Multistep load balancer where no processor topology information is used. This first performs orb partition at the node level and assigns the partitioned TreePieces to the PEs belonging to the node. Finally after this assignment, a refinement is done. More...

#include <MultistepNodeLB_notopo.h>

Inheritance diagram for MultistepNodeLB_notopo:

Public Member Functions
	MultistepNodeLB_notopo (const CkLBOptions &)

	MultistepNodeLB_notopo (CkMigrateMessage *m)

void	work (BaseLB::LDStats *stats)
	Implement load balancing: store loads and decide whether to do ORB3D load balancing.

void	balanceTPs (BaseLB::LDStats *stats)

void	balanceTPsNode (BaseLB::LDStats *stats)
	Refinement strategy to distribute TreePieces evenly among nodes. If heavily loaded nodes are detect, this moves TreePieces off of heavily loaded nodes to underloaded nodes. Note that this does not take into account communication costs.

void	work2 (BaseLB::LDStats *stats, int count)
	ORB3D load balance across nodes (as opposed to processors).

void	greedy (BaseLB::LDStats *stats, int count)

void	pup (PUP::er &p)

Additional Inherited Members
Protected Member Functions inherited from Orb3dCommon
void	orbPePartition (vector< Event > events, vector< OrbObject > &tp, int node, BaseLB::LDStats stats)

void	orbPartition (vector< Event > events, OrientedBox< float > &box, int nprocs, vector< OrbObject > &tp, BaseLB::LDStats stats, bool node_partition=false)
	Recursively partition treepieces among processors by bisecting the load in orthogonal directions.

void	orbPrepare (vector< Event > tpEvents, OrientedBox< float > &box, int numobjs, BaseLB::LDStats stats, bool node_partition=false)
	Prepare structures for the ORB partition.

void	refine (BaseLB::LDStats *stats, int numobjs)

int	partitionRatioLoad (vector< Event > &events, float ratio, float bglp, float bgrp)
	Given a vector of Events, find a split that partitions them into two partitions with a given ratio of loads.

Protected Attributes inherited from Orb3dCommon
decltype(BaseLB::LDStats::to_proc) *	mapping

decltype(BaseLB::LDStats::from_proc) *	from

CkVec< float >	procload

double	maxPieceProc

int	nextProc
	index of first processor of the group we are considering

Detailed Description

Multistep load balancer where no processor topology information is used. This first performs orb partition at the node level and assigns the partitioned TreePieces to the PEs belonging to the node. Finally after this assignment, a refinement is done.

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

MultistepNodeLB_notopo.h
MultistepNodeLB_notopo.cpp

Public Member Functions

Additional Inherited Members

Detailed Description