PeakRatioRangeScenario
Optimization scenarios · Implementation: mlsysim/engine/scenario_definitions/optimization.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
One peak rate against a baseline stated as a range, baseline_low to baseline_high.
Notes
Equation: The calculation members below define ratio_low, ratio_high.
Assumptions: Both baseline endpoints are positive rates measuring the same work as rate.
Excludes: Uncertainty distributions within the range and sustained efficiency.
References
Derivation: One peak rate against a baseline stated as a range, baseline_low to baseline_high. 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.H100OverReferenceCpu
>>> value = case.ratio_low
>>> print(f"{value:.3g}")
494Attributes
| Name | Description |
|---|---|
| baseline_high | |
| baseline_low | |
| rate | |
| ratio_high | Upper dimensionless rate ratio, using the lowest baseline rate. |
| ratio_low | Lower dimensionless rate ratio, using the highest baseline rate. |
rate
Type: Quantity.
Declared default: required.
baseline_low
Type: Quantity.
Declared default: required.
baseline_high
Type: Quantity.
Declared default: required.
ratio_low
engine.scenario_types.PeakRatioRangeScenario.ratio_low
Lower dimensionless rate ratio, using the highest baseline rate.
Against the high baseline.
Equation: (rate / baseline_high).to('').magnitude.
ratio_high
engine.scenario_types.PeakRatioRangeScenario.ratio_high
Upper dimensionless rate ratio, using the lowest baseline rate.
Against the low baseline.
Equation: (rate / baseline_low).to('').magnitude.