead2::recv::stream::emplace_reader().

class udp_reader : public spead2::recv::udp_reader_base

Asynchronous stream reader that receives packets over UDP.

Public Functions

udp_reader(stream &owner, const boost::asio::ip::udp::endpoint &endpoint, std::size_t max_size = default_max_size, std::size_t buffer_size = default_buffer_size)

Constructor.

If endpoint is a multicast address, then this constructor will subscribe to the multicast group, and also set SO_REUSEADDR so that multiple sockets can be subscribed to the multicast group.

Parameters:

owner – Owning stream
endpoint – Address on which to listen
max_size – Maximum packet size that will be accepted.
buffer_size – Requested socket buffer size. Note that the operating system might not allow a buffer size as big as the default.

udp_reader(stream &owner, const boost::asio::ip::udp::endpoint &endpoint, std::size_t max_size, std::size_t buffer_size, const boost::asio::ip::address &interface_address)

Constructor with explicit interface address (IPv4 only).

This overload is designed for use with multicast, but can also be used with a unicast endpoint as long as the address matches the interface address.

When a multicast group is used, the socket will have SO_REUSEADDR set, so that multiple sockets can all listen to the same multicast stream. If you want to let the system pick the interface for the multicast subscription, use boost::asio::ip::address_v4::any(), or use the default constructor.

Parameters:

owner – Owning stream
endpoint – Address and port
max_size – Maximum packet size that will be accepted.
buffer_size – Requested socket buffer size.
interface_address – Address of the interface which should join the group

Throws:

std::invalid_argument – If endpoint is not an IPv4 multicast address and does not match interface_address.
std::invalid_argument – If interface_address is not an IPv4 address

udp_reader(stream &owner, const boost::asio::ip::udp::endpoint &endpoint, std::size_t max_size, std::size_t buffer_size, unsigned int interface_index)

Constructor with explicit multicast interface index (IPv6 only).

The socket will have SO_REUSEADDR set, so that multiple sockets can all listen to the same multicast stream. If you want to let the system pick the interface for the multicast subscription, set interface_index to 0, or use the standard constructor.

Memory allocators

In addition to the memory allocators described in Memory allocators, new allocators can be created by subclassing spead2::memory_allocator. For an allocator set on a stream, a pointer to a spead2::recv::packet_header is passed as a hint to the allocator, allowing memory to be placed according to information in the packet. Note that if the stream_config has been configured to allow out-of-order packets then this could be any packet from the heap, rather than the first one.

class memory_allocator : public std::enable_shared_from_this<memory_allocator>

Polymorphic class for managing memory allocations in a memory pool.

This can be overloaded to provide custom memory allocations.

Subclassed by spead2::memory_pool, spead2::mmap_allocator, spead2::recv::detail::chunk_stream_allocator< CM >, spead2::unittest::legacy_allocator, spead2::unittest::mock_allocator

Public Functions

virtual pointer allocate(std::size_t size, void *hint)

Allocate size bytes of memory.

The default implementation uses new and pre-faults the memory.

The pointer type uses memory_allocator::deleter as the deleter. If the memory needs to be freed by something other than delete[] then pass a function object when constructing the pointer.

Parameters:

size – Number of bytes to allocate
hint – Usage-dependent extra information

Throws:

std::bad_alloc – if allocation failed

Returns:

Pointer to newly allocated memory

Private Functions

virtual void free(std::uint8_t *ptr, void *user)

Free memory previously returned from allocate.

This is kept only for backwards compatibility, so that existing allocators that override it will continue to work.

Parameters:

ptr – Value returned by allocate
user – User-defined handle stored in the deleter by allocate

class deleter : public std::function<void(std::uint8_t*)>

Deleter for pointers allocated by this allocator.

This class derives from std::function, so it can be constructed with any deleter at run time. In particular, this means that std::unique_ptr<std::uint8_t[], D> can be converted to memory_allocator::pointer for any deleter D.

For backwards compatibility, it can also be constructed from a shared pointer to an allocator and an arbitrary void pointer, in which case the deletion is performed by calling memory_allocator::free.

Public Functions

const std::shared_ptr<memory_allocator> &get_allocator() const

If the backwards-compatibility constructor was used, return the stored allocator.

Otherwise, returns a null pointer.

void *get_user() const

If the backwards-compatibility constructor was used, return the stored user pointer.

Otherwise, returns a null pointer.

The file examples/gdrapi_example.cu in the spead2 source distribution shows an example of using a custom memory allocator to allocate memory for heaps on the GPU.

Custom memory scatter

In specialised high-bandwidth cases, the overhead of assembling heaps in temporary storage before scattering the data into other arrangements can be very high. It is possible (since 1.11) to take complete control over the transfer of the payload of the SPEAD packets. Before embarking on such an approach, be sure you have a good understanding of the SPEAD protocol, particularly packets, heaps, item pointers and payload.

In the simplest case, each heap needs to be written to some special or pre-allocated storage, but in a contiguous fashion. In this case it is sufficient to provide a custom allocator (see above), which will return a pointer to the target storage.

In more complex cases, the contents of each heap, or even each packet, needs to be scattered to discontiguous storage areas. In this case, one can additionally override the memory copy function with set_memcpy() and providing a packet_memcpy_function.

typedef std::function<void(const spead2::memory_allocator::pointer &allocation, const packet_header &packet)> spead2::recv::packet_memcpy_function

It takes a pointer to the start of the heap’s allocation (as returned by the allocator) and the packet metadata. The default implementation is equivalent to the following:

void copy(const spead2::memory_allocator::pointer &allocation, const packet_header &packet)
{
    memcpy(allocation.get() + packet.payload_offset, packet.payload, packet.payload_length);
}

Note that when providing your own memory copy and allocator, you don’t necessarily need the allocator to actually return a pointer to payload memory. It could, for example, populate a structure that guides the copy, and return a pointer to that; or it could return a null pointer. There are some caveats though:

If the sender doesn’t provide the heap length item, then spead2 may need to make multiple allocations of increasing size as the heap grows, and each time it will copy (with standard memcpy, rather than your custom one) the old content to the new. Assuming you aren’t expecting such packets, you can reject them using set_allow_unsized_heaps().
spead2::recv::heap_base::get_items() constructs pointers to the items on the assumption of the default memcpy function, so if your replacement doesn’t copy things to the same place, you obviously won’t be able to use those pointers. Note that get_descriptors() will also not be usable.

See examples/gdrapi_example.cu in the spead2 source distribution for an example that copies data to a GPU.

Statistics

See Receiver stream statistics for an overview of statistics.

class stream_stats

Statistics about a stream.

Both a vector-like interface (indexing and size) and a map-like interface (with iterators and find) are provided. The iterators support random traversal, but are not technically random access iterators because dereferencing does not return a reference.

The public members provide direct access to the core statistics, for backwards compatibility. New code is advised to use the other interfaces. Indices for core statistics are available in stream_stat_indices.

Public Types

using key_type = std::string

using mapped_type = std::uint64_t

using value_type = std::pair<const std::string, std::uint64_t>

using size_type = std::size_t

using difference_type = std::ptrdiff_t

using reference = value_type&

using const_reference = const value_type&

using pointer = value_type*

using const_pointer = const value_type*

using iterator = detail::stream_stats_iterator<stream_stats, std::pair<const std::string&, std::uint64_t&>>

using const_iterator = detail::stream_stats_iterator<const stream_stats, const std::pair<const std::string&, const std::uint64_t&>>

using reverse_iterator = std::reverse_iterator<iterator>

using const_reverse_iterator = std::reverse_iterator<const_iterator>

Public Functions

stream_stats(): Construct with the default set of statistics, and all zero values.

explicit stream_stats(std::shared_ptr<const std::vector<stream_stat_config>> config): Construct with all zero values.

stream_stats(std::shared_ptr<const std::vector<stream_stat_config>> config, std::vector<std::uint64_t> values): Construct with provided values.

stream_stats(const stream_stats &other)

stream_stats &operator=(const stream_stats &other)

stream_stats(stream_stats &&other) = default

inline const std::vector<stream_stat_config> &get_config() const: Get the configuration of the statistics.

inline bool empty() const: Whether the container is empty.

inline std::size_t size() const: Get the number of statistics.

inline std::uint64_t &operator[](std::size_t index)

Access a statistic by index.

If index is out of range, behaviour is undefined.

inline const std::uint64_t &operator[](std::size_t index) const

Access a statistic by index.

If index is out of range, behaviour is undefined.

inline std::uint64_t &at(std::size_t index)

Access a statistic by index.

Throws:: std::out_of_range – if index is out of range.

inline const std::uint64_t &at(std::size_t index) const

Access a statistic by index.

Throws:: std::out_of_range – if index is out of range.

std::uint64_t &operator[](const std::string &name)

Access a statistic by name.

Throws:: std::out_of_range – if name is not the name of a statistic

const std::uint64_t &operator[](const std::string &name) const

Access a statistic by name.

Throws:: std::out_of_range – if name is not the name of a statistic

std::uint64_t &at(const std::string &name)

Access a statistic by name.

Throws:: std::out_of_range – if name is not the name of a statistic

const std::uint64_t &at(const std::string &name) const

Access a statistic by name.

Throws:: std::out_of_range – if name is not the name of a statistic

inline const_iterator cbegin() const noexcept

inline const_iterator cend() const noexcept

inline iterator begin() noexcept

inline iterator end() noexcept

inline const_iterator begin() const noexcept

inline const_iterator end() const noexcept

inline const_reverse_iterator crbegin() const noexcept

inline const_reverse_iterator crend() const noexcept

inline reverse_iterator rbegin() noexcept

inline reverse_iterator rend() noexcept

inline const_reverse_iterator rbegin() const noexcept

inline const_reverse_iterator rend() const noexcept

iterator find(const std::string &name)

Find element with the given name.

If not found, returns end().

const_iterator find(const std::string &name) const

Find element with the given name.

If not found, returns end().

std::size_t count(const std::string &name) const: Return the number of elements matching name (0 or 1).

stream_stats operator+(const stream_stats &other) const

Combine two sets of statistics.

Each statistic is combined according to its mode.