Namespaces
namespace	nlp_detail
namespace	srvn_detail
namespace	srvn_out

Classes
struct	BoolGraph
	A boolean sparse relation over element indices, e.g. More...
struct	Distrib
struct	LqnBalanceEquations
	The relation set of one layered model. More...
struct	LqnBoxBounds
	What lqn_boxbounds returns, in the layout of the reference's out struct. More...
class	LqnBuilder
struct	LqnCallGroup
	One routed call group: an activity, the strategy that picks among its targets, and the target ENTRIES in declaration order. More...
struct	LqnModel
	The intermediate model, and the second stage that flattens it. More...
struct	LqnMolInfo
	Everything lqn_mol reports beyond the four measure vectors. More...
struct	LqnMolOptions
	Tuning of the outer fixed point. More...
struct	LqnMolResult
	The four (nidx+1) vectors in the column convention SolverLN and LQNS report, so they line up with LN(model).getAvgTable cell for cell. More...
struct	LqnPrecedence
	One activity precedence of a task, with its activities resolved to indices. More...
struct	LqnRefCall
	One row of calls: the reference's [cidx, fromPos, toPos, aidx, vCall]. More...
struct	LqnRefGroup
	One group, the reference's R(g): the DAG below one REF task. More...
struct	LqnRefRoutes
	What lqn_ref_routes returns: the groups, or a non-empty why and no groups. More...
struct	LqnRelation
	One conservation law, as an aggregation over a node's structural neighbours. More...
struct	LqnSolution
	The iterates of a solved layered model, indexed by ABSOLUTE element index. More...
struct	LqnStruct
struct	LqnWriteReport
	What the schema could not carry, one human-readable line per loss. More...
struct	NlpLld
	A declared queue-dependent rate multiplier of one station: an expression in n, or a tabulated lldscaling row, which is refused because its interpolant has a corner at every knot. More...
struct	NlpOptions
	Options of the program: python export_nlp(lld=, alpha=, fork=). More...
class	NlpProgram
	The program of one layered model. More...
struct	NlpRate
	The queue-dependent rate of one station, r(n) = softmin(n, m) [* lld(n)], or n at a delay. More...
struct	NlpSolveResult
	What NlpProgram::solve returns: python's Result. More...
struct	ServerPools
	Heterogeneous server pools declared on a layer server, the twin of the nservertypes / servertypenames / serverspertype / servercompat / heterorates block a Network carries in sn.nodeparam{i}. More...
struct	SparseGraph
	A sparse square matrix over element indices, held as a dense vector of rows with an explicit nonzero list per row. More...

Typedefs
template<class T>
using	CdScaling
	A class-dependent scaling map, sn.cdscaling.

Enumerations
enum class	CallType
	Call kinds, with the values of MATLAB CallType. More...
enum class	LqnElement
	LQN element kinds, with the values of MATLAB LayeredNetworkElement. More...
enum class	PrecedenceType
	Activity precedence kinds, with the values of MATLAB ActivityPrecedenceType. More...
enum class	ReplacementStrategy
	Cache replacement policies, with the values of MATLAB ReplacementStrategy. More...
enum class	SchedStrategy
	Scheduling disciplines, with the values of MATLAB SchedStrategy. More...

Functions
template<class T>
LqnBalanceEquations< T >	lqn_balance_equations (const LqnStruct< T > &lqn, const LqnSolution< T > *sol=nullptr)
	Enumerate the conservation laws of the layered model lqn.
template<class T>
LqnBoxBounds< T >	lqn_boxbounds (const LqnStruct< T > &lqn)
	Evaluate the box bounds of lqn.
template<class S>
S	fj_max_exp (const std::vector< S > &m)
	E[max] of K independent EXPONENTIAL branch completion times of means m, with the reciprocals cleared so a branch of zero mean contributes zero: sum over nonempty S of (-1)^(\|S\|+1) prod_{i in S} m_i / sum_{i in S} prod_{j in S, j != i} m_j.
std::vector< double >	fj_max_exp_grad (const std::vector< double > &u)
	Gradient of fj_max_exp by complex step, one coordinate at a time.
template<class T>
NlpProgram	lqn_nlp_program (const LqnStruct< T > &lqn, const NlpOptions &opt=NlpOptions(), const std::string &name="")
	Build the nonlinear program of a layered model: python export_nlp without the emission, i.e.
template<class T>
LqnMolResult< T >	lqn_mol (const LqnStruct< T > &lsn, const LqnMolOptions &options=LqnMolOptions())
	Method of Layers on the SRVN decomposition of a layered queueing network whose entries carry no activity graph.
template<class T>
LqnRefRoutes< T >	lqn_ref_routes (const LqnStruct< T > &lqn, const std::vector< std::size_t > &callers, double maxpaths=32.0, const std::vector< std::size_t > &server_set={})
	Resolve the reference routes into the layer whose callers are CALLERS.
template<class T>
std::vector< double >	lqn_sustainable_mult (const LqnStruct< T > &l, const std::vector< double > &mult)
	Sustainable multiplicity (lsn_max_multiplicity) of every host and task of l under the declared multiplicities mult (tshift+ntasks+1).
template<class T>
LqnStruct< T >	lqn_finalize (const LqnModel< T > &m)
	Port of @LayeredNetwork/getStruct.m: flatten the model into its struct.
template<class T>
LqnModel< T >	read_lqnx_model_doc (std::unique_ptr< xml::Element > doc)
template<class T>
LqnModel< T >	read_lqnx_model (const std::string &path)
	read_lqnx_model on a .lqnx file.
template<class T>
LqnModel< T >	read_lqnx_model_text (const std::string &text)
	read_lqnx_model on a .lqnx document held in a string (the SRVN reader's translation, for one).
template<class T>
LqnStruct< T >	read_lqnx (const std::string &path)
	Read a .lqnx model.
template<class T>
LqnWriteReport	write_lqnx (const LqnModel< T > &m, const std::string &path, const std::string &model_name=std::string("LQN"), bool use_abstract_names=false, std::string *out_text=NULL)
	Write a layered model as a .lqnx document.
template<class T>
std::string	lqnx_to_string (const LqnModel< T > &m, const std::string &model_name=std::string("LQN"), bool use_abstract_names=false)
	The .lqnx document as a STRING, for a caller with no file to write to – the CLI's model-generation mode hands the XML back through its JSON envelope, and a test can compare documents without a temporary directory.
std::string	format_number (double x)
	Shortest C g spelling that round-trips x; the four codebases share this rule.
std::string	srvn2lqnx (const std::string &src, const std::string &name=std::string("model"))
	Translate SRVN text into an LQN XML document.
bool	is_srvn_file (const std::string &path)
	True when path holds SRVN text rather than XML: its first non-blank character is not '<'.
bool	is_srvn_path (const std::string &path)
	Whether path should be read as SRVN: a .in file, or a .lqn file that is not XML.
std::string	srvn_file_to_lqnx (const std::string &path)
	Translate the SRVN file path to .lqnx text, model name = its basename without extension.
template<class T>
LqnModel< T >	read_srvn_model (const std::string &path)
	Read an SRVN file (.in, or a .lqn that is not XML) into the intermediate LqnModel, via its .lqnx translation.
template<class T>
LqnStruct< T >	read_srvn (const std::string &path)
	read_srvn_model followed by lqn_finalize: the struct SolverLN consumes, as read_lqnx is.
std::string	lqnx2srvn (const std::string &xml_text, const std::string &name=std::string())
	Translate an LQN XML document into SRVN text.
template<class T>
void	write_srvn (const LqnModel< T > &m, const std::string &path, bool use_abstract_names=false)
	Write a layered model as SRVN text: write_lqnx, then lqnx2srvn, with LF line endings.

Typedef Documentation

◆ CdScaling

template<class T>

using line::lang::CdScaling

A class-dependent scaling map, sn.cdscaling.

It takes the per-class population vector at one station and returns the per-class rate multipliers, which is the signature pfqn_cdfun consumes; the alias resolves to the same std::function type as pfqn::CdScaling, so a map built here is passed straight through to the api layer.

Definition at line 736 of file lang_types.h.

Enumeration Type Documentation

◆ CallType

enum class line::lang::CallType

strong

Call kinds, with the values of MATLAB CallType.

Definition at line 474 of file lang_types.h.

◆ LqnElement

enum class line::lang::LqnElement

strong

LQN element kinds, with the values of MATLAB LayeredNetworkElement.

Definition at line 471 of file lang_types.h.

◆ PrecedenceType

enum class line::lang::PrecedenceType

strong

Activity precedence kinds, with the values of MATLAB ActivityPrecedenceType.

Definition at line 477 of file lang_types.h.

◆ ReplacementStrategy

enum class line::lang::ReplacementStrategy

strong

Cache replacement policies, with the values of MATLAB ReplacementStrategy.

Definition at line 385 of file lang_types.h.

◆ SchedStrategy

enum class line::lang::SchedStrategy

strong

Scheduling disciplines, with the values of MATLAB SchedStrategy.

Definition at line 185 of file lang_types.h.

Function Documentation

◆ fj_max_exp()

template<class S>

S line::lqn::fj_max_exp ( const std::vector< S > & m )

E[max] of K independent EXPONENTIAL branch completion times of means m, with the reciprocals cleared so a branch of zero mean contributes zero: sum over nonempty S of (-1)^(|S|+1) prod_{i in S} m_i / sum_{i in S} prod_{j in S, j != i} m_j.

Definition at line 621 of file lqn_export_nlp.h.

References fj_max_exp().

Referenced by fj_max_exp(), fj_max_exp_grad(), line::lqn::NlpProgram::residuals(), and line::lqn::NlpProgram::sweep().

◆ fj_max_exp_grad()

std::vector< double > line::lqn::fj_max_exp_grad ( const std::vector< double > & u )

inline

Gradient of fj_max_exp by complex step, one coordinate at a time.

Definition at line 642 of file lqn_export_nlp.h.

References fj_max_exp(), and fj_max_exp_grad().

Referenced by fj_max_exp_grad(), and line::lqn::NlpProgram::jacobian().

◆ format_number()

std::string line::lqn::format_number ( double x )

inline

Shortest C g spelling that round-trips x; the four codebases share this rule.

An integer below 1e15 in magnitude prints as d; anything else as the shortest %.{p}g, p = 1..17, that strtod reads back as x.

Exceptions

InputError (LINE:input:Format) when x is not finite (every caller checks first)

Definition at line 375 of file srvn_reader.h.

References format_number(), and line::InputError::InputError().

Referenced by format_number(), and line::lqn::srvn_out::num().

◆ is_srvn_file()

bool line::lqn::is_srvn_file ( const std::string & path )

inline

True when path holds SRVN text rather than XML: its first non-blank character is not '<'.

An empty file counts as SRVN, as in the reference.

Exceptions

InputError (LINE:input:File) when the file cannot be opened

Definition at line 1288 of file srvn_reader.h.

References line::InputError::InputError(), line::lqn::srvn_detail::is_space(), and is_srvn_file().

Referenced by is_srvn_file(), and is_srvn_path().

◆ is_srvn_path()

bool line::lqn::is_srvn_path ( const std::string & path )

inline

Whether path should be read as SRVN: a .in file, or a .lqn file that is not XML.

This is the line-cli and loader rule; .lqnx and .xml are never SRVN.

Definition at line 1303 of file srvn_reader.h.

References is_srvn_file(), and is_srvn_path().

Referenced by find_solver_report(), is_srvn_path(), line::io::read_layered_model(), and run_invocation().

◆ lqn_balance_equations()

template<class T>

LqnBalanceEquations< T > line::lqn::lqn_balance_equations	(	const LqnStruct< T > &	lqn,
		const LqnSolution< T > *	sol = nullptr )

Enumerate the conservation laws of the layered model lqn.

When sol is non-null every relation is also instantiated and its residual reported. On a converged fixed point they vanish to the solver's own tolerance; a residual that does not is either a documented convention difference (see mult against maxmult) or a defect in the solution.

Definition at line 466 of file lqn_balance_equations.h.

Referenced by lqn_balance_equations().

◆ lqn_boxbounds()

template<class T>

LqnBoxBounds< T > line::lqn::lqn_boxbounds ( const LqnStruct< T > & lqn )

Evaluate the box bounds of lqn.

A reference task of infinite multiplicity is treated as one customer, as the reference does: the bound is over a closed chain and an infinite population has no lower bound to report.

Definition at line 130 of file lqn_boxbounds.h.

Referenced by line::ln::SolverLN< T >::box_bounds(), and lqn_boxbounds().

◆ lqn_finalize()

template<class T>

LqnStruct< T > line::lqn::lqn_finalize ( const LqnModel< T > & m )

Port of @LayeredNetwork/getStruct.m: flatten the model into its struct.

Definition at line 501 of file lqn_reader.h.

Referenced by line::lqn::LqnBuilder< T >::build(), line::io::build_lqn_from_json(), line::opt::LineEvaluator::evaluate_layered_sensitivities(), line::io::lqn2java(), line::io::lqn2qn(), lqn_finalize(), read_lqnx(), read_srvn(), and write_lqnx().

◆ lqn_mol()

template<class T>

LqnMolResult< T > line::lqn::lqn_mol	(	const LqnStruct< T > &	lsn,
		const LqnMolOptions &	options = LqnMolOptions() )

Method of Layers on the SRVN decomposition of a layered queueing network whose entries carry no activity graph.

Parameters

lsn	LayeredNetworkStruct of an entry-only, closed, synchronous LQN
options	iteration cap, tolerance and under-relaxation factor

Definition at line 209 of file lqn_mol.h.

References line::lqn::LqnMolResult< T >::info, lqn_mol(), line::pfqn::pfqn_qdamva(), line::lqn::LqnMolResult< T >::QN, line::pfqn::QdAmvaResult< T >::R, line::lqn::LqnMolResult< T >::RN, line::lqn::LqnMolResult< T >::TN, and line::lqn::LqnMolResult< T >::UN.

Referenced by lqn_mol().

◆ lqn_nlp_program()

template<class T>

NlpProgram line::lqn::lqn_nlp_program	(	const LqnStruct< T > &	lqn,
		const NlpOptions &	opt = NlpOptions(),
		const std::string &	name = "" )

Build the nonlinear program of a layered model: python export_nlp without the emission, i.e.

_collect + _attach_lld + the script's index sets, layout, build_linear and build_residuals.

Exceptions

UnsupportedError LINE:unsupported:NlpExport naming every feature the form cannot express, a bad fork mode or alpha, or an unusable lld

Definition at line 1529 of file lqn_export_nlp.h.

Referenced by lqn_nlp_program().

◆ lqn_ref_routes()

template<class T>

LqnRefRoutes< T > line::lqn::lqn_ref_routes	(	const LqnStruct< T > &	lqn,
		const std::vector< std::size_t > &	callers,
		double	maxpaths = 32.0,
		const std::vector< std::size_t > &	server_set = {} )

Resolve the reference routes into the layer whose callers are CALLERS.

Parameters

lqn	the layered struct (after lqn_fwd_rendezvous, as SolverLN holds it)
callers	task indices that call the layer's server
maxpaths	refuse the layer above this many REF routes into it (default 32)
server_set	server elements of the layer; a prefix node whose task is one of them refuses

A caller reachable from two REF tasks is two INDEPENDENT customer pools, and the whole LAYER falls back rather than that caller alone: a refused caller may lie on another group's path, which would count its threads twice.

Definition at line 384 of file lqn_ref_routes.h.

References lqn_ref_routes().

Referenced by lqn_ref_routes().

◆ lqn_sustainable_mult()

template<class T>

std::vector< double > line::lqn::lqn_sustainable_mult	(	const LqnStruct< T > &	l,
		const std::vector< double > &	mult )

Sustainable multiplicity (lsn_max_multiplicity) of every host and task of l under the declared multiplicities mult (tshift+ntasks+1).

lqn_finalize passes l.mult; SolverLN passes them with the reference tasks trapped at a zero-capacity element set to 0, so the pools are sized without them.

Definition at line 451 of file lqn_reader.h.

Referenced by lqn_finalize(), and lqn_sustainable_mult().

◆ lqnx2srvn()

std::string line::lqn::lqnx2srvn	(	const std::string &	xml_text,
		const std::string &	name = std::string() )

inline

Translate an LQN XML document into SRVN text.

Parameters

xml_text	the LQN XML document
name	the model name for the header comment; empty means <lqn-model name> (or "model" when the root has none)

Returns: the SRVN document, newline-terminated

Exceptions

InputError	(LINE:input:Format) when the XML is malformed
UnsupportedError	(LINE:unsupported:Format) for content SRVN cannot express

Definition at line 549 of file srvn_writer.h.

References line::lqn::srvn_detail::is_space(), LINE_SRVN_ACT_ATTRS, lqnx2srvn(), line::lqn::srvn_out::Element::name, line::lqn::srvn_out::name(), and line::xml::parse().

Referenced by line::io::lqnx2srvn(), lqnx2srvn(), and write_srvn().

◆ lqnx_to_string()

template<class T>

std::string line::lqn::lqnx_to_string	(	const LqnModel< T > &	m,
		const std::string &	model_name = std::string("LQN"),
		bool	use_abstract_names = false )

The .lqnx document as a STRING, for a caller with no file to write to – the CLI's model-generation mode hands the XML back through its JSON envelope, and a test can compare documents without a temporary directory.

Definition at line 666 of file lqn_writer.h.

References lqnx_to_string(), and write_lqnx().

Referenced by lqnx_to_string(), run_generate(), and write_srvn().

◆ read_lqnx()

template<class T>

LqnStruct< T > line::lqn::read_lqnx ( const std::string & path )

Read a .lqnx model.

Parameters

path	file to read

Returns: the flattened struct SolverLN consumes

Definition at line 1752 of file lqn_reader.h.

References lqn_finalize(), read_lqnx(), and read_lqnx_model().

Referenced by line::io::read_layered_model(), and read_lqnx().

◆ read_lqnx_model()

template<class T>

LqnModel< T > line::lqn::read_lqnx_model ( const std::string & path )

read_lqnx_model on a .lqnx file.

Definition at line 1735 of file lqn_reader.h.

References line::xml::parse_file(), read_lqnx_model(), and read_lqnx_model_doc().

Referenced by read_lqnx(), line::io::read_lqnx_model(), and read_lqnx_model().

◆ read_lqnx_model_doc()

template<class T>

LqnModel< T > line::lqn::read_lqnx_model_doc ( std::unique_ptr< xml::Element > doc )

◆ read_lqnx_model_text()

template<class T>

LqnModel< T > line::lqn::read_lqnx_model_text ( const std::string & text )

read_lqnx_model on a .lqnx document held in a string (the SRVN reader's translation, for one).

Definition at line 1741 of file lqn_reader.h.

References line::xml::parse(), read_lqnx_model_doc(), and read_lqnx_model_text().

Referenced by read_lqnx_model_text(), and read_srvn_model().

◆ read_srvn()

template<class T>

LqnStruct< T > line::lqn::read_srvn ( const std::string & path )

read_srvn_model followed by lqn_finalize: the struct SolverLN consumes, as read_lqnx is.

Definition at line 1346 of file srvn_reader.h.

References lqn_finalize(), read_srvn(), and read_srvn_model().

Referenced by line::io::read_layered_model(), and read_srvn().

◆ read_srvn_model()

template<class T>

LqnModel< T > line::lqn::read_srvn_model ( const std::string & path )

Read an SRVN file (.in, or a .lqn that is not XML) into the intermediate LqnModel, via its .lqnx translation.

Definition at line 1340 of file srvn_reader.h.

References read_lqnx_model_text(), read_srvn_model(), and srvn_file_to_lqnx().

Referenced by read_srvn(), line::io::read_srvn_model(), and read_srvn_model().

◆ srvn2lqnx()

std::string line::lqn::srvn2lqnx	(	const std::string &	src,
		const std::string &	name = std::string("model") )

inline

Translate SRVN text into an LQN XML document.

Parameters

src	the SRVN document text
name	the model name written to <lqn-model name=...>

Returns: the .lqnx document, newline-terminated

Exceptions

InputError	(LINE:input:Format) on a syntax or reference error
UnsupportedError	(LINE:unsupported:Format) when the model uses a SPEX sweep

Definition at line 1239 of file srvn_reader.h.

References srvn2lqnx().

Referenced by line::io::srvn2lqnx(), srvn2lqnx(), and srvn_file_to_lqnx().

◆ srvn_file_to_lqnx()

std::string line::lqn::srvn_file_to_lqnx ( const std::string & path )

inline

Translate the SRVN file path to .lqnx text, model name = its basename without extension.

Errors name the file: "SRVN <path> (line n): ...".

Definition at line 1319 of file srvn_reader.h.

References line::InputError::InputError(), line::Error::kind(), line::lqn::srvn_detail::read_text(), srvn2lqnx(), srvn_file_to_lqnx(), line::lqn::srvn_detail::stem(), and line::UnsupportedError::UnsupportedError().

Referenced by read_srvn_model(), run_invocation(), and srvn_file_to_lqnx().

◆ write_lqnx()

template<class T>

LqnWriteReport line::lqn::write_lqnx	(	const LqnModel< T > &	m,
		const std::string &	path,
		const std::string &	model_name = std::string("LQN"),
		bool	use_abstract_names = false,
		std::string *	out_text = NULL )

Write a layered model as a .lqnx document.

Parameters

m	the intermediate model, from the builder or from read_lqnx_model
path	file to create
model_name	the name attribute of <lqn-model>
use_abstract_names	rename elements P1/T1/E1/A1, as writeXML's third argument does, for a document that is compared rather than read
out_text	when non-null, the document is SERIALIZED INTO IT and no file is written (path is then ignored). The tree cannot be returned instead: xml::Element owns its children through unique_ptr and they hold raw parent back pointers, so a moved tree dangles.

Returns: what the schema could not carry

Definition at line 220 of file lqn_writer.h.

Referenced by lqnx_to_string(), line::lqns::SolverLQNS< T >::run_analyzer(), line::io::write_lqnx(), and write_lqnx().

◆ write_srvn()

template<class T>

void line::lqn::write_srvn	(	const LqnModel< T > &	m,
		const std::string &	path,
		bool	use_abstract_names = false )

Write a layered model as SRVN text: write_lqnx, then lqnx2srvn, with LF line endings.

Parameters

m	the intermediate model
path	file to create
use_abstract_names	rename elements P1/T1/E1/A1 first, which also makes every name an LQNS identifier

Exceptions

UnsupportedError (LINE:unsupported:Format) for content SRVN cannot express

Definition at line 785 of file srvn_writer.h.

References line::InputError::InputError(), lqnx2srvn(), lqnx_to_string(), line::lqn::LqnModel< T >::name, and write_srvn().

Referenced by line::io::write_srvn(), and write_srvn().

Namespaces

Classes

Typedefs

Enumerations

Functions

Typedef Documentation

◆ CdScaling

Enumeration Type Documentation

◆ CallType

◆ LqnElement

◆ PrecedenceType

◆ ReplacementStrategy

◆ SchedStrategy

Function Documentation

◆ fj_max_exp()

◆ fj_max_exp_grad()

◆ format_number()

◆ is_srvn_file()

◆ is_srvn_path()

◆ lqn_balance_equations()

◆ lqn_boxbounds()

◆ lqn_finalize()

◆ lqn_mol()

◆ lqn_nlp_program()

◆ lqn_ref_routes()

◆ lqn_sustainable_mult()

◆ lqnx2srvn()

◆ lqnx_to_string()

◆ read_lqnx()

◆ read_lqnx_model()

◆ read_lqnx_model_doc()

◆ read_lqnx_model_text()

◆ read_srvn()

◆ read_srvn_model()

◆ srvn2lqnx()

◆ srvn_file_to_lqnx()

◆ write_lqnx()

◆ write_srvn()