TieredCheckpointScenario
Data scenarios · Implementation: mlsysim/engine/scenario_definitions/data.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
One save of workload’s state on fleet: local write per node, then a durable drain.
Notes
Equation: Compose mlsysim.physics.quantities.memory_from_params, mlsysim.solvers.TieredCheckpointModel.
Assumptions: The stated fields are inputs to mlsysim.solvers.TieredCheckpointModel; its documented validity regime, units, and modeling choices apply unchanged.
Excludes: This scenario adds no measurement, empirical calibration, or mechanisms beyond mlsysim.solvers.TieredCheckpointModel.
References
Derivation: this scenario binds its input fields to the referenced implementation contracts: mlsysim.physics.quantities.memory_from_params, mlsysim.solvers.TieredCheckpointModel. The methods below identify any additional arithmetic.
Examples
>>> from mlsysim import Scenarios
>>> case = Scenarios.Gpt3TieredCheckpoint2K
>>> type(case.result).__name__
'TieredCheckpointResult'Attributes
| Name | Description |
|---|---|
| bytes_per_param | |
| durable_bandwidth | |
| fleet | |
| local_bandwidth | |
| result | Evaluate this scenario through mlsysim.solvers.TieredCheckpointModel.solve. |
| save_interval | |
| workload |
workload
Type: Workload.
Declared default: required.
bytes_per_param
Type: Quantity.
Declared default: required.
fleet
Type: Fleet.
Declared default: required.
local_bandwidth
Type: Quantity.
Declared default: required.
durable_bandwidth
Type: Quantity.
Declared default: required.
save_interval
Type: Quantity.
Declared default: required.
result
engine.scenario_types.TieredCheckpointScenario.result
Evaluate this scenario through mlsysim.solvers.TieredCheckpointModel.solve. Equation: pass the stated fields to that implementation; its contract defines the calculation and feasibility tests. Units: each returned field follows the referenced result schema; no independent equation is evaluated by this wrapper.