lm::main::ResourceAllocator Class Reference

An object that tracks the available resources for the main simulation runner. More...

#include <ResourceAllocator.h>

Classes
class	ComputeResources
	A representation for the resources for a given node. More...

Public Member Functions
	ResourceAllocator (int processNumber, int numberCpuCores, float cpuCoresPerReplicate)
	Create a ResourceAllocator.
	ResourceAllocator (int processNumber, int numberCpuCores, float cpuCoresPerReplicate, vector< int > cudaDevices, float cudaDevicesPerReplicate)
	Create a ResourceAllocator.
virtual	~ResourceAllocator ()
virtual int	getMaxSimultaneousReplicates ()
	Get maximum number of replicates that can run at a time on the available resources based on cudaDevices and numberCpuCores.
virtual ComputeResources	assignReplicate (int replicate)
	Assign a replicate to free resources.
virtual void	removeReplicate (int replicate)
	Remove a replicate from those that are running.
virtual int	reserveCpuCore ()
	Reserve a particular core.

Static Public Member Functions
static vector< int >	getAvailableCpuCores ()
	Parse CPU affinity list from /proc/self/status.

Protected Attributes
pthread_mutex_t	mutex
int	processNumber
int	numberCpuCores
int	reservedCpuCores
vector< int >	availableCpuCores
vector< int >	cudaDevices
int	cpuSlotsPerCore
int	cpuSlotsPerReplicate
int	cudaSlotsPerDevice
int	cudaSlotsPerReplicate
int **	cpuSlots
int **	cudaSlots

Detailed Description

An object that tracks the available resources for the main simulation runner.

Constructor & Destructor Documentation

ResourceAllocator() [1/2]

lm::main::ResourceAllocator::ResourceAllocator	(	int	processNumber,
		int	numberCpuCores,
		float	cpuCoresPerReplicate )

Create a ResourceAllocator.

Parameters

processNumber	The process identifier
numberCpuCores	The number of cores on teh resource
cpuCoresPerPrelicate	The number of cores to be used by each simulation replicate

Here is the call graph for this function:

ResourceAllocator() [2/2]

lm::main::ResourceAllocator::ResourceAllocator	(	int	processNumber,
		int	numberCpuCores,
		float	cpuCoresPerReplicate,
		vector< int >	cudaDevices,
		float	cudaDevicesPerReplicate )

Create a ResourceAllocator.

Parameters

processNumber	The process identifier
numberCpuCores	The number of cores on teh resource
cpuCoresPerPrelicate	The number of cores to be used by each simulation replicate
cudaDevices	A set of the identifiers for the available CUDA devices
cudaDevicesPerReplicate	The number of GPUs to be used by each simulation replicate

Here is the call graph for this function:

~ResourceAllocator()

lm::main::ResourceAllocator::~ResourceAllocator ( )

virtual

Member Function Documentation

assignReplicate()

ResourceAllocator::ComputeResources lm::main::ResourceAllocator::assignReplicate ( int replicate )

virtual

Assign a replicate to free resources.

Parameters

replicate

The replicate identifier

Returns

A class representing the compute resources given to the replicate

Here is the call graph for this function:

Here is the caller graph for this function:

getAvailableCpuCores()

vector< int > lm::main::ResourceAllocator::getAvailableCpuCores ( )

static

Parse CPU affinity list from /proc/self/status.

Returns

Vector of available CPU core indices

Here is the caller graph for this function:

getMaxSimultaneousReplicates()

int lm::main::ResourceAllocator::getMaxSimultaneousReplicates ( )

virtual

Get maximum number of replicates that can run at a time on the available resources based on cudaDevices and numberCpuCores.

Here is the caller graph for this function:

removeReplicate()

void lm::main::ResourceAllocator::removeReplicate ( int replicate )

virtual

Remove a replicate from those that are running.

Parameters

replicate

The replicate identifier

Here is the caller graph for this function:

reserveCpuCore()

int lm::main::ResourceAllocator::reserveCpuCore ( )

virtual

Reserve a particular core.

Returns

The identifer fo the core that has been reserved

Here is the caller graph for this function:

Member Data Documentation

availableCpuCores

vector<int> lm::main::ResourceAllocator::availableCpuCores

protected

cpuSlots

int** lm::main::ResourceAllocator::cpuSlots

protected

cpuSlotsPerCore

int lm::main::ResourceAllocator::cpuSlotsPerCore

protected

cpuSlotsPerReplicate

int lm::main::ResourceAllocator::cpuSlotsPerReplicate

protected

cudaDevices

vector<int> lm::main::ResourceAllocator::cudaDevices

protected

cudaSlots

int** lm::main::ResourceAllocator::cudaSlots

protected

cudaSlotsPerDevice

int lm::main::ResourceAllocator::cudaSlotsPerDevice

protected

cudaSlotsPerReplicate

int lm::main::ResourceAllocator::cudaSlotsPerReplicate

protected

mutex

pthread_mutex_t lm::main::ResourceAllocator::mutex

protected

numberCpuCores

int lm::main::ResourceAllocator::numberCpuCores

protected

processNumber

int lm::main::ResourceAllocator::processNumber

protected

reservedCpuCores

int lm::main::ResourceAllocator::reservedCpuCores

protected

---

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

• /data2/LM_zls_github/Lattice-Microbes/src/core/ResourceAllocator.h
• /data2/LM_zls_github/Lattice-Microbes/src/core/ResourceAllocator.cpp