Allow to configure benchmark from CLI

[GalRogozinski]

Description

We have added the following flags to benchmark runner:
-Dforks - How may time should the jvm fork. Running more than one fork will cause more iterations but should reset JVM optimization profiles. May reduce variance. Default Value: 1.
-DwarmUpIterations - How many warm up iterations we should run before we stop measuring. Default value: 5
-DmeasurementIterations - How many measurement iterations we should have. Default value: 10
-Dthreads - How many threads will run the test (concurrency). Default value: # of cores.

Example of how to run from CLI:
mvn -Dtest=BenchmarkRunner#launchDbBenchmarks -Dforks=2 -DmeasurementIterations=6 -DwarmUpIterations=2 -Dthreads=3 test

Fixes #1651

Type of change

• Test Enhancement (a non-breaking change which adds functionality)

How Has This Been Tested?

Locally run the example line. The following was printed:

# Warmup: 2 iterations, 10 s each
# Measurement: 6 iterations, 10 s each
# Timeout: 10 min per iteration
# Threads: 3 threads, will synchronize iterations
# Benchmark mode: Average time, time/op
# Benchmark: com.iota.iri.benchmarks.dbbenchmark.RocksDbBenchmark.deleteBatch
# Parameters: (numTxsToTest = 10)

Checklist:

• My code follows the style guidelines for this project
• I have performed a self-review of my own code

[kwek20]

Its not actually faster now right? Youre just able to change settings to make it faster :D
Or does the .threads() change anything? I assumed it would always take all processors

[GalRogozinski]

True,
Devops said that they want to accelerate the test by utilizing more cpus

I am still waiting for their feedback