Title:
HARDWARE CIRCUIT FOR DEEP LEARNING TASK SCHEDULING
Document Type and Number:
WIPO Patent Application WO/2022/246639
Kind Code:
A1
Abstract:
Apparatuses, systems, and techniques for scheduling deep learning tasks in hardware are described. One accelerator circuit includes multiple fixed-function circuits that each processes a different layer type of a neural network. A scheduler circuit receives state information associated with a respective layer being processed by a respective fixed-function circuit and dependency information that indicates a layer dependency condition for the respective layer. The scheduler circuit determines that the layer dependency condition is satisfied using the state information and the dependency information and enables the fixed-function circuit to process the current layer at the respective fixed-function circuit.
More Like This:
Inventors:
ZHOU YAN (CN)
ZHANG YILIN (CN)
CHEN GENG (CN)
ZHOU YAN (CN)
FAN QIFEI (CN)
GAIKWAD PRASHANT (US)
ZHANG YILIN (CN)
CHEN GENG (CN)
ZHOU YAN (CN)
FAN QIFEI (CN)
GAIKWAD PRASHANT (US)
Application Number:
PCT/CN2021/095739
Publication Date:
December 01, 2022
Filing Date:
May 25, 2021
Export Citation:
Assignee:
NVIDIA CORP (US)
ZHOU YAN (CN)
ZHOU YAN (CN)
International Classes:
G06F7/57
Foreign References:
CN107844833A | 2018-03-27 | |||
CN108256636A | 2018-07-06 | |||
CN109409510A | 2019-03-01 | |||
US20200357099A1 | 2020-11-12 |
Attorney, Agent or Firm:
P. C. & ASSOCIATES (CN)
Download PDF:
Previous Patent: M-TYPE HIDDEN CONNECTING PIECE
Next Patent: DISPLAY SUBSTRATE, FABRICATION METHOD THEREFOR, AND DISPLAY APPARATUS
Next Patent: DISPLAY SUBSTRATE, FABRICATION METHOD THEREFOR, AND DISPLAY APPARATUS