MeanVersusTailScenario
Statistics scenarios · Implementation: mlsysim/engine/scenario_definitions/statistics.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
A benchmark’s mean latency against a production tail stated as a range.
Notes
Equation: The calculation members below define gap_low, gap_high.
Assumptions: Mean and tail endpoints are supplied latency quantities with positive mean.
Excludes: Inferring a tail distribution or equating benchmark and production workloads.
References
Derivation: A benchmark’s mean latency against a production tail stated as a range. Each member records the corresponding algebra or named physics function.
Examples
Evaluate the named configuration and display a modeled result.
>>> from mlsysim import Scenarios
>>> case = Scenarios.ModelOnlyMeanVersusProductionP99
>>> value = case.gap_low
>>> print(f"{value:.3g}")
10Attributes
| Name | Description |
|---|---|
| gap_high | Upper production-tail latency divided by benchmark mean, dimensionless. |
| gap_low | Lower production-tail latency divided by benchmark mean, dimensionless. |
| mean | |
| tail_high | |
| tail_low |
mean
Type: Quantity.
Declared default: required.
tail_low
Type: Quantity.
Declared default: required.
tail_high
Type: Quantity.
Declared default: required.
gap_low
engine.scenario_types.MeanVersusTailScenario.gap_low
Lower production-tail latency divided by benchmark mean, dimensionless.
Equation: (tail_low / mean).to('').magnitude.
gap_high
engine.scenario_types.MeanVersusTailScenario.gap_high
Upper production-tail latency divided by benchmark mean, dimensionless.
Equation: (tail_high / mean).to('').magnitude.