org.gridgain.grid.benchmarks
Class GridLocalNodeBenchmark

java.lang.Object
  org.gridgain.grid.benchmarks.GridLocalNodeBenchmark

All Implemented Interfaces:

Serializable

public class GridLocalNodeBenchmark
extends Object
implements Serializable

Auto-benchmarking facility for local node. This class represent Spring-friendly JavaBean that performs requested benchmarks and exposes their results (scores) via getters. This bean can be used by Spring configuration to initialize local node attributes with benchmark values that can be later utilized by map/reduce logic in heterogeneous environment to perform a weighted split. Constructor of the JavaBean allows to specify which benchmarks need to be executed, and method GridLocalNodeBenchmark.start() actually executes them or loads them from local file store if they were previously saved (in Spring this method would be an initialization method for this bean).

Note that benchmark scores don't depend directly on the size of the specific benchmark (since its score is weighted by the time). However, as a rule of the thumb a single benchmark that takes less than several seconds is likely to be inaccurate, and a single benchmark that takes more than a several minutes to complete is likely to be excessive.

Using default benchmark sizes will result in approximately 4 minutes running total for all 5 benchmarks on Pentium 4 3.0 GHz. It is recommended that benchmark caching is turned on (basically flags loadResults and saveResults are set to true.

You can initialize local node benchmarks by adding/uncommenting the following section in GridGain Spring XML file:

 <property name="userAttributes">
     <map>
         <entry key="grid.node.benchmark">
             <bean class="org.gridgain.grid.benchmarks.GridLocalNodeBenchmark" init-method="start"/>
         </entry>
     </map>
 </property>
 

You can configure load balancing to automatically use node benchmarks to distribute jobs on the grid. Here is an example of how Spring XML configuration would look like:

 <property name="loadBalancingSpi">
     <bean class="org.gridgain.grid.spi.loadbalancing.adaptive.GridAdaptiveLoadBalancingSpi">
         <property name="loadProbe">
             <bean class="org.gridgain.grid.spi.loadbalancing.adaptive.GridAdaptiveBenchmarkLoadProbe">
                 <!-- Specify name of benchmark node attribute (the same as above). -->
                 <property name="benchmarkAttributeName" value="grid.node.benchmark"/>

                 <!-- Benchmarks scores to use. -->
                 <property name="useIntegerScore" value="true"/>
                 <property name="useLongScore" value="true"/>
                 <property name="useDoulbeScore" value="true"/>
                 <property name="useIoScore" value="false"/>
                 <property name="useTrigonometryScore" value="false"/>
             </bean>
         </property>
     </bean>
 </property>
 

Please make sure to properly initialize this probe to use exactly the scores you need in your grid. For example, if your jobs don't do any I/O, then you probably should disable I/O score. If you are not doing any trigonometry calculations, then you should disable trigonometry score.

Original idea of these benchmarks is by Christopher W. Cowell-Shah.



See Also:

Documentation
Email Support
Online Forums
Issue Tracking

Author: 2005-2008 Copyright © GridGain Systems. All Rights Reserved. ver. 2.1.0

	Nikita Ivanov, GridGain Systems nivanov@gridgain.com
	Dmitriy Setrakyan, GridGain Systems dsetrakyan@gridgain.com
	Alexander Magdenko, GridGain Systems amagdenko@gridgain.com
	Boris Suchodoev, GridGain Systems bsuchodoev@gridgain.com
	Denis Kharlamov, GridGain Systems dkharlamov@gridgain.com
	close

See Also:

Serialized Form

Field Summary

static long	DFLT_DOUBLE_SIZE Default benchmark size.
static long	DFLT_INT_SIZE Default benchmark size.
static long	DFLT_IO_SIZE Default benchmark size.
static long	DFLT_LONG_SIZE Default benchmark size.
static String	DFLT_PATH Default file path where benchmark results placed (value is work/benchmarks/localnode.xml).
static long	DFLT_TRIG_SIZE Default benchmark size.

Constructor Summary

GridLocalNodeBenchmark()
Creates new bean with all benchmarks turned on.

GridLocalNodeBenchmark(boolean runInt, boolean runLong, boolean runDouble, boolean runTrig, boolean runIo)
Creates new bean with specific benchmarks turned on and off.

GridLocalNodeBenchmark(boolean runInt, boolean runLong, boolean runDouble, boolean runTrig, boolean runIo, boolean loadResults, boolean saveResults)
Creates new bean with specific benchmarks turned on and off.

GridLocalNodeBenchmark(boolean runInt, boolean runLong, boolean runDouble, boolean runTrig, boolean runIo, boolean loadResults, boolean saveResults, String filePath)
Creates new bean with specific benchmarks turned on and off.

Method Summary

long	getDoubleScore() Gets relative score of double operations for this benchmark on the local node.
long	getIntegerScore() Gets relative score of int operations for this benchmark on the local node.
long	getIoScore() Gets relative score of I/O operations for this benchmark on the local node.
long	getLongScore() Gets relative score of long operations for this benchmark on the local node.
long	getTrigonometryScore() Gets relative score of trigonometry operations for this benchmark on the local node.
void	runBenchmarks() Runs all requested benchmarks.
void	setDoubleBenchmarkSize(long doubleSize) Sets double benchmark size.
void	setIntegerBenchmarkSize(long intSize) Sets int benchmark size.
void	setIoBenchmarkSize(long ioSize) Sets I/O benchmark size.
void	setLongBenchmarkSize(long longSize) Sets long benchmark size.
void	setTrigonometryBenchmarkSize(long trigSize) Sets trigonometry benchmark size.
void	start() Runs all requested benchmarks or load if benchmark results exist.
String	toString()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

DFLT_INT_SIZE

public static final long DFLT_INT_SIZE

Default benchmark size.

See Also:

Constant Field Values

DFLT_LONG_SIZE

public static final long DFLT_LONG_SIZE

Default benchmark size.

See Also:

Constant Field Values

DFLT_TRIG_SIZE

public static final long DFLT_TRIG_SIZE

Default benchmark size.

See Also:

Constant Field Values

DFLT_DOUBLE_SIZE

public static final long DFLT_DOUBLE_SIZE

Default benchmark size.

See Also:

Constant Field Values

DFLT_IO_SIZE

public static final long DFLT_IO_SIZE

Default benchmark size.

See Also:

Constant Field Values

DFLT_PATH

public static final String DFLT_PATH

Default file path where benchmark results placed (value is work/benchmarks/localnode.xml).

See Also:

Constant Field Values

Constructor Detail

GridLocalNodeBenchmark

public GridLocalNodeBenchmark()

Creates new bean with all benchmarks turned on.

Caching of benchmark results is turned on which means that every time an instance of this class is created, it will try to upload stored benchmark results from file system, if if not found, it will calculate benchmarks and then store them on the file system.

GridLocalNodeBenchmark

public GridLocalNodeBenchmark(boolean runInt,
                              boolean runLong,
                              boolean runDouble,
                              boolean runTrig,
                              boolean runIo)

Creates new bean with specific benchmarks turned on and off.

Caching of benchmark results is turned on which means that every time an instance of this class is created, it will try to upload stored benchmark results from file system, if if not found, it will calculate benchmarks and then store them on the file system.

Parameters:

runInt - true to turn int benchmark on, false otherwise.

runLong - true to turn long benchmark on, false otherwise.

runDouble - true to turn double benchmark on, false otherwise.

runTrig - true to turn trigonometry benchmark on, false otherwise.

runIo - true to turn I/O benchmark on, false otherwise.

GridLocalNodeBenchmark

public GridLocalNodeBenchmark(boolean runInt,
                              boolean runLong,
                              boolean runDouble,
                              boolean runTrig,
                              boolean runIo,
                              boolean loadResults,
                              boolean saveResults)

Creates new bean with specific benchmarks turned on and off.

Parameters:

runInt - true to turn int benchmark on, false otherwise.

runLong - true to turn long benchmark on, false otherwise.

runDouble - true to turn double benchmark on, false otherwise.

runTrig - true to turn trigonometry benchmark on, false otherwise.

runIo - true to turn I/O benchmark on, false otherwise.

loadResults - true to load benchmark results from file, false ignore file.

saveResults - true to save benchmark result in file, false otherwise.

GridLocalNodeBenchmark

public GridLocalNodeBenchmark(boolean runInt,
                              boolean runLong,
                              boolean runDouble,
                              boolean runTrig,
                              boolean runIo,
                              boolean loadResults,
                              boolean saveResults,
                              String filePath)

Creates new bean with specific benchmarks turned on and off.

Parameters:

runInt - true to turn int benchmark on, false otherwise.

runLong - true to turn long benchmark on, false otherwise.

runDouble - true to turn double benchmark on, false otherwise.

runTrig - true to turn trigonometry benchmark on, false otherwise.

runIo - true to turn I/O benchmark on, false otherwise.

loadResults - true to load benchmark results from file, false ignore file.

saveResults - true to save benchmark result in file, false otherwise.

filePath - Benchmark results file path. This should be either absolute or relative path within GRIDGAIN_HOME. Benchmarks results will be loader and saved into this file depending on flags loadResults and saveResults. Note that default value is DFLT_PATH.

Method Detail

start

public void start()
           throws IllegalArgumentException

Runs all requested benchmarks or load if benchmark results exist. In Spring context it should be a initializer method.

Throws:

IllegalArgumentException - If loading of results is turned on and results could not be loaded for whatever reason.

runBenchmarks

public void runBenchmarks()

Runs all requested benchmarks. In Spring context it should be a initializer method.

setIntegerBenchmarkSize

public void setIntegerBenchmarkSize(long intSize)

Sets int benchmark size.

Parameters:

intSize - int benchmark size.

setLongBenchmarkSize

public void setLongBenchmarkSize(long longSize)

Sets long benchmark size.

Parameters:

longSize - long benchmark size.

setDoubleBenchmarkSize

public void setDoubleBenchmarkSize(long doubleSize)

Sets double benchmark size.

Parameters:

doubleSize - double benchmark size.

setTrigonometryBenchmarkSize

public void setTrigonometryBenchmarkSize(long trigSize)

Sets trigonometry benchmark size.

Parameters:

trigSize - Trigonometry benchmark size.

setIoBenchmarkSize

public void setIoBenchmarkSize(long ioSize)

Sets I/O benchmark size.

Parameters:

ioSize - I/O benchmark size.

getIntegerScore

public long getIntegerScore()

Gets relative score of int operations for this benchmark on the local node. Note that this score is only relative to other int scores obtained from this benchmark on this or other nodes. Note that scores for this benchmark can vary even on the same node from run to run as they are executed in the same VM as grid instance itself and are thus subject to all VM optimization configurations.

Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

Returns:

Relative score of int operations for this benchmark. Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

getLongScore

public long getLongScore()

Gets relative score of long operations for this benchmark on the local node. Note that this score is only relative to other long scores obtained from this benchmark on this or other nodes. Note that scores for this benchmark can vary even on the same node from run to run as they are executed in the same VM as grid instance itself and are thus subject to all VM optimization configurations.

Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

Returns:

Relative score of long operations for this benchmark. Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

getDoubleScore

public long getDoubleScore()

Gets relative score of double operations for this benchmark on the local node. Note that this score is only relative to other double scores obtained from this benchmark on this or other nodes. Note that scores for this benchmark can vary even on the same node from run to run as they are executed in the same VM as grid instance itself and are thus subject to all VM optimization configurations.

Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

Returns:

Relative score of double operations for this benchmark. Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

getTrigonometryScore

public long getTrigonometryScore()

Gets relative score of trigonometry operations for this benchmark on the local node. Note that this score is only relative to other trigonometry scores obtained from this benchmark on this or other nodes. Note that scores for this benchmark can vary even on the same node from run to run as they are executed in the same VM as grid instance itself and are thus subject to all VM optimization configurations.

Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

Returns:

Relative score of trigonometry operations for this benchmark. Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

getIoScore

public long getIoScore()

Gets relative score of I/O operations for this benchmark on the local node. Note that this score is only relative to other I/O scores obtained from this benchmark on this or other nodes. Note that scores for this benchmark can vary even on the same node from run to run as they are executed in the same VM as grid instance itself and are thus subject to all VM optimization configurations.

Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

Returns:

Relative score of I/O operations for this benchmark. Higher score means better performance (faster execution of the benchmark). Minimum value is zero.

toString

public String toString()

Overrides:

toString in class Object