Patent · US Active

Application programming interface to wait on matrix multiply-accumulate

US12204897B2 · kind B2 · utility

0Cited by

7References

20Claims

0Family size

Assignee

NVIDIA Corporation · US

Inventors

Harold Carter Edwards · Campbell, US
Kyrylo Perelygin · Santa Clara, US
Maciej Piotr Tyrlik · Durham, US
Gokul Ramaswamy Hirisave Chandra Shekhara · Kanchinakote, IN
Balaji Krishna Yugandhar Atukuri · San Jose, US
Rishkul Kulkarni · Austin, US
Konstantinos Kyriakopoulos · Weinsberg, DE
Edward H. Gornish · Palo Alto, US
David A. Berson · Marietta, US
Bageshri Sathe · Pune, IN
James Player · Campbell, US
Aman Arora · Bengaluru, IN
Alan Kaatz · Santa Clara, US
Andrew Kerr · San Francisco, US
Haicheng Wu · Cary, US
Cris Cecka · San Jose, US
Vijay Thakkar · Boston, US
Sean J. Treichler · Sunnyvale, US
Jack Choquette · Palo Alto, US
Aditya Avinash Atluri · Redmond, US
Apoorv Parle · San Jose, US
Ronny Meir Krashinsky · San Francisco, US
Cody Addison · Cedar Park, US
Girish Bhaskarrao Bharambe · Pune, IN

Key dates

Filing date	Nov 30, 2022
Grant date	Jan 21, 2025
Priority date	—
Expiry date	Nov 30, 2042

Classification

Technology area (CPC G)Physics
CPC primaryG06F17/16
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

Apparatuses, systems, and techniques to perform computational operations in response to one or more compute uniform device architecture (CUDA) programs. In at least one embodiment, one or more computational operations are to cause one or more other computational operations to wait until a portion of matrix multiply-accumulate (MMA) operations have been performed.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.