Class Benchmark

Run the benchmark algorithm.

Inheritance
object
Benchmark
Inherited Members
Namespace: Lucene.Net.Benchmarks.ByTask
Assembly: Lucene.Net.Benchmark.dll
Syntax
public class Benchmark
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

Constructors

Benchmark(TextReader)

Run the benchmark algorithm.

Declaration
public Benchmark(TextReader algReader)
Parameters
Type Name Description
TextReader algReader
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

Properties

Algorithm

Returns the algorithm.

Declaration
public virtual Algorithm Algorithm { get; }
Property Value
Type Description
Algorithm
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

RunData

Returns the runData.

Declaration
public virtual PerfRunData RunData { get; }
Property Value
Type Description
PerfRunData
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

Methods

Exec(string[])

Utility: execute benchmark from command line.

Declaration
public static void Exec(string[] args)
Parameters
Type Name Description
string[] args

Single argument is expected: algorithm-file.
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

Execute()

Execute this benchmark.

Declaration
public virtual void Execute()
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.

Main(string[])

Run the benchmark algorithm.

LUCENENET specific: In the Java implementation, this Main method was intended to be called from the command line. However, in .NET a method within a DLL can't be directly called from the command line so we provide a .NET tool, lucene-cli, with a command that maps to this method: benchmark
Declaration
public static void Main(string[] args)
Parameters
Type Name Description
string[] args

Benchmark config and algorithm files.
Remarks
  1. Read algorithm.
  2. Run the algorithm.

Things to be added/fixed in "Benchmarking by tasks":
  1. TODO - report into Excel and/or graphed view.
  2. TODO - perf comparison between Lucene releases over the years.
  3. TODO - perf report adequate to include in Lucene nightly build site? (so we can easily track performance changes.)
  4. TODO - add overall time control for repeated execution (vs. current by-count only).
  5. TODO - query maker that is based on index statistics.