Open Models Almanac
Methodsinference & servinginference scheduling

single mention · inference & serving

Priority-based, topology-aware GPU job scheduling

Uses priority queueing and topology-aware placement to keep jobs within one InfiniBand domain.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference scheduling for the methods filed beside it.

Filed under inference & serving › inference scheduling