gcl::CollectiveBalancedReorder Class Reference

Helper class to reorder a tensor in a per-tile-balanced fashion such that each replica obtains (for inputs to AllGather or outputs of ReduceScatter) an equally sized 1D tensor with equally sized regions. More...

#include <CollectiveBalancedReorder.hpp>

Public Member Functions
	CollectiveBalancedReorder (poplar::Graph &graph_, poplar::Tensor tensor_, unsigned replicationFactor_, const poplar::DebugNameAndId &dnai_, bool allowElementMap=false, unsigned grainSize=1)
	Constructor. More...
poplar::Tensor	createReplicaSlice (const poplar::Type &type)
	Create a tensor mapped efficiently over the same tiles as the reference tensor. More...
poplar::Tensor	createCollectivesTensor (const poplar::Type &type, const std::string &debugPrefix)
	Create a tensor mapped efficiently over the same tiles as the reference tensor. More...
poplar::Tensor	undoRearrangeForCollective (const poplar::Tensor &tensor) const
	Reorder tensor back into the expected IR tensor shape and order. More...
std::vector< std::size_t >	getReferenceShape () const
	Get the shape of the reference tensor. More...
const CollectiveBalancedHostRearrangement &	getHostRearrangement () const
	Get a helper class that implements allows to apply the rearrangement on the host. More...
void	zeroPaddingInCollectiveTensor (poplar::Tensor &collectiveTensor, poplar::program::Sequence &prog) const
	Zero the padding of a collective friendly tensor. More...

Detailed Description

Helper class to reorder a tensor in a per-tile-balanced fashion such that each replica obtains (for inputs to AllGather or outputs of ReduceScatter) an equally sized 1D tensor with equally sized regions.

This helper class reduces the memory used by the syncful collective. The reordering process:

• Flattens the input tensor
• Analyses the tile mapping
• Determines reordering strategy and required internal padding
• Can rearrange and undo the rearrangement on any tensor that has the same tile mapping
• Can rearrange and undo the rearrangement on host tensors that are to be copied into CBR-rearranged RemoteBuffers

Constructor & Destructor Documentation

CollectiveBalancedReorder()

gcl::CollectiveBalancedReorder::CollectiveBalancedReorder	(	poplar::Graph &	graph_,
		poplar::Tensor	tensor_,
		unsigned	replicationFactor_,
		const poplar::DebugNameAndId &	dnai_,
		bool	allowElementMap = false,
		unsigned	grainSize = 1
	)

Constructor.

Parameters

graph_	The poplar graph.
tensor_	The reference tensor to rearrange.
replicationFactor_	The replication factor of the graph.
dnai_	Debug name and id.
allowElementMap	Allow alternative representation of the host rearrangements. Sometimes it is beneficial to collapse all intervals into simple 1-to-1 element map. This flag should be set true in all new code and deprecated when all frameworks implement serialisation of newly added elementMap field.
grainSize	The grain size to use when padding the tensor.

Member Function Documentation

createCollectivesTensor()

poplar::Tensor gcl::CollectiveBalancedReorder::createCollectivesTensor	(	const poplar::Type &	type,
		const std::string &	debugPrefix
	)

Create a tensor mapped efficiently over the same tiles as the reference tensor.

The returned tensor has the size of the input of the reduce scatter and of the result of the all gather.

Parameters

type	The type to use when creating the tensor.
debugPrefix	The debug prefix.

Returns

The efficient tensor created from the reference.

createReplicaSlice()

poplar::Tensor gcl::CollectiveBalancedReorder::createReplicaSlice

(

const poplar::Type &

type

)

Create a tensor mapped efficiently over the same tiles as the reference tensor.

The returned tensor has the size of the result of the reduce scatter and of the input of the all gather.

Parameters

type	The type to use when creating the tensor.

Returns

The efficient tensor created from the reference.

getHostRearrangement()

const CollectiveBalancedHostRearrangement & gcl::CollectiveBalancedReorder::getHostRearrangement ( ) const

inline

Get a helper class that implements allows to apply the rearrangement on the host.

Returns

The helper class for host rearrangement.

getReferenceShape()

std::vector< std::size_t > gcl::CollectiveBalancedReorder::getReferenceShape ( ) const

inline

Get the shape of the reference tensor.

Returns

The shape of the reference tensor.

undoRearrangeForCollective()

poplar::Tensor gcl::CollectiveBalancedReorder::undoRearrangeForCollective

(

const poplar::Tensor &

tensor

)

const

Reorder tensor back into the expected IR tensor shape and order.

Parameters

tensor

The tensor to rearrange.

Returns

The tensor with the rearrangement undone.

zeroPaddingInCollectiveTensor()

void gcl::CollectiveBalancedReorder::zeroPaddingInCollectiveTensor	(	poplar::Tensor &	collectiveTensor,
		poplar::program::Sequence &	prog
	)		const

Zero the padding of a collective friendly tensor.

Parameters

collectiveTensor	The collective tensor to zero the padding of.
prog	The sequence to add the zeroing program to.

---

The documentation for this class was generated from the following file:

• include/gcl/CollectiveBalancedReorder.hpp