#include <bitplane_rze_stage.h>

Inheritance diagram for fz::BitplaneRZEStage:

Public Member Functions
void	setInverse (bool inv) override

bool	isGraphCompatible () const override

void	execute (cudaStream_t stream, MemoryPool pool, const std::vector< void > &inputs, const std::vector< void * > &outputs, const std::vector< size_t > &sizes) override

std::string	getName () const override

std::vector< size_t >	estimateOutputSizes (const std::vector< size_t > &input_sizes) const override

std::unordered_map< std::string, size_t >	getActualOutputSizesByName () const override

size_t	getActualOutputSize (int index) const override

uint16_t	getStageTypeId () const override

uint8_t	getOutputDataType (size_t) const override

uint8_t	getInputDataType (size_t) const override

size_t	serializeHeader (size_t, uint8_t *buf, size_t max_size) const override

void	deserializeHeader (const uint8_t *buf, size_t size) override

size_t	getMaxHeaderSize (size_t) const override

void	saveState () override

Public Member Functions inherited from fz::Stage
virtual size_t	getRequiredInputAlignment () const

virtual std::vector< std::string >	getOutputNames () const

int	getOutputIndex (const std::string &name) const

virtual void	setDims (const std::array< size_t, 3 > &dims)

virtual void	onFinalize (size_t, MemoryPool *)

virtual size_t	estimateDeviceFootprintBytes (size_t) const

virtual size_t	estimatePinnedFootprintBytes (size_t) const

virtual void	postStreamSync (cudaStream_t stream)

virtual size_t	estimateScratchBytes (const std::vector< size_t > &input_sizes) const

Detailed Description

Fused bitplane-transpose + zero-group run-zero encoder (FZ-GPU's lossless back-end), as a pipeline stage.

In one shared-memory pass per 4096-byte chunk the forward kernel:

bit-transposes a 32×32 matrix of uint32_t words (a bitshuffle at 4-byte element width), then
eliminates all-zero 4-byte groups, reserving compacted output space with a per-block atomicAdd. Input is interpreted as uint16_t symbols (two per word), matching FZ-GPU's quantizer codes; the typical pipeline is QuantizerStage<float,uint16_t> → BitplaneRZEStage. The transpose is built in, so a separate BitshuffleStage is not needed in front of it.

Relationship to BitshuffleStage + RZEStage

This stage is functionally close to BitshuffleStage(element_width=4) → RZEStage(levels=1), but it is not equivalent: (1) the transpose result never leaves shared memory, so it avoids a full global-memory round-trip (faster); (2) zero elimination is at 4-byte-group granularity rather than per-byte; (3) it emits FZ-GPU's exact self-describing archive (per-block atomic start positions + 128-byte header). It exists for FZ-GPU fidelity and throughput, not to add new functionality. See docs/stages/bitplane_rze.md.

Output is a self-describing byte archive (uint8_t): a 128-byte header followed by the bitflag bitmap, per-block start positions, and the compacted bitstream.

Note: Not graph-compatible. Both directions do a host-blocking cudaStreamSynchronize + D2H copy (the forward to read the compacted length, the inverse to read the archive header).; Fixed uint16_t input — no template parameter (matches the reference kernels). u8/u32 support would need new kernels.; Prior work: the encode/decode kernels are the FZ-GPU lossless codec (Boyuan Zhang et al., HPDC '23), BSD-3-Clause. The stage wrapper and memory-pool integration are FZGPUModules code. See THIRD_PARTY.md.

Member Function Documentation

◆ setInverse()

void fz::BitplaneRZEStage::setInverse ( bool inverse )

inlineoverridevirtual

Switch between forward (compression) and inverse (decompression) mode. Affects getNumInputs()/getNumOutputs() for stages with asymmetric port counts.

Reimplemented from fz::Stage.

◆ isGraphCompatible()

bool fz::BitplaneRZEStage::isGraphCompatible ( ) const

inlineoverridevirtual

Whether this stage is safe inside a CUDA Graph capture.

A stage is graph-compatible if execute() enqueues only device-side work (kernel launches, cudaMemcpyAsync D2D/H2D) and makes no host-synchronous calls. Override and return false if execute() contains D2H copies or dynamic decisions based on device data — the DAG will throw at setCaptureMode(true) time rather than producing a broken graph.

Default: true. Inverse-mode stages that do D2H reads (e.g. RZE inverse) must return false.

Reimplemented from fz::Stage.

◆ execute()

void fz::BitplaneRZEStage::execute	(	cudaStream_t	stream,
		MemoryPool *	pool,
		const std::vector< void * > &	inputs,
		const std::vector< void * > &	outputs,
		const std::vector< size_t > &	sizes
	)

overridevirtual

Execute the stage. Inputs, outputs, and sizes are device pointers/bytes.

Stages may call cudaStreamSynchronize(stream) or issue blocking D2H copies when the algorithm requires it (e.g. Huffman histogram readback for codebook construction, ANS renormalization tables). Such stages must return false from isGraphCompatible() and must document the sync points.

Note: the DAG dispatches sibling nodes (same topological level) via a sequential CPU loop, each enqueuing to its own stream. A sync inside execute() blocks the CPU from dispatching subsequent siblings until the synced stream is idle — this delays parallel branches in wide DAGs. In a linear pipeline there are no siblings and no extra cost.

Implements fz::Stage.

◆ getName()

std::string fz::BitplaneRZEStage::getName ( ) const

inlineoverridevirtual

Human-readable name used in error messages and debug output.

Implements fz::Stage.

◆ estimateOutputSizes()

std::vector< size_t > fz::BitplaneRZEStage::estimateOutputSizes ( const std::vector< size_t > & input_sizes ) const

overridevirtual

Estimate output buffer sizes given input sizes. Used for buffer allocation planning in PREALLOCATE mode — must be a safe upper bound; under-estimation causes buffer overruns.

Implements fz::Stage.

◆ getActualOutputSizesByName()

std::unordered_map< std::string, size_t > fz::BitplaneRZEStage::getActualOutputSizesByName ( ) const

inlineoverridevirtual

Actual output sizes after execute(), keyed by output port name.

Implements fz::Stage.

◆ getActualOutputSize()

size_t fz::BitplaneRZEStage::getActualOutputSize ( int index ) const

inlineoverridevirtual

Actual size of a single output by index after execute(). Avoids constructing the map for the common single-output case. Default delegates to getActualOutputSizesByName(); override to return directly from an internal field.

Reimplemented from fz::Stage.

◆ getStageTypeId()

uint16_t fz::BitplaneRZEStage::getStageTypeId ( ) const

inlineoverridevirtual

Stage type identifier written into the FZM file header.

Implements fz::Stage.

◆ getOutputDataType()

uint8_t fz::BitplaneRZEStage::getOutputDataType ( size_t output_index ) const

inlineoverridevirtual

DataType enum of the given output port.

Implements fz::Stage.

◆ getInputDataType()

uint8_t fz::BitplaneRZEStage::getInputDataType ( size_t ) const

inlineoverridevirtual

Expected DataType of the given input port.

Used by Pipeline::finalize() to detect type mismatches between connected stages before any execution. Return DataType::UNKNOWN to opt out of checking — byte-transparent stages (Bitshuffle, RZE, RRE) and mock stages must return UNKNOWN; finalize() skips any connection where either side is UNKNOWN.

Reimplemented from fz::Stage.

◆ serializeHeader()

size_t fz::BitplaneRZEStage::serializeHeader	(	size_t	output_index,
		uint8_t *	header_buffer,
		size_t	max_size
	)		const

inlineoverridevirtual

Serialize stage config into header_buffer (max 128 bytes) for the FZM file. Return the number of bytes written, or 0 if the stage has no config.

Reimplemented from fz::Stage.

◆ deserializeHeader()

void fz::BitplaneRZEStage::deserializeHeader	(	const uint8_t *	header_buffer,
		size_t	size
	)

inlineoverridevirtual

Restore stage config from header_buffer during decompression.

Reimplemented from fz::Stage.

◆ getMaxHeaderSize()

size_t fz::BitplaneRZEStage::getMaxHeaderSize ( size_t output_index ) const

inlineoverridevirtual

Maximum bytes this stage writes into its per-output FZM header slot.

Reimplemented from fz::Stage.

◆ saveState()

void fz::BitplaneRZEStage::saveState ( )

inlineoverridevirtual

Save/restore config state around a decompression pass. deserializeHeader() overwrites the stage's forward-pass config; saveState() is called before and restoreState() after so the stage returns to its original configuration.

Reimplemented from fz::Stage.

Public Member Functions

Detailed Description

Relationship to BitshuffleStage + RZEStage

Member Function Documentation

◆ setInverse()

◆ isGraphCompatible()

◆ execute()

◆ getName()

◆ estimateOutputSizes()

◆ getActualOutputSizesByName()

◆ getActualOutputSize()

◆ getStageTypeId()

◆ getOutputDataType()

◆ getInputDataType()

◆ serializeHeader()

◆ deserializeHeader()

◆ getMaxHeaderSize()

◆ saveState()