Nvidia Corporation

Yogesh Kulkarni

Gongyu Zhou

Oren Rubinstein

Thomas Melvin Ogletree

Saket Arun Jamkar

Abhijat Ranade

Shounak Kamalapurkar

Zheng Yu Zheng

Yudong Tan

A processing unit can include a performance monitor for monitoring the performance of the processing unit and associated sub-units. The performance monitor includes a logic analyzer, and implements a state machine via state machine data entries stored in a memory associated with the performance monitor. A state machine data entry includes output signals associated with state transitions. The output signals include a next state and a trigger to the logic analyzer. The performance monitor implements logic circuits that determine, based on input signals and the state machine data entries, the next state to transition and associated output signals. If a state transition includes a trigger to the logic analyzer, the trigger is transmitted to the logic analyzer. In response to the trigger, the logic analyzer assembles and samples input signals and stores the sampled input signals into the memory associated with the performance monitor, overwriting the state machine data entries.

Shared structure for a logic analyzer and programmable state machine

A processing unit can include a performance monitor for monitoring the performance of the processing unit and associated sub-units. The performance monitor can include a state machine. The state machine can be implemented via state machine data entries stored in a memory associated with the performance monitor. A state machine data entry includes information indicating a state transition condition and output signals. The state transition condition includes a current state and input signals required to meet the condition. The output signals include a next state, one or more counter actions, and one or more triggers. The performance monitor implements logic circuits that determine, based on input signals and the state machine data entries, the next state to transition and associated output signals. The state machine data entries can be written and re-written by a user.

Programmable state machine for a hardware performance monitor

An integrated circuit, in accordance with embodiments of the present technology, includes a plurality of engines, a plurality of engine level power gating (ELPG) controllers, and a power gating arbiter for implementing engine level power gating arbitration techniques. The power gating arbiter may receive requests from one or more ELPG controllers to turn on their respective engines or portions therein. The power gating arbiter prioritizes the request and sends an acknowledgment to a given ELPG controller to turn on or off its corresponding engine according to the prioritized predetermined order. After receiving the acknowledgement, the given ELPG controller turns on or off its corresponding engine and returns an indication to the power gating arbiter that the corresponding engine is turned on or off. The process may be iteratively repeated for each received request after receiving the indication from the previously serviced ELPG controller that its corresponding engine is turned on or off.

Engine level power gating arbitration techniques

Growing community of inventors

Sunnyvale, CA, United States of America

Yogesh Kulkarni