The architecture of the runtime and driver that let host software load and run programs on the accelerator: memory management, scheduling, communication, error handling and the interfaces to firmware and frameworks.
Architecture Work product in the AI accelerator lifecycle
The runtime and driver sit between customer software and the hardware. Fixing their architecture with the hardware architecture makes sure the hardware provides what the software needs and lets software work start before silicon.
The full lifecycle detail, with its charts, flows, tables and every section, is for registered members. Registration is free and takes a minute. You come back to this page afterwards.