Module stack 

Opt-in, shared autotuning above individual kernels.

Tensor operators and fused graphs use the same controller as explicit application plans. Existing LocalTuner behavior is unchanged until enable_stack_autotune is called; only sets declaring an exact workload signature and reference participate in the new path.

Structs

Candidate

CandidateReport

Decision

Problem

RuntimeEnvironment

Supply extra tags for a deployment’s compiler flags and load/topology/power regime. Set these BEFORE constructing devices/starting inference; changing them live is unsupported.

StackPolicy

StackTuner

No state mutex is held while running a candidate, validator, profile, or disk operation. Contenders use the declared reference instead of waiting (also avoids nested-tuning deadlocks).

Stats

Tolerance

TuneFailure

TuneReport

Enums

DecisionSource

FailureKind

Mode

Scope

Timing

Validation

Traits

TrialRunner

Benchmarks MUST use isolated state. A completed measurement includes all device work whose cost belongs to the plan. A submit-only host timestamp is not a valid measurement.

Functions

cache_key

enable_stack_autotune

Install once, before worker threads/model loading. No implicit environment mutation or I/O.

fields

Canonical, length-delimited encoding prevents concatenation aliases such as (ab,c)/(a,bc).

is_tuning

median

paired_score

runtime_environment

stack_autotuner

try_execute_stack

Fallible full-stack path. Actual request execution is performed ONCE after selection; an execution error invalidates future choices but is never silently replayed on another kernel.

workload_digest

This is NOT a cryptographic authenticity check. Caches must be in a trusted user directory.