Packages

final case class CausalProfiler(iterations: Int = 100, candidateSelector: (String) => Boolean = _ => true, samplingPeriod: zio.Duration = 20.millis, minExperimentDuration: zio.Duration = 1.second, experimentTargetSamples: Int = 30, warmUpPeriod: zio.Duration = 30.seconds, coolOffPeriod: zio.Duration = 2.seconds, zeroSpeedupWeight: Int = 10, maxConsideredSpeedUp: Int = 100, sleepPrecision: zio.Duration = 10.millis) extends Product with Serializable

A causal profiler inspired by https://github.com/plasma-umass/coz, that can be used to instrument code and automatically figure out which areas should be optimized to increase overall throughput.

The profiler will periodically select regions of code for optimization and will artifically speed them up by slowing down all code that runs concurrently with it. It will determine throughput based on the number of times that zio.profiling.causal.CausalProfiler.progressPoint has been invoked during an experiment.

Only zio.profiling.CostCenter will be considered for optimization, so make sure to tag the code.

iterations

the number of experiments to perform

candidateSelector

when a new experiment is about to be performed, this is used to decide on a candidate for optimization

samplingPeriod

how often data about fiber locations gets collected. Lower values increase the overhead of the profiler.

minExperimentDuration

minimum duration of an experiment

experimentTargetSamples

number of times every progress point should be invoked during an experiment. will be used to adjust duration of the next experiment.

warmUpPeriod

period of time to delay start of the profiler.

coolOffPeriod

period of time to wait between experiments.

zeroSpeedupWeight

rate of selecting no speedup for the experiment. A zeroSpeedupWeight of n means that that every nth experiment will have no speedup applied.

maxConsideredSpeedUp

maximum speedup percentage that will be selected for experiments. Values over 100 do not have any practical use.

sleepPrecision

estimated precision of Thread.sleep. The profiler will busy spin once it enters this proximity to the target time in order to increase accuraccy.

Linear Supertypes
Serializable, Product, Equals, AnyRef, Any
Ordering
  1. Alphabetic
  2. By Inheritance
Inherited
  1. CausalProfiler
  2. Serializable
  3. Product
  4. Equals
  5. AnyRef
  6. Any
  1. Hide All
  2. Show All
Visibility
  1. Public
  2. Protected

Instance Constructors

  1. new CausalProfiler(iterations: Int = 100, candidateSelector: (String) => Boolean = _ => true, samplingPeriod: zio.Duration = 20.millis, minExperimentDuration: zio.Duration = 1.second, experimentTargetSamples: Int = 30, warmUpPeriod: zio.Duration = 30.seconds, coolOffPeriod: zio.Duration = 2.seconds, zeroSpeedupWeight: Int = 10, maxConsideredSpeedUp: Int = 100, sleepPrecision: zio.Duration = 10.millis)

    iterations

    the number of experiments to perform

    candidateSelector

    when a new experiment is about to be performed, this is used to decide on a candidate for optimization

    samplingPeriod

    how often data about fiber locations gets collected. Lower values increase the overhead of the profiler.

    minExperimentDuration

    minimum duration of an experiment

    experimentTargetSamples

    number of times every progress point should be invoked during an experiment. will be used to adjust duration of the next experiment.

    warmUpPeriod

    period of time to delay start of the profiler.

    coolOffPeriod

    period of time to wait between experiments.

    zeroSpeedupWeight

    rate of selecting no speedup for the experiment. A zeroSpeedupWeight of n means that that every nth experiment will have no speedup applied.

    maxConsideredSpeedUp

    maximum speedup percentage that will be selected for experiments. Values over 100 do not have any practical use.

    sleepPrecision

    estimated precision of Thread.sleep. The profiler will busy spin once it enters this proximity to the target time in order to increase accuraccy.

Value Members

  1. final def !=(arg0: Any): Boolean
    Definition Classes
    AnyRef → Any
  2. final def ##: Int
    Definition Classes
    AnyRef → Any
  3. final def ==(arg0: Any): Boolean
    Definition Classes
    AnyRef → Any
  4. final def asInstanceOf[T0]: T0
    Definition Classes
    Any
  5. val candidateSelector: (String) => Boolean
  6. def clone(): AnyRef
    Attributes
    protected[lang]
    Definition Classes
    AnyRef
    Annotations
    @throws(classOf[java.lang.CloneNotSupportedException]) @native() @IntrinsicCandidate()
  7. val coolOffPeriod: zio.Duration
  8. final def eq(arg0: AnyRef): Boolean
    Definition Classes
    AnyRef
  9. val experimentTargetSamples: Int
  10. final def getClass(): Class[_ <: AnyRef]
    Definition Classes
    AnyRef → Any
    Annotations
    @native() @IntrinsicCandidate()
  11. final def isInstanceOf[T0]: Boolean
    Definition Classes
    Any
  12. val iterations: Int
  13. val maxConsideredSpeedUp: Int
  14. val minExperimentDuration: zio.Duration
  15. final def ne(arg0: AnyRef): Boolean
    Definition Classes
    AnyRef
  16. final def notify(): Unit
    Definition Classes
    AnyRef
    Annotations
    @native() @IntrinsicCandidate()
  17. final def notifyAll(): Unit
    Definition Classes
    AnyRef
    Annotations
    @native() @IntrinsicCandidate()
  18. def productElementNames: Iterator[String]
    Definition Classes
    Product
  19. def profile[R, E](zio: ZIO[R, E, Nothing])(implicit trace: Trace): ZIO[R, E, ProfilingResult]

    Profile a program and get the result.

    Profile a program and get the result. The program must run long enough for the profiler to complete and will be interrupted once the profiling is done.

    The profiled program must use the progressPoint function to signal progress to the profiler.

  20. def profileIterations[R, E](zio: ZIO[R, E, Any])(implicit trace: Trace): ZIO[R, E, ProfilingResult]

    Profile a zio program by running it in a loop and using completion of an iteration as the progress metric.

    Profile a zio program by running it in a loop and using completion of an iteration as the progress metric.

    profileIterations {
  val fast = ZIO.succeed(Thread.sleep(20)) <# "fast"
  val slow = ZIO.succeed(Thread.sleep(60)) <# "slow"
  fast <&> slow
}
  21. val samplingPeriod: zio.Duration
  22. val sleepPrecision: zio.Duration
  23. def supervisor(implicit trace: Trace): ZIO[Scope, Nothing, Supervisor[ProfilingResult]]

    Create a supervisor that can be used to profile a zio program.

    Create a supervisor that can be used to profile a zio program. Getting the value of the supervisor will suspend until profiling to complete. For profiling to work correctly, OpSupervision must be enabled for the effect.

  24. final def synchronized[T0](arg0: => T0): T0
    Definition Classes
    AnyRef
  25. final def wait(arg0: Long, arg1: Int): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws(classOf[java.lang.InterruptedException])
  26. final def wait(arg0: Long): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws(classOf[java.lang.InterruptedException]) @native()
  27. final def wait(): Unit
    Definition Classes
    AnyRef
    Annotations
    @throws(classOf[java.lang.InterruptedException])
  28. val warmUpPeriod: zio.Duration
  29. val zeroSpeedupWeight: Int

Deprecated Value Members

  1. def finalize(): Unit
    Attributes
    protected[lang]
    Definition Classes
    AnyRef
    Annotations
    @throws(classOf[java.lang.Throwable]) @Deprecated
    Deprecated

Inherited from Serializable

Inherited from Product

Inherited from Equals

Inherited from AnyRef

Inherited from Any

Ungrouped