Skip to main content

RecoveryBudgetScenario

Reliability scenarios · Implementation: mlsysim/engine/scenario_definitions/reliability.py.

Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.

Cold restarts of workload after failures of n_devices components of component_mtbf.

recovery is the four-term budget profile (Systems.Reliability.Recovery). Added 2026-10-20.

Notes

Equation: Compose mlsysim.physics.quantities.memory_from_params, mlsysim.physics.reliability.calc_mtbf_cluster, mlsysim.solvers.RecoveryBudgetModel.

Assumptions: State size uses the specified bytes-per-parameter convention and independent component hazards determine failure cadence.

Excludes: Unlisted restart stages, correlated faults, and runtime measurement are absent.

References

Derivation: this scenario binds its input fields to the referenced implementation contracts: mlsysim.physics.quantities.memory_from_params, mlsysim.physics.reliability.calc_mtbf_cluster, mlsysim.solvers.RecoveryBudgetModel. The methods below identify any additional arithmetic.

Examples

>>> from mlsysim import Scenarios
>>> case = Scenarios.Gpt3RecoveryBudget1K
>>> str(case.checkpoint_bytes)
'2450000000000.0 B'

Attributes

Name Description
bytes_per_param
checkpoint_bytes State bytes that must be restored after failure. Equation:
component_mtbf
n_devices
recovery
result Evaluate this scenario through mlsysim.solvers.RecoveryBudgetModel.solve. Equation:
workload

workload

Type: Workload.

Declared default: required.

bytes_per_param

Type: Quantity.

Declared default: required.

recovery

Type: Any.

Declared default: required.

component_mtbf

Type: Quantity.

Declared default: required.

n_devices

Type: int.

Declared default: required.

checkpoint_bytes

engine.scenario_types.RecoveryBudgetScenario.checkpoint_bytes

State bytes that must be restored after failure. Equation: mlsysim.physics.quantities.memory_from_params(parameters, bytes_per_param). Units: bytes; the specified state convention determines what is included.

result

engine.scenario_types.RecoveryBudgetScenario.result

Evaluate this scenario through mlsysim.solvers.RecoveryBudgetModel.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.

Back to top