Enum anira_rt_verdict#

enum anira_rt_verdict#

Whether a plan keeps up with the stream on this machine (anira_plan_info.rt_verdict), from its real-time factor over the inferences that run at once: the load, rt_factor / rt_parallel, the share of the session’s parallel capacity one hop’s inference takes. An estimate at prepare, measured under that moment’s load: below the limit a likelihood, above it a certainty. anira owns the thresholds, so every host applies the same rule; an offline render, which does not run against the clock, is never gated by it.

Values:

enumerator ANIRA_RT_UNMEASURED#

No pass and no measured figure (an explicit budget with ANIRA_WARMUP_NONE or FIXED 0): rt_factor is the declared budget’s, a statement about the contract, not the machine.

enumerator ANIRA_RT_COMFORTABLE#

Load below 0.75: misses are unlikely while the machine stays as it was at the prepare.

enumerator ANIRA_RT_MARGINAL#

Load from 0.75 to below 1: the stream keeps up on average, but load the measurement could not see (other handlers, other applications, the clock speed) can push it over: misses are likely under load.

enumerator ANIRA_RT_NOT_REALTIME#

Load 1 or more: the pool cannot finish the inferences the stream submits, so misses are certain; the miss policy fills them. The session is sized for one host block’s burst; the prepare does not fail.

enumerator ANIRA_RT_VERDICT_FORCE32#